Kodachrome 21

Summer of Film

2011-09-29T06:34:00.000-07:00

Earlier this summer, I struggled with my photography. I bought into a new camera system, the Sony SLT-A55, which I used at my grandfather's 100th birthday and our trip to Disney World. I was generally happy with the photos, and the camera performed fantastically at the difficultly-lit birthday party, but the photos (while good) lacked something I couldn't put my finger on... The modern "travel zoom" SEL18-250 was a real wonder, with a fantastically useful zoom range and generally excellent performance, but the images often lacked a certain je ne se quois I search for. It was a bit "heavy handed" in its manner of drawing, and I overall preferred it for people photography over documentary and travel, oddly enough. I did get some photos I love, as with any camera I own; like this one:

My interest in black and white photography was rekindled, and I wasn't satisfied with the results as above--DxO film pack gets close, but not quite "real," in my opinion.

I stopped off at an Olympus XZ-1 for everyday photos, and struggled getting what I liked out of the small sensor. The lens was really great, though, and the camera certainly takes the "prettiest" small-sensored photos I've seen since my G9..

That said... I wasn't thrilled by it. I returned the camera.

One day I was down in my basement rearranging items in our deep freezer and found a box... Inside was at least 50 rolls of color film I bought at Costco back in 2009. I got out an old Minolta SRT-101 with a Rokkor 135/2.8, and went to a park... Sasha at the park:

I felt at home using film again. I loved almost the whole roll of images. In the meantime, searching for a "digital fix," I bought a Minolta 50/1.7 for my Sony A55, which I was VERY pleased with. Granted, my copy was sitting new in its box (at a local shop for 20+ years!) and in perfect shape, but I really enjoyed it:

Although I loved that lens, ~75mm doesn't really work as my "only" lens. So I searchd for more older Minolta lenses... I ended up with a 35-70 F4 from the "beercan" era. I was instantly reminded of older vintage lenses I used to use on my Nikkor... There was a soft feeling underlaid with a quick-drawing sharpness characteristic of lenses with a little more aberrations wide open than modern lenses. The resulting feel was sometimes a bit more pastel, more 3-dimensional and "rounded" than a hard-drawing modern lens. I liked it:

In the meantime, I shot more and more film. After realizing that Costco not only develops and scans color but will scan any negative you bring in, I decided to get back into developing film myself and grabbed an old roll of Tri-X... I was hooked.

I stopped by my favorite local shop one afternoon, Chick's Camera Exchange, and had a talk with Steve (The owner). He had set aside a Leica M3 for me to look at. I had been lusting after Leica's since I tried one in college, on loan from a professor... The few rolls I shot back then stuck in my head as being everything I wanted from photography, and I knew I had to have one someday. For years I lied to myself that digital was "better" and SLR's were "the truth." 3 seconds into handling that Leica M3 I knew this was my camera. I bought it, along with a 50mm 1.4 Summilux II and 135/4.5 Hektor, and later added a 35/2 Summicron (with "goggles" for my M3). I sold all my digital cameras except the fantastic and fantastically convenient Sony HX9V. A veritable orgy of film has followed, and I've enjoyed every minute of it. I've had the best summer of photography since my first summer after college, shooting with my dearly departed (stolen!!) Minolta XD11 and Rokkor 50/1.7. The Summilux is such a fantastic lens.. It pulls in light like no other, and even wide open draws sharply with excellent contrast. Bokeh can be a little "articulated" for modern tastes, but I love it. Reportage:

Portraiture:

I am just marveled at how my "hit rate" has grown from a low of 10-15% with digital to 80%+ with the Leica M3... And this is even with the fact I have no meter on the M3 (and thus "meter in my head" or from a quick reference meter on my iPhone) and have to manually focus. I absolutely adore this camera.

Since that time I've bought 3 screwmount bodies (looking for 1 good one!) and some screwmount lenses (with adapter to use on M3, also): 50/2 Summar, 50/2 Summitar, 50/1.5 Summarit, and 35/3.5 Summaron with external finder.. I also bought and returned a Zeiss Biogon 21mm 2.8; I loved the focal length, didn't like the modern rendering...

Because my M3/Summilux/Summicron combo is worth a fair amount of money, I decided to chase down a screwmount setup for our recent family vacation to Kiawah Island. Although the screwmount camera and lenses had more limitations than the M mount stuff I own, I totally enjoyed the experience, and resulting photos. More to come on that, and individual lens reviews (as I learn the characteristics and pros/cons of each lens) soon. I'm shooting film until it sunsets...

Olympus XZ-1: the real successor to the Canon G9? (Review Part I)

2011-07-05T10:16:00.000-07:00

Excepting the dark ages of digital, when digital SLR's were new, very expensive, and I didn't own one, I've never been a big fan of compact cameras. The first compact camera I really loved was my Canon G9, which I used for quite a few years to great success. It was the first compact camera that I really "relied" upon during vacations, and I actually left my DSLR at home on 3 separate trips (a trip to Vermont, Wyoming, and Disney World) to go light and fleet with the G9. I get this feeling that I'm "working" sometimes when I'm using a DSLR, and would rather relax and enjoy my vacation than "work the photos." A compact camera kinda lets me do that, although it must still have good enough handling to please a photographer, and I would prefer it to have good image quality, as I sometimes like prints of my vacation photos.

My Canon G9 did quite well in that regard. I got to know the camera quite wel, and I'd say its output at base ISO in good light was shockingly close to a good DSLR's, as long as you shot RAW and minded the exposure. In poor light, the G9 fell down... But no matter, it was so freeing to be able to travel with a camera in my pocket rather than a camera slung over my shoulder. I worked within the limitations the G9 imposed, and when I did so, the images rarely failed to please.

Now, on a one-in-a-lifetime trip like my trip to Europe in 2007, or our trip to China to adopt our daughter in 2009, I didn't take a compact. I lugged a big 'ol DSLR (Nikon D200 to Europe and Canon 5D Mark II to China), and loved every minute of it. I knew I'd be taking photos in tough lighting, and didn't want to compromise with a compact. I'm glad I didn't. Sometimes, you really do need a "big gun."

The G9 wasn't without its downfalls, though. I found it hard to nail the metering sometimes, and the dynamic range of the sensor was very slim, like slide film; skies were often hopelessly lost in a sea of white when overexposed. The autofocus was OK, but not totally up to keeping up with moving kiddos. ISO 80 was fantastic, anything over that truthfully went downhill fast. ISO 400 was as high as I'd shoot it, and much of the wonderful 3-dimensional quality of its images was lost at that point. Its lens was only 35mm on the wide end--great for everyday "documentary" photos, but hardly wide enough to capture that sweeping vista you see in front of you on vacation. It was a little sluggish when shooting RAW files, which I did almost all the time. Generally, it was just getting a little long in the tooth. So I bought the newfangled Sony NEX5 to replace it. APS-C sensor in a package (with pancake) not a ton bigger than the G9? I thought I'd be in heaven.

But I wasn't. You'll notice I didn't blog that year... To be honest, I didn't take many photos! Ok, so having a young baby certainly curtailed my photo taking (and processing!) time, and I got off on a film tangent (very enjoyable, BTW)... But the fact remains I often didn't take a camera with me... So got not photos, obviously! The NEX5 with pancake may have been generally pocketable, but I found I didn't really like 24mm all that much, and with the kit zoom it was MUCH larger than my old jeans-pocket G9. I happened to grab a Sony HX5V last summer, and really enjoyed that camera. I upgraded to the new HX9V this summer, which has been an improvement. Generally I've been very impressed with the HX9V.

However impressed I am with the HX9V's photos, it's not the panacea of small cameras for me. To get the best out of it, I can't shoot in aperture priority and do all my standard mental calculations to make photos look the way I want them to; in fact, there is no aperture priority mode, and P and M modes don't allow you to take advantage of some of that fantastic sensor wizardry (image stacking to reduce noise and create in-camera HDR images). So I "ride the modes" and shoot iAuto, Superior Auto, Custom (with HHT, AMB, and HDR setup as my 3 settings), or "Background Defocus" mode. It's just a different experience, and one that impinges a bit on my creativity. Never mind there's no real control of depth of field (outside of Background Defocus mode) unless you're at silly long focal lengths... And the photos, although quite good (FANTASTIC when printed, to tell you the truth!), don't have quite the "snap" that the G9's did.

So I started looking for another small camera to supplement the HX9V. I don't intend to get rid of the HX9V, as it's fantastic at what it does and still the best camera for me at times. But I wanted something more centered on the classic photographer, hoped for some control over depth of field, and hoped to get slightly better image quality to boot. Although I truly enjoy the freedom of the HX9V's 24-385mm zoom, I'd be fine with something around the standard DSLR "kit" 28-80mm.

I considered the Canon G12, Panasonic LX5, Samsung EX-1, and Olympus XZ-1. I settled on the Olympus. The G12 was a front runner, and IMO best at high ISO's (800 & 1600), but honestly I felt wasn't as good as the G9 at base ISO. The LX5's colors were "off" to my eyes. The Samsung had horrendous JPEG processing (soft!), and no support in my RAW software of choice (DxO), as well as the fact it was the largest of the 4. The Olympus had best base ISO image quality, and provided a modicum of control over depth of field with its fast lens.

After shooting with it this July 4th Weekend, I am happy to say that the Olympus XZ-1 files remind me very much of the Canon G9's at base ISO, and perhaps are even better. The camera itself handles a bit better; focus is faster, RAW recording doesn't seem to slow it down, and it does do a pretty good "disappear out of the photographer's way" to let me shoot what my mind's eye sees.

I'm even enjoying the Black and White I get out of the XZ-1. Here's one straight from the camera:

And one with mild tweaks in Olympus Camera Viewer (DxO doesn't read XZ-1 yet, either!):

I love the color output, also, it has great skin tones:

The color output is a little like my old Canon 5D Mark II, especially in the blues and greens; but the reds and yellows are a little different, and very pleasing. Operationally, the camera is quite responsive and although focus isn't fantastically quick like the HX9V, you are able to capture candid moments fairly well:

You can control depth of field somewhat with the very fast lens, although don't expect DSLR-levels of background blur, and bokeh isn't the smoothest I've seen:

Bokeh isn't bad, by any stretch, but sometimes out of focus areas of high contrast can show up a little harsh:

The lens is capable of smooth bokeh, though, and here's an example of a photo that would have totally failed with the HX9V's Background Defocus mode:

Not the best photo, by any stretch, I was just playing around... But the transition from in focus to out of focus is quite smooth, and generally bokeh here is nice, IMO. And here, at a short focal length:

Generally the lens does a nice job setting off the subject when wide open, and you can accentuate this with the in-camera "Pin Hole" vignetting filter:

All in all, I'm pretty impressed with the XZ-1. Sadly, it's not nearly as fantastic indoors in low light; if you can keep the ISO down to 200, it will please greatly... But as the ISO creeps up to 400, quality takes a noticeable nosedive. Perhaps if I used a different RAW processor I'd be happier with 400-800, hopefully DxO will add support at some point. Today, though, I will shoot the XZ-1 much like I shot the G9--400 and under only, and preferably base ISO if at all possible. Luckily the fast lens actually opens up more opportunities for this than you'd think; I did a test inside a dim church with the Sony HX9V's HHT mode vs the XZ-1, and the XZ-1's output was noticeably better; the Olympus was able to shoot at ISO 200 whereas the HX9V needed ISO 800 due to the slower lens. I printed both at 12x18, and the Olympus photo was noticeably nicer. That said... The XZ-1 didn't really touch the NEX5, which also shot at ISO 800 (with the slow kit lens), but had the benefit of image stacking HHT mode which rendered an almost noise-free shot. In outdoor tests, the XZ-1 generally excelled, but the HX9V was really close and sometimes even better, with far less photographer "effort." That said, I do find the XZ-1 allows me to follow my creative "mind's eye" more, and therefore, the photos are generally noticeably superior.

I feel like I have my G9 back, with a faster lens, better base ISO image quality, and even better handling... In a very similar sized package. Yay! More to follow soon.

Sony HX9V Review Part IV: The Everywhere Pocket Rocket

2011-07-01T14:35:00.000-07:00

Because the HX9V is small enough to fit in almost any pocket, it can go everywhere with you, capturing interesting memories when your DSLR is snug at home in it's case :)

Take this killdeer, for example. I had gone to my dad's house to pick up my daughter after playing some tennis, and happened to hear the distinctive call of a killdeer in his backyard. I had the HX9V in my pocket of my work pants, so I grabbed it to capture a few frames. This is at full optical zoom, and except for a little busy-ness in the bokeh, I was impressed with the HX9V's capture. The HX9V seems a little sharper at full zoom (385mm) than the HX5V was at full zoom (250mm). It is MUCH sharper than the HX5V at 250mm, so this is a bonus all around. Because of diffraction (wide open is f5.9!), it's a little softer than the lower focal lengths, and you will battle atmospheric effects at 385mm, but you can still get some nice images.

I don't tend to use the HX9V at 385mm much, though; to me, the amount of compression at that length renders photographs quite "flat" and uninteresting in many cases. I'm just not very accomplished at using long focal lengths; perhaps I will grow into them over time. One thing's for sure, though, for certain things (wildlife, concerts, etc), it's hugely useful to have. You don't have to use it if you don't want to; so don't just rack it out to 385mm and shoot. Try shorter lengths, too, the frame may be more interesting for it (and it will certainly have more of a 3-dimensional look). Take this frame, for example:

I had been shooting my dad and daughter at near max focal length, and I wasn't thrilled with the frames I had captured. I backed off a bit, and immediately my eye started picking up much more interesting stories to tell.

Because of the flexibility of the HX9V's focal range, I captured a shot of my daughter from my dad's deck above:

Then this scene at a wider focal length:

Then a full zoom scene of Sasha walking along the shore:

That is just a huge focal range to have in your pocket!

As a teaser, some more HX9V stuff coming soon... I printed some 12x18's at Costco of the Sony HX9V vs Sony NEX5 vs Olympus XZ-1... And the results were somewhat surprising! Stay tuned...

Samsung NX100 Quick Review

2011-06-27T10:35:00.000-07:00

As I blogged about last month in my quick Sony HX5V recap, I have been quite pleased with small cameras lately. The HX5V image quality was "good enough" to enjoy photography for its own sake, and certainly good enough to capture family moments. The quality is still quite some way from 35mm film or a good DSLR, but the fact remains that I'm not taking photos if I'm leaving my camera at home... And with a toddler in tow, I found myself leaving my DSLR at home.

While the change from the HX5V to the HX9V was a solid jump forward in image quality and creative control, I'll be the first to admit that shooting with a compact is still an experience very alien to most photographers. When I'm in photo mode, in the back of my head I'm balancing aperture for DoF, shutter speed for motion, and ISO to make the first two work for me. Other than small tweaks to WB, drive mode, or using AEL or exposure compensation, I'm really just concentrating on capturing what my mind's eye sees. A camera that "gets out of my way" is usually what I'm looking for. Compacts aren't quite there yet; there's no real DoF control with a tiny sensor compact with a relatively "slow" lens like my HX9V, so you resort to the "background defocus" setting. Compacts struggle in low light, so I use Handheld Twilight, AntiMotionBlur, or let Superior Auto make up its own mind. Compacts have limited dynamic range so I like the in-camera High Dynamic Range image-stacking of the HX9V to help extend DR somewhat. For the most part, these advanced options work quite well... But I would love to have a more pocketable camera that felt more natural to take photographs with, for fun.

I've owned the Sony NEX5 since it was released, but I've not really gotten on with it. It's not that I don't like the image quality; in fact, I think the 18-55 kit lens (or, at least my copy of it) is quite good. Bokeh, in particular, excels for such a cheap optic... And I really like the light and fast manner in which it draws. The pancake 16/2.8, though, has been a disappointment for me. It's not a terrible lens, in fact I've gotten some images I really like from it, like this grab from my morning walk into work:

Or this shot of my daughter which isn't terribly well framed, but I enjoy, anyway (FWIW, I've found the NEX only tracks focus well center of frame, so you either crop later, or leave it like this...):

The problem I have with the SEL1628 is that I'm not a huge fan of 24mm as an "all around" lens. 24 is SO much wider than 28 or 30... Or my all-time favorite walk around ~35mm equivalent... Yes, the SEL1628 isn't super high quality, giving fairly soft results compared to the kit lens, lots of distortion, vignetting etc... But the real problem is ME, not the lens. I struggle with all that a 24mm pulls into a frame. Certainly as an option the 24mm is fun and fine; but Sony hasn't seen fit to release ANY MORE PANCAKES for the e-mount NEX's which I think is insane. Check out the ridiculous Sony lens roadmap here. Where are the pancakes? Why would I want DSLR-sized lenses on my NEX?? I already found myself either grabbing a compact, or a DSLR... and never the NEX unless I "told" myself to. I can just fit the NEX5 with the 18-55 in the pocket of my cargo shorts/pants, but in practice I rarely do so. I simply grab the HX9V.

In any case, I noted on Flickr that photos from the Samsung NX cameras and in particular the 30mm f2 pancake were quite good. The 20mm pancake has somewhat harsh bokeh at times, but is sharp as heck... And a very useful focal length. Even the 20-50 and 50-200 shots look quite beyond their price... In particular, the bokeh of most of the Samsung lenses seems quite good. With a 16mm of their own on the way, I thought a NX100 with 16, 20, and 30 pancakes would be quite the fun little art/travel toy, and prices on the NX100 are almost ridiculously low right now. You can get the NX100 with 20-50, flash, and either the 30/2, 20/2.8, or 50-200 for around $600 right now. That's a whole lot of kit for not much more than a Canon G12. So I jumped.

Well... The NX100 had some high points, such as the tremendously configurable picture/color settings. I had a particular fondness for the "calm" picture mode (sorry for the ridiculously bad photo, I promised not to post the late night portrait of my wife that actually came out quite nice, and this was my only other photo taken with this picture mode!)... this one with the 50-200:

In standard mode, colors were quite different from Canon, Nikon, or Sony.. Being closer to my Sony HX9V than any DSLR I've used. But they did have pleasant pop without being too over the top... This one with the 20-50:

I didn't like the black and white rendition at all, though; to my eyes it was either too blueish-gray:

Or generally lacking in the microcontrast that I love about good B&W:

Not that poor B&W in-camera is a dealbreaker for me, but when my A55V does it so well, and so effortlessly... I find it hard to give up:

There is a very "1.0" feeling to the Samsung gear, also. Plastic quality and build is quite a bit behind my Sony HX9V, and updating the firmware and lens firmware was tweaky (you have to load a lens.frm file on the root of the memory card for each lens you want to update, only one at a time please). The "fn" quick access wasn't quick or easy with the reticent scroll wheel on the back of the camera. In fact, in general, the interface and "flow" just didn't fit me well, either. I felt more like I was using a compact camera than a DSLR... The iFunction idea had real promise, as I LOVE manual aperture rings on my film cameras, but the implementation had absolutely no tactile feedback... leaving me prefer just changing options from the menu. And why Samsung chose to make the top wheel NOT control aperture, I'll never know... As that was the best feeling control on the whole camera. Even the shutter was a bit nonintuitive, with a very distinct half press stop. The lenses, also, weren't as refined as current Canon, Nikon, or Sony gear; the 50-200 was pretty good in the build quality department, but was as loose as a 20 year old Quantaray zoom, the lens creep wasn't creep at all but full on race to 200mm as soon as the camera tilted down. The 20-50 felt insubstantial and toylike, the Sony 18-55 has it all over the Samsung here (even the Panasonic 14-42 feels substantially better).

However, I'm usually willing to overlook little flaws like this, especially in a "fun" camera... The problem was, the NX100 wasn't really fun to shoot, and the image quality was decidedly behind the times... I knew the high-ISO photos would be fairly poor, but I really wasn't prepared for how poor they were (or, rather, how far Sony has come on APS-C); my NEX5 and A55V simply blew the Samsung out of the water at any comparable ISO. The Sonys even maintain a somewhat 3-dimensional "feel" to images, even at ISO3200 and up... Whereas the Samsung's image would "flatten" and look very "digital" from anything over 400. I'm not a real stickler for uber-sharp images and/or no noise... But the Samsung's images were just completely uninspiring at higher ISO's.

At base ISO, I could live with it, and honestly, Samsung's lenses (at least the 30/2 and 50-200) have quite good bokeh which I really appreciate. My particular copy of the 20-50 was well below the 50-200 in sharpness; I'm not really a stickler there, but I note it for others who are. It was ok, but worse than the Sony 18-55. Bokeh wasn't really all that great, either, although better than some kit lenses. Samsung is also quite aggressive in pricing, and for that I can forgive some of the "1.0" foibles.

Size-wise, I was surprised how much larger the NX100 was once I had it in-hand. It felt huge compared to the NEX5, and with the 20-50 on it I couldn't even fit it in my cargo pocket without the lens forming an odd protrusion (the nex, on the other hand, actually fits better even with it's longer 18-55, as it's less tall, and sits flush against my leg like it's sitting on a table). With the 30/2 pancake I wanted, it would be much better... And I knew it wouldn't be the size of my HX9V... But once I had it in hand I realized the NX100 wasn't really the answer to my size need, either, as especially with the really very nice 50-200 it wasn't any smaller than my Sony A55V, really.

But the fact remained I really wasn't inspired by this camera... Add in the specific deficiencies vs. my NEX5, and it's the first camera I've ever returned to the store I bought it.

I picked up my A55V later that day to shoot some frames for my birthday... And immediately I knew I had made the right choice... it was so much more intuitive to use, and the color and 3-d "look," even with a cheap old $35 minolta lens, make it so much of a better choice:

That said, I'm still wanting a "pocketable" fun camera, expecially since my wife has appropriated our HX9V since returning from New York City last week. I'm really thinking of selling my NEX, although I'm waiting to see how "pancakey" the 24mm zeiss really ends up being... In the meantime, I have an Olympus XZ-1 on the way, and I'm actually considering the crazy Pentax Q...

Sony HX9V Review Part III: Using the Sony HX9V at a Concert

2011-06-07T06:10:00.000-07:00

This weekend (on June 5, 2011) I saw Phish at Riverbend Amphitheater near Cincinnati, Ohio. I decided to take the new Sony HX9V as I wasn't sure how Riverbend was with SLR's.

Last summer I shot a few concerts with the HX5V. I found it to be decent, but not stellar. The image stacking modes that worked so well in low light didn't work quite as brilliantly at a concert; Handheld Twilight (HHT) mode sometimes didn't stitch the moving main subject together well, and Anti-Motion Blur (AMB) was quite grainy and color was poor. The sensor's dynamic range was quite taxed by the main stage lights and dark surroundings, which meant that exposure had to be nailed the first time around as there wasn't much latitude in post processing. On the plus side, the Smart Zoom (zoom to 350mm at a resolution of 5 megapixels) was quite handy, bringing main subjects close. Overall, I was happy to capture some memories, but not thrilled with the performance:

Earlier in the evening, in better light, results were more pleasing:

I am happy to report that I am more pleased with the HX9V's output. Overall, the camera is still a pocket digital camera with a tiny sensor, and therefore will not be equivalent to a DSLR, and nor should I expect it to be. However, the image stacking HHT and AMB modes worked quite well, producing shots with less noise and more clarity, sharpness, and color than the HX5V's.

Here was my vantage point, taken in Superior Auto mode (which used HHT image stacking), ISO 250:

And about 40mm effective focal length, Superior Auto, ISO1600 (HHT):

That is something new--HHT mode never used to let ISO climb to 1600 on the HX5V, as far as I can recall.

Here is one at approximately 185mm focal length, again Superior Auto, ISO 320:

Here's one at full optical zoom, Superior Auto, ISO 800:

I have found the camera to be very good in Superior Auto mode, inexplicably capturing better, sharper, more colorful images than if I set it directly to HHT mode. I have started using the HX9V on Superior Auto mode quite frequently. Whenever the light comes down, this seems to produce the best results. Daytime I'm not sure of yet, although I've been using Superior Auto a lot, sometimes I get better results by forcing the camera to capture a HDR image.

Through the night at Phish, though, as the light levels decreased, I noted that the Superior Auto shots were starting to show a lot of blurred main subjects. I just shot multiple images for awhile, capturing the sharpest result, but in the end started to get better results after switching to AMB mode.

And later in the show, when it was darker, I forced it into AMB mode using the "MR" setting on the dial, and selecting my saved AMB preset. This one was about 185mm, ISO3200:

Another in AMB mode, ISO3200, at full optical zoom:

Of note, I did not use the Smart Zoom feature as much. I found my results to be, at least on the camera's LCD, about the same as zooming in (cropping) a full 16 megapixel image, so I didn't use it much. Also, the full res images were coming out well enough, I stuck with those. I also think it was difficult to hand hold the full 28x Smart Zoomed images, which would be about 672mm effective focal length!! Perhaps I could have persisted and used AMB solely (as I think I was mostly using smart zoom with Superior Auto, which wasn't detecting subject motion and thereby didn't choose AMB), and garnered good results. Maybe next time.

I also shot some video at the show:

Video results were decent, but the camera really did struggle when the light came down. At the end of the evening, I couldn't get it to reliably focus at full optical zoom, so I gave up and just enjoyed the show. I recall the HX5V being slightly better in this regard. Another thing the HX5V was better at was zoom smoothness in video mode; the HX9V's zoom isn't as smooth when racking in and out. Sometimes, it's very smooth. At other times, it "jumps" a little. That said, especially at wider focal lengths, the video is fairly good in low light and the mics don't distort as easily as the HX5V's did. Also, the wind cut feature seems to work. And the location of the mics is MUCH better--I frequently put my fingers on top of the HX5V's mics. All in all, the video quality of the HX9V isn't as fantastic in low light as it is in good light... But it is SO good in good light that maybe my feelings are skewed here a bit. It did a solid job. If only I could hold the camera more steady :)

One note: every time you turn off the HX9V, your exposure compensation setting is zero'd out. You can save a setting to your "MR" location, which is nice, but if you use Superior Auto, Movie Mode, or any other auto mode frequently... Remember to re-set your EV every time you restart the camera! I often turn it off between shots, so this is something I frequently forgot to do. For concerts, I've found -.7 to -1.3 to be the most commonly needed settings, depending upon stage lighting. I sometimes needed as much as -1.7EV. The HX9V's programmable custom button makes this very easy to access.

I'm quite pleased with this set of images from the HX9V. I never expected DSLR quality images, but the images I got are actually printable I think. I will try to print some at 8x10, 11x14, and maybe even 13x19 to see what they look like, just for fun. I'll post later.

One last photo to leave you with... This one using the new "Background Defocus" mode on the HX9V:

Sony HX9V Part II: Sweep Panorama High Resolution Mode: 43 megapixel images!

2011-06-03T13:36:00.000-07:00

One of the very interesting features about the Sony HX9V is the new "Sweep Panorama High Resolution Mode." This mode allows for a 180-degree panorama shot at 10480 x 4096 pixels (43 megapixels!). This is a big improvement over the HX5V's 4912 x 1920 mode, which seemed to use the video recording feature and produced ok but inferior prints. They were soft, lacking detail, and didn't have the "pop" of the HX5V's standard photos.

All that has changed on the HX9V; the prints are sharp, full of detail, and seem to have much better color and contrast than the HX5V. Here is a panorama I shot at Disney's Saratoga Springs pool:

And here is a 100% crop from the far right side; look for the red umbrella with the hula hoopers on the main pic for the area covered by this crop:

There's a huge amount of detail in these panoramas.

That said, there are some negatives, and most relate back to Sony's image processing. The heavy noise reduction is still present (grasses blend to mush; the HX5V did this also), and although this is a 16-megapixel sensor, and you do get good detail out of it, don't expect DSLR levels of detail and clarity. Still, it's pretty remarkable what it does capture. You could make HUGE prints that would look pretty fantastic.

Panorama's are tough, though, especially 180 degree and greater; stitching errors are almost always found somewhere in the pano shot; the HX9V does a pretty good job with fairly distant people, even moving people, but other moving objects (like the hula hoops) and buildings or other straight lines can prove especially problematic. Generally, the more distant your subjects, the best... Make sure you watch around your feet for curbs, railings, etc, as those almost always stitch poorly. And areas with moving people all around, like the panos I tried to shoot at the Magic Kingdom, rarely came out decent. Too many close moving people and buildings don't work out well for the high resolution mode. It is also somewhat hard to keep the horizon level, so I had to adjust this one and most others a degree or so in post processing (thereby cropping off a few megapixels). In more challenging environs, switch to the lower res mode which 1) isn't 180 degrees), 2) seems to stitch close people a little better, and 3) is easier to keep level.

I think the sweep panorama high resolution mode is a neat option, though, especially for landscapes and vistas... I would love to use this function in the mountains, death valley, grand canyon, yellowstone, etc. It's so easy to do, though, it's worth a try wherever you are. If they don't work out, just grab some 24mm shots as backups (although 24mm will "feel" really restrictive after a 180-degree shot!), which is what I did, as my "keeper" rate was fairly low (maybe 25%). The ones that work, though, have a unique manner of recalling memories of being in a location, which is really nice.

Sony SLT-A55 First Shots Review, Part 2

2011-05-21T13:36:00.000-07:00

In my last post, I was guardedly optimistic about the SLT-A55. Today, I'm much more pleased with it; I'd say I'm very pleased with the body and generally pleased with the SAL18-250 lens.

Last night my daughter Sasha came downstairs wearing a super cute hat, so I grabbed the camera and did a quick "model shoot." Nevermind the spaghetti sauce on her face, and mud from playing in the backyard...

This one I turned into a B&W with DxO's "TriX" film look, partially because I envisioned it that way, and partially because the spaghetti sauce looked really heinous in this shot haha!

Although I didn't nail the composition on this one, I like the shot a lot, and will likely crop it a little:

Being Silly:

My little model:

And from earlier in the evening, I have MUDHANDS!

Focus accuracy was MUCH better if I just turned on Continuous Autofocus. The Automatic AF (which should switch from one shot to continuous on its own) wasn't doing a great job with the unpredictable starts and stops of my toddler. With AF-C on, it was VERY good. I also utilized the 6fps (or is it 7?) drive mode, which was quite useful. I have seen many reviews mention the buffer hangs after letting go of the shutter, but I never felt it laggy. In fact, the camera is a little laggy in one-shot mode, but put it continuous drive, and it's very very snappy--both during the shots (of course) and after. It will lag if you decide to do some chimping... But I don't do that a ton while shooting moving subjects, so to me the camera felt fairly responsive.

I used this knowledge to great effect today when shooting my grandfather's 100th birthday party. I used AF-C and continuous drive with subjects that were moving quite a bit, and AF-A for relatively static folks. Because I was shooting indoors in poor mixed lighting, and ISO1600 was barely enough to keep the shutter speeds up with wide open aperture let alone stopped down, I went back and forth between continous drive and bumping the ISO. With the former, usually one shot out of 2-4 was sharp (others would sometimes show motion blur), the latter of course dropped the quality but always got the shot. My feeling is this camera is a little worse at ISO6400 than my old Canon 5D Mark II, which I have shot in the same location, so I'd like either a faster lens than the 18-250, a flash, or I'll simply use continuous to get some keepers. That's a great photog. tip, BTW, for when you're in a bind and your subject matter isn't so dependent upon specific frame in time captures (like a wedding or something)...

I also tried "smile shutter" and it actually worked fantastically. It was kind fun to walk around with smile shutter on, AF-C, and direct the camera towards people I wanted to capture. Invariably they'd laugh at some point, and the camera always got the shot. Pretty wild and fun to play with. I used this for maybe 10-15% of my shots.

FWIW, I shot about 400 photos, a few minutes of video, and used the LCD almost exclusively for framing and review... I still had 46% left on the battery, which was quite shocking to me. I had an extra battery along frm my NEX in case I needed it, but obviously I didn't. Fantastic performance, and I may actually get a full Disney day out of a battery... We shall see.

I'll post some photos from that party later. Now, I prepare for Disney World.

Sony SLT-A55 First Shots Review

2011-05-20T12:59:00.000-07:00

With an impending summer full of family and vacation fun, I decided to add a "vacation lens" to our library of camera gear. When I bought my Nikon D200, I got the kit with the Nikkor 18-200VR (series I) lens, and my wife Jennifer later used that lens exclusively on her Nikon D40. After she appropriated it, I missed the lens for its useful Vibration Reduction feature, and overall useful zoom range (about 27-300mm on a APS-C DSLR). We particularly enjoyed using this lens on our trip to Europe; I used the 17-55DX on my D200, and Jennifer the 18-200 on the D40. Upon arriving home, I was continually impressed with the shots from the 18-200... For much of my purposes, it is "good enough" in relation to image quality, and the zoom range sure was useful...

I heavily considered the Sony E-Mount 18-200 for my Sony NEX5. But this lens/body combo looked almost ridiculous, with the NEX5 looking like a pack of cards with the 18-200 attached. Additionally, I have found the small NEX5's grip to not be particularly useful in a way a normal DSLR's grip is; I often shoot the NEX5 looking down at it, like an old TLR camera; with my thumb on top of the shutter button. I love this method... But it wouldn't work well with a heavy 18-200 attached, where the photographer would more likely want it at eye level to stabilize the long lens.

In any case, for a few hundred dollars more than the 18-200 for the Sony NEX series, you can get a Sony A55 with 18-250mm lens kit from B&H. I figured this would give a better platform for usage of a relatively heavy vacation lens, as well as open up the ability for some more prime lenses, which are my favorite by far... I really love my NEX5, but I am extremely disappointed in the lack of lenses available for the E Mount! I had hoped by now that Sony would at least come out with a normal prime and a portrait prime...

I received my A55 this week, and have only had a few short opportunities to shoot with it. My first impressions are the body is of good quality, but does not show the fine detailing and quality feel of the NEX5, or the excellent ergonomics and robust build as my departed Canon 5D Mark II and Nikon D200 DSLR's. Not that I expected it to, as with lens, it's less than 1/3 the cost of a 5D2...

But where the rubber meets the road is performance. What do I think so far? I'm guardedly optimistic. Unlike the Sony HX9V I blogged about yesterday, this one isn't a clear and over-the-fence winner. I've found indoor white balance to be problematic, and focus accuracy and speed isn't quite what I had hoped; sometimes it's lightning fast, and other times it's just OK. This may very well be related to the rather dim travel zoom, and I will flesh that out. An example of poor AWB and not quite nailing the focus (granted, this is a tough AWB challenge, and it was fairly dim in the room... And it's ISO1600, so some focus error may simply be strong NR!)

Image quality-wise, I have been somewhat pleased, although not over-the-moon. Images straight out of camera are quite soft; some playing around in a photo editor shows that the lens is resolving WAY more detail than appears, as they take sharpening very well. On a whole, I'm not one for biting sharpness, which I find unnatural. As mentioned, the images right out of the A55 seem poorly focused sometimes, because of low sharpening and high noise reduction. Since NR isn't really tweakable unless you shoot RAW and post-process, I will play with sharpening in camera to see what I can get.

Look and Feel is important to me, and this is one area I'm curious about with the A55. Generally, the images feel somewhat "vintage" to me--more like an old film camera than a "technically excellent" new DSLR. I like this aspect!! I am hoping this bears out, but here's an example:

That just "feels" a little vintage, like film to me. I like it.

This one is also a little vintage, both due to color, look & feel, and the swirly bokeh I've seen from some old Nikkors:

Bokeh on the 18-250 can be nice and smooth, or kinda swirly and busy. I'll flesh out over time the strengths and weaknesses.

Overall, I'm just getting used to the UI, handling, and IQ. I'll know a lot more after I shoot with it for a week in Disney World, and for my grandfather's 100th birthday party tomorrow...

Sony HX9V Part I First Shots Review

2011-05-19T13:00:00.000-07:00

As mentioned yesterday, I have had a really great time shooting with the Sony HX5V over the past 9 or 10 months. Even though the image quality was a little lacking when pixel peeping, as compared to the DSLR's I was used to, I was in general quite happy with the performance when actually printed at "normal" print sizes. And the various intelligent image-stacking modes of the HX5V gave the ability to take photos that normal compacts couldn't do well.

However, I did have some quibbles. Chiefly among them:
1. images really needed some post processing (contrast and color boost, especially)
2. for sports, I would have liked some extra long zoom... and better high-iso performance to freeze motion
3. focus speed and accuracy was OK, but often didn't accurately focus on kids & pets which tend to move unpredictably
4. It was hard to achieve pleasingly defocused backgrounds when shooting portraits. Of course this is due to the small sensor. Bokeh when fully zoomed was rather busy, and of course compression at 250mm wasn't ideal for portraits
5. White Balance was often fooled indoors, and needed frequent tweaking post-capture

After seeing images of the HX9V on the web, I decided to give it a try. It seemed to attempt to solve all my quibbles, plus added better build quality and some thoughful UI revisions.

After about a week of use, I'm very excited about this camera. Sony seemed to be "listening" to my points of pain, and their solutions have proved mostly beneficial.

Focus in good light is lightning fast--I am not speaking in hyperbole here, but it feels very close to a good DSLR focus system, and accuracy is even better than most. Face detection recognizes faces at odd angles, and nails the eyes in focus (my preference most of the time) almost every time.

Accuracy of tracking moving subjects is fantastic--I get at least 8 out of 10 "keepers" during 10fps bursts, even indoors in tricky lighting:

(I was also quite pleased with white balance in the above shot--that is a TOUGH place for a camera to nail AWB).

Focus is also spooky accurate with quickly moving dogs--even good DSLR's struggle with this (my Canon 5D Mark II was very poor at tracking this sort of action, my old Nikon D200 quite a bit better):

Background Defocus control is a bit tweaky, in that you have to play with focal lengths vs. subject distance and background distance to get the best results (or any results at all, sometimes the camera will tell you "could not perform defocus" and luckily just give you the straight shot)... but the effect is quite pleasing, and can very well simulate a relatively large aperture lens on a DSLR:

Here's an example of background defocus gone awry.. Usually the camera just won't give you a defocused "try" if it didn't get it right, but occasionally you get some really weird effects:

The camera isn't perfect... And it's not a DSLR. Certainly, low light brings out the worst in compacts, and the output in my testing (none online yet, I'll get some images up eventually) shows the camera to be a little better than the HX5V but not fantastically so. And the overactive noise reduction, even at base ISO, still exists as carryover from the HX5V; grass, for instance, doesn't really look like grass but greeb blobs. As you'd guess from the almost ridiculous 16-megapixels packed onto a small sensor, Sony is doing this out of necessity. But sharpening seems improved, and it's more intelligent.

Generally, though, the camera maintains its fun-to-shoot personality, and I'm THRILLED with the contrast & color I'm getting... No post processing needed at all! That's what a compact should be--less work for busy parents :)

Also, I'm quite pleased with the build; excepting the rattly and plasticky rear control wheel which I find very hard to use, most of the UI improvements are indeed an improvement. I was initially dismayed that I lost access to HHT, HDR, and AMB on the quick control dial, but the new "MR" (Memory Remember) setting stores 3 profiles, and I have assigned all 3 of those to the 3 memory settings for easy access. That said, Auto+ actually works quite well... I may end up using that setting more frequently if it usually nails the mode I would have chosen eventually. My only reservation is it seems very quick to choose HDR mode outside... And HDR's often need some post processing to look good. That could be annoying if you're going through 500+ shots from a family trip. Yes, I'm allergic to lots of post processing nowadays, and prefer to shoot like I'm shooting with film--get it right in camera the first time, if at all possible. I am a reformed RAW-only shooter.

More to come...

Sony HX5V Review

2011-05-18T05:44:00.000-07:00

I've been remiss in posting for quite some time, so it's high time I posted a few updates. The first of which is a quick review of a camera I've really enjoyed using over the past 9-10 months; the Sony DSC-HX5V.

Since the arrival of my daughter, I have had predictably less time to fool around with cameras... Especially scanning film. At the time of our trip to China, I had 2 Canons: Canon 5D Mark II DSLR, and my old trusty Canon G9. After our trip to Disney World last summer, where I carried only my G9, I decided that small & light was way more important to me than near-perfect image quality... And sold the 5D Mark II.

Our G9 was getting long in the tooth, and I longed for a wider angle, better high-ISO performance, and (especially) faster focusing. At base ISO, outdoors, the G9's image quality didn't disappoint (provided you utilized the smallish dynamic range in the best way possible)... But I really did have a hard time keeping up with fast moving kids & dogs. And the ~35mm "wide" end of the G9 felt so constrictive. Because I did enjoy the longer zoom of the G9, especially for sports, I looked at all the "compact travel zoom" cameras available last summer, and chose the Sony HX5V.

Although the Sony wasn't as good at base ISO as the G9 when pixel-peeping, I found it had some serious strengths; chiefly among them better metering, better dynamic range (or tone curves for contrasty situations), and after I printed some 13x19's at Costco taken with both cameras on the same day.. The Sony's images were preferred by all I showed the prints, including myself. Yes, upon very close inspection the HX5V had some serious "smoothing" going on for noise reduction, which is noticeable even in daylight. But on a 13x19, at normal print viewing distances, you couldn't see that. And exposure, color, and dynamic range of the HX5V far surpassed the G9... Which WAS noticeable at standard print viewing distances. Since I rarely print at anything over 13x19... I decided not to fret that the HX5V didn't look great when pixel peeping, and didn't look back.

The HX5V also had some real tricks up its sleeve due to the superfast readout times of the CMOS "Exmor R" sensor... It stacked images in camera to reduce noise (Anti Motion Blur and Handheld Twilight), increase effective dynamic range (HDR), and had a speedy 10 frames per second burst mode for action photography. Add to that a super useful zoom range of 25-250mm, generally speedy performance, and some surprising results from the image-stacking modes, and I had a really fun time shooting with the HX5V!

This is the one thing that sticks out in my head about the HX5V: as long as you don't pixel peep, it was a FUN camera to shoot with. That's important.

Here are a few images I really liked out of the HX5V (hard to choose, I liked a lot of my shots)... Some HDR shots:

And some "normal" shots:

The Handheld Twilight mode proved to be quite useful at times:

Anti-Motion Blur was fairly successful as well, and while the quality goes way down as compared to HHT, you get shots that you otherwise wouldn't get without horrendous amounts of noise. The motion-stoppping algorithms were pretty good, too, and usually chose the proper things to "freeze" (usually people):

With 10fps and a fairly long zoom, I was excited to have the HX5V in pocket for sporting events. However, in practice, although you could get some decent shots:

...In practice, the camera was disappointing for "sports" shooting. It often lost focus when tracking the subject, focusing on the background instead:

...and its performance with very fast moving objects (like cars) at long zooms left a lot to be desired:

The HX5V wasn't awful in those circumstances, and I certainly got some nice shots, but I tried printing some for my wall and they looked rather poor at 11x14, 8x10 was about as high as you could go. Now I admit: this was pushing this little camera so far...

I had some other niggles with the HX5V after quite a bit of usage. First and foremost: contrast and color almost ALWAYS needed post processing to make the images shine. Frequently, white balance and/or sharpening also needed attention. Back when I had a ton of free time, I really enjoyed pouring over each image and optimizing it... I admit, I'm lazy nowadays. I don't want to post process much, if at all. Because I always felt the need to postprocess the HX5V's images, this was a bit of a drag... Especially because the Sony NEX5 I also use rarely *needed* post-processing at all (although if I wanted to PP, I could always get a little better image, the delta wasn't as great as with the HX5V). Physically, there were a few things I didn't love about the HX5V, also: the lack of a grip coupled with the smooth (and cheap) plastics made it less comfortable to shoot for extended periods. The screen was acceptable, but nothing else... It looked about 3-5 years behind the times to me, the coarseness and lack of color fidelity made it difficult to review images in camera.

I recently bought the Sony HX9V, which I will be reviewing shortly... and I can say Sony listened to all us users on the niggles of the HX5V, and at least attempted to address them.

That said... I absolutely LOVED my HX5V. Although the image quality when pixelpeeping always disappointed a little... The fact that I could take it ANYWHERE, when coupled with truly excellent exposure for people photos, the fantastic image-stacking modes, and a really useful zoom range, made the HX5V my go-to camera for quite some months! As long as I put aside my prediliction to be overly critical about image quality... I was always very pleased with what I got back from that little camera. It was "good enough" most of the time, and really fun to use, always.

Nikkor-Q 135/3.5 Mini Review, Part II

2009-02-23T03:50:00.000-08:00

Datasheets for the Nikkor 135 f/3.5 Q from "The Nikkormat Handbook" by John Cooper.

One of the neighborhood turkeys, shot with the 135/3.5 Q at full aperture.

Two of the neighborhood Tom's, shot with the 135/3.5 Q at full aperture.

I hadn't shot wide open much with the 135/3.5 Q, but I had an opportunity last week when I walked outside to take out the garbage, and saw the turkeys roaming my neighbor's yard. I happened to have my Nikkor 28/2.0 on my F2, with the 135/3.5 Q in my case ready to go, and film in the camera... I thought briefly about running inside to grab the 85/1.8 H I've been testing lately, but after a quick meter reading decided that 1/125 @ f3.5 was just about perfect, anyway... So I thought I'd give the 135/3.5 Q a try.

I think the Turkey photos turned out pretty well, actually. At full aperture the 135 Q is definitely softer than at 5.6/8.0, but it's not so terribly soft as to be unuseable. There is still some of the "bite" and "motion" or "tension" noticeable that makes this lens such an interesting choice. Some of the "swishy" nature of the bokeh is tamed at full aperture, although the loss of the incredible sharpness evident at f5.6 is probably not worth the tradeoff, IMO, unless you want to soften the bite a bit (as for a portrait).

Shot at 1/250 @ f8 in the morning on the way in to work.

I've been meaning to get a few photos of the LeVeque tower in its early morning glow, and sometime I need to walk a bit closer to get some detail shots. The angle I'm at here precluded a good composition due to a parking garage and other unsightly things just out of view. But it's evident that the 135/3.5 Q can display SERIOUS sharpness when stopped down a bit. In my last review post, I stated the lens wasn't quite as bitingly sharp when shot at infinity focus... I believe, though, that some of this "softening" is due to my inability to sharpen my 30-megapixel Canon MP980 scans enough... I need to figure out a good 2-pass sharpening technique. Here, I simply took a 1.7-megapixel scan and sharpened in DxO as I normally would a 30-megapixel scan: 500/.8/0. It ended up a bit oversharpened, but obviously I could tweak this if I should so wish... I should be able to get the large scans to be similarly sharp, with some practice. Even in this small scan, though, you can see the incredible detail that this lens can draw at a distance. No, it's not as bitingly sharp as a close-range shot... But it certainly isn't terrible.

This weekend I went to a Columbus Blue Jackets hockey game; I briefly considered taking the 135/3.5 Q, but I am actually glad I did not. I shot my Minolta MC 135/2.8, and needed every bit of that wider aperture. My photos were well-exposed at 1/250 and f2.8 (ISO 400), and some were useable even though the shutter speed was a tad slow and motion blur crept in. The 135/3.5 Q would have struggled with a 1/125 shutter speed in this situation. So that's the big negative of this Nikkor; it's not quite fast enough for indoor use.

Really, it's an outdoor lens. And a very unique and interesting one, at that. I hope I have been able to give a bit of this unique "flavor" that this lens imparts to photos. It's well worth buying and enjoying.

Nikkor 135/3.5 Q 135mm Lens Mini-Review

2009-02-17T07:39:00.000-08:00

First in a series of old Nikkor lens reviews, I present to you the Nikkor-Q 135/3.5 (above, on the Nikkormat which is the left camera).

I bought this lens for $50 or $55 from a local camera shop, Chick's Camera Exchange. It was in pristine shape, and came with the original metal clip-on lens hood (not shown). This lens isn't known as one of the best Nikkor's of the era, although most seem to agree that it is pretty good; here, Bjorn Rorslett:

The venerable old-timer from the Nikon 'F' epoch can still be used on modern Nikon bodies (that's lens continuity), and is capable of giving quite sharp images from f/5.6 onwards. There is a roundness of detail rendition that suits many scenes quite well. Image contrast isn't extremely high by modern standards and the lens flares quite easily since it is single-coated.

Since it isn't known as one of the best Nikkor's ever, prices are very very low. Add to this that the photographer's kit from the 60s and 70s often included a 35, 50, and 135; so many of these lenses were sold... And you can buy them quite cheaply, as I have demonstrated. 135mm is a pretty long focal length, so many of these 60s SLR newbies didn't use the lenses much... So you can often find them in really good shape.

So, should you buy one? Yes! I like this lens a lot. Its manner of drawing is quite a bit different from most modern lenses. The very simple 4-element in 3 groups design, derived from the Nikkor rangefinder 13.5cm/3.5 lens designed in the 40s or 50s, so it is quite an old design. My model is a pre-ai lens with the original coating and not the later multicoating. The photo above demonstrates the strengths and weaknesses (or uniqueness) of this lens quite well: sharpness is outstanding and bokeh is interesting.

The sharpness of this lens is absolutely astounding.

From f5.6 on, sharpness is on par with ANY modern lens. High-contrast or busy backgrounds can be just that. The manner in which this lens draws reminds me of a rapier: quick, sharp, and a little busy. There is sometimes a "motion" or some "activity" in the bokeh that might offend modern sensibilities, but on balance I actually like on occasion. Further, this isn't always the case:

The 135/3.5 Q isn't the highest contrast Nikkor; in fact, it is generally much lower in contrast than most other Nikkors of the period. Either use this to your advantage, or switch to another lens. It "tones down" high contrast situations well:

But make sure you use the lens hood! This lens is extremely susceptible to flare and loss of image contrast, especially when stopped down a bit (f8 or more, especially). Since the lens starts with low contrast anyway, keep this in mind as it is quite noticeable:

Perhaps the later 135/3.5 QC version is a little better in that regard. If you like to do distant cityscapes, etc. that the compressed 135mm focl length can give you, I'd recommend looking for the QC, AI, or AIS versions for that reason. I may have to try a late version to see if they improve contrast a little yet retain the very special "rapier" manner of drawing that I so enjoy.

You can see in the photo above that sharpness isn't as apparent at infinity. Although atmospheric haze can certainly affect this, my general feeling after using this lens for the last 6 weeks is that this lens is optimized for the <20 foot range, where its exceptional sharpness really comes through.

Handling and shooting with this lens is a little difficult, so I thought it worth mentioning. Focus throw is rather long, and the ring on mine fairly stiff. Add to that the somewhat dimmer-than-normal view through the viewfinder due to the max f3.5 aperture, and you can be challenged a bit when photographing moving objects.

Would this lens be a good portrait lens? I'm not sure; the balance towards super-sharpness and the slightly "busy" bokeh may provide an edge that really isn't so welcome in a portrait. Perhaps, though, if you choose your backgrounds carefully, and shoot wide open or at f4, the portraits could turn out decent. You definitely do NOT want to shoot at f5.6 or higher; sharpness is such that it would render facial imperfections in a not-so-pleasing manner. The 135 focal length is too long for indoor portraits, so they would likely be outdoors and backgrounds would have to be chosen carefully. I personally like shorter focal lengths for portraits (85-105mm), so this wouldn't be my first choice. But I will try some and report back someday.

In short, this lens is a smashing bargain at the price it sells for today. I really like the manner in which it draws. It's my "rapier" and I enjoy it thoroughly.

Best for: pets, hiking (a great closer-range nature lens), travel in cities (shoot short landscapes, inifinity landscapes not a strong point), perhaps short-length animal photography, where exceptional sharpness can bring great pop to the images.

Lessons Learned from eBay: Buying Old Camera Gear

2009-02-10T05:57:00.000-08:00

Tree and shadow outside my home after a large snowfall. Taken with Nikkor-S 50/1.4, Nikon FE, and Fuji Superia X-Tra 400. Converted to B&W and toned in DxO.

OK, so since I've been in old-camera-acquistion mode lately, I thought I'd share a few tips if you're looking for old camera gear on eBay:

1. If it says it's "mint/near mint" it likely isn't. Look for high-res pics and a return policy. If it's mint, the seller won't balk at either.
2. If it says "no battery to test" ANYWHERE in the auction, THE CAMERA IS BROKEN. No matter what the person says ("it worked last time I tried it" "my uncle said it worked" etc)
3. If the seller says "I don't know a thing about cameras nor how to test" IT IS BROKEN.
4. If the seller says ANYWHERE that the item "works" but is sold "AS-IS" IT IS BROKEN.
5. If the seller says "clean" this does not mean it works. It could be broken and "clean."
6. If the seller says "clean glass" this does not mean: no cleaning marks, no element separation, no fungus, no dust inside, no oil on aperture baldes... UNLESS SPECIFICALLY STATED. If the seller does not specifically state these things, ASSUME THEY EXIST and bid accordingly.

Some stinkers I've received:
1. "clean" camera which had mold inside, and frozen selector switches (water damage?)
2. "near mint" lens with tons of visible fungus inside. Seller courteously refunded my money after returning the lens (I was out shipping) and promptly re-listed as NEAR MINT again! Stay away from "jfkamera" on eBay. This is the actual photo in the listing... Note no frontal view of the lens which would show the fungus:

3. a Minolta XD11 camera which "sounds as if it is snapping a picture when I push the button" which arrived in completely unoperable state. If an XD11 will flip up the mirror and fire the shutter, it's working. The XD11, with it's electro-mechanical shutter either works, or it doesn't. This XD11 arrived to me with a completely frozen frame advance lever, and an inoperable shutter release. The camera would not respond at all, to any input. The seller offered a refund on the purchase price only, not the shipping... I don't agree that this is equitable, since the item was not delivered in the state advertised. I'm still trying to convince this seller of my position, and may have to file a claim with my cc or ebay. So... Lesson learned here: if the listing doesn't specifically state "meter is working, shutter is working, mechanically operational" or the user claims "I don't know enough about cameras and lenses to guarantee that everything is in perfect working condition" then ASSUME THE CAMERA IS BROKEN.
4. A lens advertised as "Optics clear and scratch free" exhibiting a fairly sizable dust blob (dark in color) and the beginning of element separation on a cemented pair of elements. Also, the lens rattles in the hand when slowly turned over, and the aperture ring feels odd when turned: an issue with the spring? Don't know, I haven't paid someone to open this up yet. This seller was very honest, didn't know the flaws, and offered a part refund. So things can turn out OK, although if you want near-mint stuff you will get some disappointments.
5. A lens advertised as "collector quality" with "clean glass, no marks" which had an absolute CROSSHATCH of cleaning marks on the front element.

I will put together some buying guides soon on some of these old products. There are a lot of nice cameras and lenses out there on eBay; if you are OK with pre-ai Nikon lenses, or anything Minolta, you can get some really excellent deals on what I would call "near mint" stuff. Lenses with very slight handling marks and perhaps a few tiny pieces of inconsequential dust sell for as low as $6 for a Minolta Celtic, or $40 for a Nikkor. Cameras with very little use and wear can be found for as low as $20-45 (Minolta SRT's and Nikkormat's especially). Sometimes, you can get someone's entire "kit" with wide, normal, tele lenses (usually 35/50/135) and a mechanical camera for $50-75. There's some great deals out there in ebay photo land. But also some real stinkers. Be prepared, do your homework, don't bid impetuously.

Costco Film Scan Auto Corrections

2009-02-05T06:38:00.000-08:00

Yesterday I mentioned Costco's auto-corrections on film scans, and some of the interesting things they do to a scan. I uploaded a few images to Flickr which REALLY show this in action.

Costco scan:

Canon MP980 scan:

This image makes the following quite apparent:
1. vignetting correction! The Nikkor-S 35mm 2.8 lens was shot wide open here, and shows fairly heavy vignetting. Costco's scanner nearly eliminates that effect.
2. Color correction: Costco's scans come out a little too yellow for my tastes. My Canon MP980 scans too red. I'd like to find something a little more in the middle.
3. Noise reduction: look at the yellow and white chair just above Henry's head; the Costco scan shows a lot less noise. This has pros and cons in some shots, but overall, if you're allergic to grain... You'll like the Costco scans. I don't mind it.
4. Active exposure correction, like Nikon D-Lighting, DxO Lighting, etc. Costco's scan definitely shows selective brightening. Fairly effective, at that. I used DxO Lighting on the Canon shot, the original scan would have shown this better.
5. Sharpening. Try as I might, I can't match the fine sharpening that Costco's Noritsu seems to be able to do. I don't own Photoshop, though, so perhaps Photoshop's USM could do better... But DxO, ArcSoft Photo Studio, and Capture NX all cannot equal Costco's sharpness. It's epic. If you like biting sharpness... You'll love the Costco scans. (although a caveat: the 6.5mp scans are sharper than the 26 megapixel scans... but remember the print sizes you're looking at!)

The Costco scans are really good for the amount of time you spend scanning them (which is to say, no time!)... Although I'm not 100% happy with them. The scans from the Canon MP980 have a more 3-dimensional, less manipulated look; even at the 8x12 print sizes I'm looking at. I don't like the red cast from the Canon, though I have taken steps to minimize it (in the original scan of Henry, he's a deep brick red... not realistic!), I have more work to do. The big problem is one of speed... The Canon takes 7-8 minutes to scan one negative at full res. Multiply that by a roll of 36... And the Costco scans look very attractive :)

These photos taken on Fuji Superia X-Tra 400, with a Nikkor-S 35mm 2.8 lens, wide open, through a Nikon FE.

26 Megapixel Film Camera? At Costco, YES

2009-02-03T06:39:00.000-08:00

As I mentioned in a previous blog, Costco scans your film for a super-cheap $2.99 per roll at time of development. I have had mixed results with this, and thought it a good time to codify what I've found in case others are interested.

1. Cost of film at Costco has dropped to $1 per roll. It was $6.97 for 6, and now is $5.97 for 6. This film is Fuji Superia X-Tra 400 (24 exposures), which I have found to be quite good.

2. Cost of development is $1.59 fixed (for anything up to 40 exposures).

3. Cost of scanning is $2.99 fixed (for anything up to 40 exposures).

4. If you get prints with your scan, the size of the scans will be anywhere from 1.7 (with 4x6 prints) to 6.5 megapixels (with 5x7 prints). (I will discuss how to tweak this later).

5. If you don't get prints with your scan, and don't help the Costco employee "fool" the Noritsu for high-res scans, you will get 1.7 megapixel scans.

6. The Noritsu at my Costco will spit out 26 MEGAPIXEL scans if you really "fool" it. This seems to be the native scan resolution.

How to "fool" the Noritsu: when I realized that scans were higher-res with 5x7's, I wondered if this was as high as the machine would scan and drop to CD. I also didn't want prints with my scans, as I usually just toss the folder in a shoebox and aside from the few 4x6/5x7 that end up on the 'fridge, they go to waste. So I asked my friendly Costco employee (Mickey) to see if he could "fool" the scanner into dropping the 6.5 megapixel scans by telling it I was printing 5x7's, but not print those 5x7's. This worked, wonderfully! The 6.5mp scans are wonderful. To do this, ask your friendly Costco employee to "scan at 5x7 to CD, no prints."

Around this time, I had a 8x12 printed out for my Grandfather. I scanned this on my Canon MP980 flatbed, and noticed there was a WHOLE LOT MORE than 6.5mp of resolution. So I started thinking... The largest my Costco could print was 12x18... What if Mickey set the scanner to "scan at 12x18 to CD, with no prints" ? What resulted was ASTONISHING:

26 megapixel scans.

So there you go. For a paltry $5.58 per roll ($1 for film, $1.59 for dev, $2.99 for scans) you get 26 megapixel film scans that are really REALLY good. The positives:
1. You waste no time brushing off film, scanning, etc etc at home. Hours upon hours of work for each roll are saved.
2. The scans show acuity that I've never seen from a consumer flatbed scanner. Even the expensive film scanners don't seem to extract this much detail from scans (click through and open the large scan)... The detail is quite literally astonishing:

3. There is a VERY sophisticated set of "Digital ICE" occurring. I have compared with film scans at home (ad nauseum, for hours!) and I notice the following in all Costco scans: very sophisticated USM which picks out fine detail and not much grain, sophisticated noise reduction which virtually eliminates grain in some areas of images, as-expected exposure adjustment which will make anything -2 to +2 stops seem "average," and even vignetting correction for old lenses (!!)... In the image below, you can see a slight bit of vignetting, but this was taken wide open with the lovely Nikkor 35/2.8, and I can assure you vignetting was actually much heavier than shown in this shot:

That said, there are some negatives:
1. Color is just a tad off from reality; it seems slightly "yellow" to me. The comparison between the Olympus E-420 and the Nikkormat/50 1.4/Fuji Superia X-Tra 400 I posted earlier is a good example:

I "warmed" the digital shot quite a bit, and still it doesn't quite match the honeyed glow of the Coscto scan. Sometimes, this honeyed glow works well. Other times, not so much. You can, of course, tweak balance in post, so this isn't a problem you can't overcome, but in some shots it can become a little onerous.
2. Exposure adjustments can be highly ANNOYING. For anything that is far outside the realm of "average" you will be disappointed. For instance, black dog shots. The auto-adjustment at Costco will try to make your black dog "properly exposed" and when you've put him at Zone II/III he'll be a Zone V! And your background will blow. Like this:

And this is how I exposed the frame (Canon MP980 scan, sorry for not cleaning the neg first!):

This isn't only a problem when shooting a black dog, for instance. Snow photos have come out VERY poor from Costco... At least compared to my artistic vision. Essentially: if you're a Zone photographer, or you craft images to drop into specific exposure areas, you WILL be disappointed if your efforts are somewhat away from Zone V.

I am going to work with Costco on this last point. I think it's likely that auto-adjust can be shut off...

I will also be posting a series about scanning with the Canon MP980 soon. There are some things that really need to be followed to get the best out of this scanner. When I do these things, scans are very good, indeed. It is a fair amount of work, though; so if you really just want the easy way out, use Costco!

First High-ISO Test of DxO/Canon 5D Mark II hits the web

2009-01-27T12:01:00.000-08:00

Last week I mentioned I had downloaded the DxO Optics Pro update that allowed me to natively "develop" my Canon 5D Mark II RAW files. I wasn't too thrilled with the high-ISO results in my initial look, and I haven't had time enough to really figure out what's going on with DxO and the 5D II. But as expected, results are hitting the web, including Focus Numerique (translated from French to English with Babelfish).

They post a photo comparison of DxO vs Canon Digital Photo Professional:

On the left, DxO. On the right, DPP. OK, so you can see that DxO is really cleaning up these high ISO images. But what Focus Numerique doesn't show in this "micro test" is what I have found problematic: although noise is certainly alleviated, so is detail. Perhaps I'm being too harsh? Yes, at 100% detail is definitely reduced by DxO Noise. But it does look good, as a whole...

The other thing the test image won't show is banding: the 5D II shows serious banding in the shadows of ISO 12,800+ shots. This can be mitigated somewhat in post-processing, but if you're shooting in very yellow tungsten light, you may not like the results (blue banding is prevalent). DxO does nothing to eliminate this, that I see.

Anyway, I will print some images and make up my own mind. Look for more here in the coming week.

DxO Optics Pro Adds Canon 5D Mark II Support

2009-01-23T06:11:00.000-08:00

Yesterday I received an email from DxO that support for the 5D Mark II (and Canon G10) has been added. Bravo! And only about 6 weeks after they hit the street in the US. That's much faster than previously, so their "doubled capacity" may prove to be true.

I downloaded the update last night, and played with a few photos. I thought I'd post a few on my Flickr stream to share, but it's clear I will need to learn how to tweak them to my liking. What I am seeing is not that positive at the moment.

Noise: I ran an ISO 12,800 image through DxO's correction modules. I opened it expecting to see the very fine-grained noise/detail that I have so enjoyed with the correction module for the Nikon D200. Instead, what I got was plastic. No detail, all noise obliterated. Playing the the sliders didn't really help that much. This is very much up for consideration.

Color: the color seems quite different than Canon's own DPP app. I will need to do some side by side testing, but there are positives and negatives here. Generally, the color response is very much like the Sony A900 to my eyes.

Tone: the default tone curve in DxO 5.3.2 is VERY steep. Steeper than Canon's! Yikes.

White Balance: click white balance with a subtle adjustment enabled me to absolutely NAIL my dog Henry's color in compact flourescent light. I couldn't do this with DPP. And the color is more saturated than DPP.

Color may be better with DxO... Noise and Tone are not.

More testing is in order, and I will share photos as I have some good examples.

Nikon Film Bargains in the Digital Age: the Golden Age of Film?

2009-01-22T10:55:00.000-08:00

The world of Nikon F mount cameras is large and varied. I have been doing a lot of reading at Michael Liu's wonderful set of pages which detail much of the intricacies of the F system. I have a lot to learn about what's what, what's good, and what to stay away from; but I can say that I feel we're in a sort of golden age in the history of film photography. Film, film cameras, and wonderful old lenses are fantastic bargains if you know where to look and are prepared to be a patient treasure hunter. In the photo above, you see the following (in paren's what I paid for the items), from left to right:
Nikon Nikkormat FTn chrome camera body ($29)
Nikkor-Q 135mm f3.5 on the Nikkormat ($55 with shade, absolutely mint)
Nikkor 35-70 f3.5 Macro behind the Nikkormat ($65, absolutely mint)
Nikon FE body ($46)
Nikkor-S 50/1.4 ($free with Nikkormat)

That's a grand total of $195 for 2 bodies and enough lenses to keep you in photos for a good long time. What the cost (or the photo) doesn't do a good job of conveying is the quality feel that is so lacking in today's plastic-clad cameras. Sure, cameras like the Nikon D200, Canon 5D Mark II, and the like are certainly well built and solid. And they do feel very good in the hand. But there's something about these old chunks of metal! These old bodies and lenses are SO well constructed, and a joy to use.

What's more, I forgot how truly easy it is to manually focus a camera lens on a "full frame" SLR. I haven't done this much with my 5D Mark II, since the autofocus is so darn handy, but it really is a non issue with these old Nikons, their excellent split-prism finders, and the well-damped and smooth MF rings on the old Nikkors.

If you don't mind manual focus and manual exposure, you can find some stellar bargains as I show above. Sure, some of the old Nikon gear is expensive; collectors snap up really great pro bodies and superfast lenses. And many people still shoot and buy Nikon's later MF gear, making the late Ai and Ai-S lenses quite expensive as well. And for some inexplicable reason everyone wants the black cameras... Even though they don't wear as well as the Chrome. I haven't figured that out yet. I mean, sure, I like the look of black... But if chrome is 1/2 the cost, and easier to find in better shape (maybe that's the key), I'm all for that...

But you can find some stellar bargains if you shop pre-Ai lenses and the older bodies. The FE shows above is quite modern; it has an aperture-priority mode, and a very good centerweighted meter. To use a pre-Ai lens on it you must use "stopdown" exposure measurement. No big deal, really... But if you want the ultimate in retro film mechanical durability... Go for an old Nikkormat and some pre-Ai lenses. I don't know if you'll find one as cheaply as I did ($29 Nikkormats seem hard to come by nowadays, even though they were a dime a dozen just a few years ago), but you certainly won't spend a lot.

So why is this the golden age? When a photographer can spend $200 and get a heck of an assortment of fun gear which will last a lifetime if cared for... And film costs $1 a roll at Costco (Fuji Superia X-Tra 400, 6 rolls, $5.97)... And development costs $1.59 at Costco...

Sure feels like a golden age to me. Enjoy it while it lasts.

Shooting Film in 2009: Nikon Nikkormat FTn vs. Olympus E-420 and The Pancake

2009-01-16T16:40:00.000-08:00

Me shooting the Nikkormat with a Nikkor-S 50/1.4 in extremely cold weather, as seen by the Olympus E-420.

I did a little digging tonight, and found I bought the Nikkormat in December of 2000 from Midwest Photo Exchange in Columbus, OH. I thought I recalled purchasing it for about 50 dollars, but I was shocked when I found the entry in my check register... $31.73. $32 for the Nikkormat WITH a lens! I remember thinking that was cheap enough to buy "on a whim." "I'll say!" as my grandfather often says.

Over the years, I used the Nikkormat infrequently, usually keeping it stoked with Tri-X, Plus-X, or Ilford Delta 100. I never put color film through the Nikon; I had "digital" for color. And film B&W was so much better.

Fast-forward to last week when I discovered Costco developed film. For $1.59! Crazy. To shoot a roll of film and have it developed and scanned at Costco it will set you back the princely sum of 6 bucks. That includes the film. I thought, "I'm crazy if I don't take advantage of this, and shoot some film before it disappears altogether." I also thought it might be a nice reality check to test my theory that digital pretty much equals 35mm film in all respects now. I'm always up for a good photographic test.

Costco only sells one film: Fuji Superia X-Tra 400, in a 6-pack for $6.97. I picked up a box, loaded the 'ol Nikkormat, and went shooting...

A view North on High St. in Columbus, OH during a snowstorm.

Decrepid railroad bridge across Rt.33 entering downtown Columbus.

One of the first homes in Northwest Columbus, off of Rt. 33 and Fishinger Road near Upper Arlington.

La Pavoni Europiccola Millenium. My morning coffee machine. Pulling a double of Espresso Brioso from Cafe Brioso in Columbus, OH. Tasty!

Doreen hard at work, with natural light streaming in her 3rd floor window.

I had a great time shooting film again. In this day and age of Digital SLR's, I thought perhaps I might have trouble shooting an all-manual film camera (manual exposure, manual focus, manual everything). But the Nikkormat was like putting on an old tennis shoe... It just fits.

I realized something: I shoot digital rather like I shot my old manual film cameras. I really don't "chimp" much with digital, and honestly, I don't even look at the info in the viewfinder... I shoot just like I did in the film days; take some test exposures, set exposure compensation if necessary (or do so on the fly, with manual cameras like the nikkormat!), and just go shoot. I'm much more into the act of capturing images than the actual craft of the capture... If that makes sense. My hands do the work, and I simply follow my eye's vision. As long as a camera "gets out of the way" and lets me work, I'm fine with it.

The Nikkormat "got out of the way" just fine! I thought it might be tough to get used to manual focus again, but it wasn't, really. The split-prism rangefinder viewscreen on my Nikkormat absolutely rocks, and MF really wasn't an issue. I am certain I don't always nail MF as well as I could with an AF camera, but does it really matter at the sizes I normally print? If I'm planning on printing at 20x30 I'll make sure it's nailed. Otherwise, if its shooting for fun, you have a little leeway anyway.

The results? Pretty good. The Nikkor-S 50/1.4 has always been a wonderful lens for me. Sharp as heck from f2.8 on up, with great color and contrast. f1.4 is a little dreamy, f2 sharper and great for portraits. Just a really great lens, and proved to me that it can do "color" as well as I remember it doing black and white.

Costco did a very good job developing, and a pretty good job scanning. I say pretty good because I did find some blemishes on the scans (dust). Not a ton, but a little. Not a huge deal to me because I can always re-scan if I want to print large. The scans are 3130x2075, and appear to be good to print at least a 8x12. I will have to try some 13x19 reprints to see how they hold up; I think they should be fine at that size. If you want to enlarge any further, you'll have to invest in a film scanner.

The Fuji Superia XTra 400 isn't bad film at all. In the following 100% crop, you can see a "shootout" between my Olympus E-420 and the Superia as shot by the Nikkormat. I used the E-420 because 1) it was pocketable, 2) I felt the 10mp E-420 would be most similar to a 6.5mp film scan, and 3) wondered how "the Pancake" would do against a famous vintage Nikkor. I've always really enjoyed the E-420/Pancake combo, feeling it was rather like the digital version of my old film "friends"... So I decided to see if it was up for the challenge. I actually carried and shot both the Nikkormat and the E-420 for a few days; I will probably put up some more comparos in the coming days. I really don't enjoy shooting 2 cameras, though, it simply takes away from the vision...

Immediately it is obvious the digital is holding more detail. Not even close here. The Fuji 400 speed film is at a big disadvantage to the E-420; the E-420 can, of course, be shot at whatever ISO you want, for each frame. Not so with film, of course. At ISO 100, the E-420 is essentially noise-free, and this really shows in the sky crop, and the shadows on the house.

But less immediately obvious is that the film shot had MUCH better color and more attractive contrast (to me). I had to work for about 10 mins. on the RAW file from the Olympus to get it close in color/contrast to the Nikkormat/Fuji shot. When I opened the pack of film from Costco (I splurged and got 5x7s), I was taken aback by the excellent color and contrast. Very punchy, but subtle where it needed to be. A fun film. And as you can see, it actually did well under varied light sources which means you don't actually have to carry a stack of filters to make it look ok.

The above shot of the house is a good example... Have a look at the "original" size and you can see that it's quite contrasty, but holds detail very well in both the shadows and the highlights. I don't have the E-420 version up, but no amount of work I tried made the E-420 photo quite that perfect; the Olympus seemed to wash out the tones in the tree bark too quickly.

So what do I think? The Olympus has some strengths over the Nikkormat film setup, notably speed, portability, storage space (3-400 RAW images on one card), and is a wonderful compositional and learning tool. I feel a bit more "free" to experiment with digital, since I'm not thinking in the back of my mind "that shot cost 40 cents." And I've gotten great color and "look" from the Olympus with some fiddling of the RAWs and in DxO. I also think the Pancake is a great lens; it has a sort of vintage look to the photos itself, which harks back to the time of the great Nikkors. A bit more bold, perhaps, but still a more vintage manner of drawing than most modern lenses.

But film still has much to be desired: great color and contrast, and the "cost factor" as well as the full manual controls make one really think about each photo. It's nice to slow down sometimes! I have also found that one can buy old Nikkor lenses for around $50-75 apiece in absolutely wonderful shape. These old lenses are absolutely stunning in their build and design; they easily feel much better than any modern counterpart. As long as you don't go for the super-rare or super-fast models, and you don't mind full manual controls, there are HUGE bargains to be had in non-AI Nikkors. I just picked up an absolutely mint 135mm/3.5 Nikkor-Q with hood and original caps for about $55, a near-new 200mm f4 Nikkor-Q for $49, and a similarly mint and probably never used 35-70 f3.5 macro Zoom-Nikkor (OK, AI but a good deal) for about $65. I wouldn't normally buy a zoom for a manual camera, but... Why not at that price? I'm on the hunt for a 35/2 or 35/2.8 and a 105/2.5. The latter holds its value a little better, as its pretty famous...

Anyway, more to come soon on that front, including photos. 'Till next time...

Ken Rockwell Tests D3x and 5D Mark II against film!

2009-01-14T11:13:00.000-08:00

Ah, Ken Rockwell. The guy is the much-bemoaned evil Nikon-biased prognosticator twin of "Thom Hogan." Thousands of people a year on various internet fora are whipped into a frenzy at his ridiculous postings. Everything to Ken is Black and White (overpumped colors, really, but...). It "is", or it "isn't." It's "the best" or "the worst." And my favorite: what is the "best" today, is "the worst" in a short time, as soon as something new comes out.

Really, though, I think the guy's a genius. AFAIK, he makes his living from his ridiculous website. So how does one drive traffic to that website? Create controvery! If you think about it, simply take what he says with a grain of salt knowing that it's there to stir controvery (and thus, drive traffic, and donations) and there are some actual truths in his writing. In fact, he says so himself here.

The Canon boys are up in arms about his D3x vs 5D Mark II test. Really, if you simply take what Ken has said and divide by 42 "Ken Embellishment Units," you arrive at the truth. The D3x has better JPEGs. So what? Come on, Ken even states:

If you wanted to screw around with raw files, you probably could make the Canon 5D Mark II look about as good as the D3X.

So there you go. Why does it hurt so much? No one shoots JPEG, anyway.

So, with the requisite 42 KEU applied, we get:
1. The 5D Mark II needs to be shot RAW to get the best out of the sensor
2. The 5D Mark II renders color a little differently from the Sony A900 and D3x (not worse, IMO, just like a different "film")
3. 5D Mark II ($2700) has essentially the same image quality as D3x ($8000).

Interestingly, Ken updated the comparison with shots from 645 and 35mm film cameras. What does this show? It takes about 21-24 megapixels to "equal" 35mm film. This is something I've long guessed to be true, and others are starting to echo the same. We're "there" now, with digital.

So I say this as I shot my first roll of film in many years. What was I thinking? Well, I decided to have some enlargements of Christmas photos done for my grandfather at Costco, and when I was there I noticed film development chemicals... And a very cheap development price. For about $9 you can have your film developed, scanned, and printed. And it just so happens Costco sells 6 rolls of Fujufilm 400 Superia X-Tra for $6.97. Nuts! That makes a roll of film, with development, printing, and scanning, cost $10. Crazy cheap.

So I dug out the old Nikkormat FTn with the only lens I have for it, a Nikkor-S "Nippon Kogaku" 50/1.4. I have never shot color on the Nikkormat, only Black and White:

I loved that camera for black and white. Without a spotmeter, I had a harder time using the Zone System to place my tones... But the lovely contrast and pop of the images usually camer through unless I mucked it up pretty badly. The old 50mm lens (single coated) certainly exhibits more flare, ghosting, and definitely not round OOFh, but sometimes this can be used to great effect:

Really a perfect combo for high contrast B&W, which I was really enjoying during the '99-2002-ish period.

I don't know how it will do with color. I'd love to shoot some Kodachrome 25... It would fit that camera so well. I wonder if I can find any?

In any case, I will post some images soon.

Shooting Dogs in the Snow with the Canon 5D Mark II

2009-01-10T11:11:00.000-08:00

Unusually we have had a lot of rain this winter. When it finally changed to snow yesterday morning we decided to get the dogs out for some much-needed hard play. They've been a little cagey as of late.

My wife threw the ball, and I decided to try to get some shots of the dogs playing. Shooting with a new camera, especially in a difficult situation (white snow, black dog, moving fast!), is always a challenge, and this proved no less of a challenge. So what did I learn?

Custom Functions: I have C2 setup for "street" shooting; ISO 400, D+, motor drive, AI Focus. I started here for the dog shoot, figuring I'd just switch on AI Servo and be done with it. After a few test snaps, I upped the ISO to 800, and moved on. Well... One of the neat things about Custom Functions on the 5D Mark II is that when you switch your camera off, when it powers on, it resets back to all the settings you stuck in your Custom Function position. The idea here is you can use your settings as a base, modify for the shoot, and next time you fire up your camera it's back to normal. Well, that bit me in the butt: when the camera goes to sleep, your custom settings revert as if you'd turned the camera off! Oops. This put me back at ISO 400. So my shutter speeds were a little too slow to freeze motion well.

D+/Highlight Tone Priority: I'm not liking this setting ON at all. Basically, it uses a slightly higher than normal shutter speed, which protects highlights, and uses selective gain enhancement on the shadow portions of an image to bring them back to normal. Essentially, it intends to increase dynamic range of the sensor, although at a slight increase in noise in the shadows. I figured this was ideal for a snow shot, especially with a black dog, and that a little extra noise was no big deal with the buttery smooth 5D2. I'm not liking what I'm seeing in the shadows, though; there is an odd crosshatch-type pattern in flat shadows. Click through this file and view the original size:

Check just left of the "P" in pink on my wife's pantleg. Sorry for the overly-compressed example (Flickr at work here) but you should be able to see what I mean. This is an ISO 400 shot in broad daylight; while it could have some noise in the shadows, this isn't the normal random noise we're used to seeing. It has some odd patterning. I'm turning off Highlight Tone Priority/D+ from now on.

AF Servo Performance: is ok. Not quite up to the level of the Nikons, I think. I did some dog-in-flight shooting with our D200 and the keeper rate was probably 10-15% higher than with the 5D Mark II. The 5D2 was just about as good as the D200 when shooting Henry, but when shooting Tucker the black Lab it just had a harder time. Here's an example:

You can see the plane of focus just at Tucker's back legs. A so-so performance here, but generally what I expected of the 5D2. I will have to try some more shooting of moving dogs soon; I realized since I used my C2 setting I actually didn't have the "focus assist points" turned on, so this should help in tracking.

Frame Rate: plenty for my needs. I suppose 6fps would be better, but really, ~4fps is just fine for me. I didn't catch any better shots with the 5fps of the Nikon D200. I did notice the slightly longer viewfinder blackout, but again, it really was a non issue and I'm used to it now.

Megapixels: I'll admit, I really enjoyed having carte blanche with respect to cropping:

That is a 12-megapixel crop from the full 21-megapixel image. Since we are "switchers" back to the Canon camp, and at this point only have one lens (the 24-105L), I didn't have much choice which lens to use :) Usually, I would have used around 150-200mm on the D200 (225-300mm effective) on the D200 to get close action shots. But you use what you have... And here, the 24-105L did great! Even cropped to simulate a 200mm shot, I was able to get a 12 megapixel (larger than the D200) image out of it. That's real freedom. I'm going to be honest about this: I'm not shooting for National Geographic here, I'm just having fun in the backyard with my dogs. 12 megapixels is plenty. I'm really enjoying the freedom of cropping that the 5D2 gives!

Next time, I'll get my shutter speed up, turn on the focus assist, turn off D+, and have at it. I will post again after I give it another whirl.

It's a dog's life...

Hamilton 365: Photos of a Community

2009-01-06T06:37:00.000-08:00

The Online Photographer always has excellent links to very good photographic work. Although I may not agree with Mike's fervent support for certain products (as well as I'm certain he wouldn't agree with mine :), he certainly has a good eye and knows good work when he sees it. I really appreciate his wandering links to interesting stuff.

Recently he posted a link to "Hamilton 365," a photo-a-day series of the residents of Hamilton, Ontario, by photographer Larry Strung. Browsing through the series (home page here) I am struck with the honesty of Larry's work. It is obvious that Larry really got to know each of his subjects: each photo seems to convey a real sense of who the person is, and where they happened to be in life on the day he met them. There are some absolutely fantastic portraits in this list, such as Beth:

Yuan:

John:

Another John:

Adam:

Frankly, there are so many great photos, that it would be impossible to post them all. The thing I like about Larry's photography is each one is unique and tells a story about the person. What a neat idea, this. I'm sure the residents of Hamilton really enjoyed it, and I'm very happy Mike linked the site.

September 11, 2001: Patriotic Response to Uncertain Times

2009-01-06T06:04:00.000-08:00

Last weekend, as I looked through Canon G1 photos for a few to feature in my first blog of the New Year, I found a photo that really surprised me. I had actually forgotten that I took this photo, although the memory of the scene hasn't faded one bit. Sometimes this happens to me; I make photographs without even realizing I made a photo. I'm simply capturing what I feel or am thinking.

In this case, I'm very glad I made this capture. It occurred on September 11, 2001. I was in Idaho with my father and wife (then fiancee, we were married on November 17, 2001), on our annual flyfishing vacation. We woke to a frantic phone call from my stepmother, who explained to us what had happened in New York. We didn't have a television, so we turned on the radio and listened to reports stream in as I attempted to dial my laptop into the internet (modem days...) to get some information. We were only able to see a few photos as the phone lines and internet saturation was very high. It was a very scary and uncertain time.

I recall driving to Victor, Idaho to have lunch at, I believe, the White Horse Cafe. Or something like that. It was an odd lunch; when we opened the screen door to the dining room it creaked loudly and everyone turned to face us. They turned to see who was walking in, and what we were up to. Suspicion was high, and outsiders were treated cautiously. We were outsiders.

Upon returning from this odd and slightly discomforting experience, we noticed a farmer had cut a giant "USA" into his field. I walked to the end of the gravel drive of the cabin we were staying in, to the old cattle roundup area, to take a closer look. It was a very awe-inspiring site; although I felt a bit odd at the strange response of fellow citizens at lunch, I felt uplifted by the farmer's gesture. I thought maybe things would be OK, after all.

This photo isn't an award-winning photograph by any stretch. I'm happy I found it in my archives, though. I applied the DxO "Cross-Processed Kodak Elite 100" film look, as it seemed to match the mood and tone I felt the best. The colors are familiar, but a little disquieting. I think this fits. I hope you enjoy it.

More Canon G10 Thoughts

2009-01-05T06:58:00.000-08:00

I'm sorry I have no photos to post today, I have a bit of a backlog of sorts at the moment. I spent my photo-editing time this weekend organizing and moving files around on my file server; with 3 cameras in regular rotation (G9, E420, 5D2) and 3 separate RAW editors (DxO, Olympus Master, and Canon DPP, respectively) I need to find a somewhat better organizational system as things are getting a bit out of hand. You know how it goes.

Anyway, so no new and interesting photos to post. Lately, when I've been hunting for photography (equipment) news, I have been heading to 1001 Noisy Cameras. Its a kitchy, freewheeling news site for the purpose of aggregating camera review and commentary all in one spot. Although Noisy doesn't offer commentary of substance, it is a nice place to see if you missed anything for the day.

This morning as I sipped my coffee, I found a link off Noisy to a quick G10 review by Mountain Trail Photo. Nothing groundbreaking was revealed in the article, per se, but I like the practical style of review. It's how I "review" cameras, typically. For the "how does it work in pratice?" is the real question about a new camera. It is obvious Guy Tal likes the G10 (and, in fact, states so in his first paragraph). It sounds like a nice camera, and he has some very nice photos in his gallery to prove that it does "get along" with a photographer. Enjoy, I like Guy's style.

Also off Noisy, I found a link to Serious Compacts G10/DLux4 (Panasonic LX3 rebadged) test #4, a high ISO comparison. The Pansonic is definitely better than the Canon, which is something we knew. Something interesting is Amin's comment on the ISO disparity that DxO measured (the LX3 was essentially much less sensitive than its specified ISO, by about one stop). He equalized exposures (same aperture, shutter, and ISO) and the files came out essentially equal to the G10's exposure. Amin concludes:

assuming that 1) neither manufacturer is misrepresenting the shutter speed and/or aperture setting, and 2) that DxOmark is correct in their sensor analysis, it follows that somewhere in the processing pipeline, the LX3 data is getting pushed to the appropriate exposure index

Seems reasonable to me, based upon the processing that Panasonic is performing on the lens deficiencies.

I do think the Panasonic LX3 is a little better noise-wise than the G10. ISO 400, in particular, looks much nicer to me and perhaps would produce better prints than the G9/G10 at ISO 400 (ISO 400 is as far as I will take my G9). However, ISO 800 still looks like a "step beyond" to me in poor lighting, which of course, is where it would be used. So it's a little moot, as the Canon is definitely resolving more detail at base ISO, which is where I'd use it most.

I've been pondering the G10 lately. Now that I have my G9 back from Canon, all shiny and new, I've been pondering whether to stick it up on eBay. I usually do this just before the newest model arrives, to get as much of my investment back as possible. But regular readers of this blog will know that I hem-hawed a bit too long, and here I am. Still hem-hawing. I don't really see enough of an improvement in image quality over the G9's files (save ISO 1600, which looks much better... but I wouldn't use that setting anyway) to make me run out and buy one, and there are pros and cons to both models. The wideangle has been tempting me as of late. But really, I would mostly be buying another year of warranty. This would be useful, as it appears the G10 still has lens dust issues. Bummer. I wish that Canon had fixed that.

Of course, the extra year of warranty would be somewhat of a wash, as I'll lose a little money upgrading the to G10 (unless the Live.com rebates come back). So really, it has come down to: wideangle or not? I haven't made up my mind.

That, and the Olympus micro four thirds prototype keeps me thinking: could that perhaps replace the Canon G series, for me? Certainly it won't be jeans-pocketable as the G9 is. But it is definitely coat-pocketable (as is the E420, for me), so this is probably good enough. The big question will be lens selection. Obviously, a zoom is going to make the Olympus micro four thirds camera too large for a pocket. So it will be all about the primes. I could easily live with a 28/35mm and 85mm prime. They'd have to be pancakes, though, to keep the size down... And Olympus hasn't announced any more primes for Four Thirds, so...

Don't know. I still haven't made up my mind.

--------------------------------------------

In other news, I found a review of the Canon MP980 all-in-one printer/scanner/copier this weekend at Imaging-Resource. I have been considering a new multifunction device, as I'd love to have something that scanned film negatives easily. This one seems to, and got pretty good marks by Imaging Resource. I will do some more research.

A Moment of Reflection: How Far We've Come (in Digital)

2009-01-03T18:32:00.000-08:00

As 2008 wound down new and interesting photography announcements took a pause as many vacationed or spent time with family. I didn't have a ton of time to get out and shoot personal work, but had some time around the New Year to sit around and ponder what 2009 might bring. In a "surf around the net" moment, I happened upon an interesting site called DigicamHistory.

I must say, the first thing that struck me was the mention of "digicam." It has been at least a few years now since I have heard digital cameras called "digicams." Yet this was such a common thing back when Film still ruled the roost, and the mere fact I'd forgotten that term made me chuckle.

The site is a little klunky, but an absolute wealth of interesting history. I had an enjoyable hour or so browsing through and remembering "the beginnings" of digital photography.

I started pondering that moment. When did I become cognizant of "digital cameras?" Certainly not until after I graduated from college in 1997... Pre '97 I was solidly in the film camp. I think I recall hearing about work in the digital realm, but more on the scientific/astronomy end of things, and less with respect to photography. I guess I recall news photographers getting into them in the early to mid 90s... But in terms of "art" I don't think they hit my radar until the end of the century when resolution started to climb, and prices began coming down.

By mid 1999, I recall starting to think about buying a digital camera. At that time, I would have been using 3 film cameras; my wife's (girlfriend at the time!) Canon Rebel X (I think that was the model... have to dig it out sometime), my Canon Elan IIe, and a Nikon Nikkormat FTn. I usually kept the Rebel loaded with Kodak Gold 200, the Elan IIe with Ilford Delta 100, Kodak TMax 100, or Tri-X 400, and the Nikon almost always got Kodak Plus-X. As you can see, I photographed mostly in black and white those days. In late '99 digital cameras broke the 1-megapixel mark for well under $500, they became a reality for a just-out-of college guy on a small budget. So I made the leap.

January of 2000 (Y2K was a total non-issue, remember that?) I bought a Kodak DC215. I believe I paid somewhere around $350 for that camera, which seemed like a pretty good deal at the time. There were other cameras I wanted more, but I didn't have quite enough to buy them, so settled for the best ~$300 camera I could find, which I thought to be the Kodak:

It was well-built, having a nice metal body with rubberized grip points. The interface was klunky, in the way, and the camera itself was very SLOW. The fixed focus lens didn't provide good performance at infinity, and the lack of resolution made landscapes an iffy proposition anyway... So I tended to use it for social snaps, like this one:

I vividly recall taking the DC215 to the Detroit Auto Show in early 2000. I had never visited the Detroit show, and made a short vacation out of it with a few friends and my wife (girlfriend). I didn't even take a film camera... Probably the first trip I ever put full "trust" in digital imagery. I got a lot of shots like this:

And this:

Upon arriving home, I pretty much decided the DC215 was worthless for "real" photography. With no real manual controls, no autofocus, and no way to set ISO... I really felt like I was "fighting" the camera constantly to get it to do what I wanted to do. The photos weren't awful, but by Spring 2000 I had printed enough of them that I realized anything over a 4x6 was pretty much junk. Even 4x6s looked a little suspect if you looked at them at arm's length. My use of the DC215 dwindled through 2000, and I carried a film camera along with the digital on our annual flyfishing vacation out west. For me, digital hadn't "arrived" yet.

The progress made in the next 18 months was astonishing at the time, and even more so when I look back now. Resolution jumped to 1.3, 1.7+, 2.0, 3.0 megapixels, and even higher in 2000. I still couldn't come close to affording the Digital SLR's of the time, but one camera hit my radar and looked very promising: the Canon G1:

Rangefinder looks to remind me of the Leica I always wanted, a metal body which felt good in the hand, f2.0 lens, manual controls, adjustable ISO, and RAW. So much promise! I bought one in December of 2000, after printing photos from various test sites across the web. 8x10s were very pleasing with that camera. I used the G1 from 12/2000 through 12/2003, and captured many good photos with it, such as these:
Ilford XP2 Film by DxO

Kodak Elite 400 Film by DxO

Kodak Elite 200 Film by DxO

Kodak Elite 200 Film by DxO

And even a photo of our first house, just after move-in in the Fall of 2001:

As an aside: my wife and I were married in the Fall of 2001, and eschewed digital photography for "real film." We specifically sought out a photographer who was deeply rooted in film, and requested both Medium Format color (for posed shots) and 35mm B&W (for candids) were used. Mohsen Shirzadian of Accent Photographic did a wonderful job, and I am still very much enjoying that decision to this day. For in 2001, digital hadn't quite arrived yet. But I digress.

As you can see, I decided to play with these G1 files in DxO a little this afternoon. I just upgraded to the FilmPack 2.0, which gives me a nice standalone DxO application to apply film looks to camera files which aren't supported by DxO. I have never tried the G1's files in DxO, and I am very impressed with what it did with them. The G1's JPEGs aren't as stretchable as the Canon G9 RAWs processed directly with DxO, but simply applying the Film Looks worked well unless you get a little too crazy as the files don't pull that well. But I sure wish I had DxO back in 2000, to play with the G1's files... Anyway, I enjoyed using the G1, and it was probably the first camera that really had me putting down my film cameras. Digital was just so convenient; with the ability to instantly review your photos, as well as stamping on each file that details aperture/metering/etc, they are so much better as learning tools (no more keeping a notebook of aperture, shutter, etc for each photo!), and I enjoyed that. I was fairly pleased with the photo quality of the G1, it sure was light years ahead of that Kodak... But still not "film."

In the end, my enthusiasm for the G1 waned a bit. With the tools of the time, I couldn't really get the Black and White looks I really loved about film, and the sensor itself didn't have great dynamic range so you really had to nail everything to get decent tones. Even then, the tonal depth just didn't seem "right." Color was pretty good, however, and the G1 actually brought about a bit of a renaissance for me in terms of color photography. Man, I really wish I had DxO at the time. It took me about a minute or two per image to process the above files; I would push and pull files in photoshop for hours back in the day to get results like that.

But the G1 itself still "fought" me as a photographer; it was still slow, electronic zoom still annoyed me, and the user interface just wasn't as fluid as a film SLR. So in the end, I did go back to my film cameras on occasion. But the writing was on the wall, for me at least; the G1's printed output at anything up to 11x14 seemed to rival film, and that was as large as I usually printed, anyway. So the convenience and learning tool aspect of digital really started to win me over. And in 2003, when the Canon Digital Rebel arrived, I bought one and never looked back. I haven't shot more than 12 rolls since.

Maybe I should shoot some film before it's all gone! I recently read that Polaroid has stopped producing film. How sad! Polaroids were a part of my childhood; I never owned a Polaroid camera (I had everything from 110 to 35mm to medium format film, no instant Polaroid), but I used them many times and was always flat-out amazed at the things. I will never forget the sound of that film popping out, and then flopping it around in your hand as it "dried" (developed). Memories.

Anyway, as I look back on the early days of digital, I'm happy to be where we are now. Cameras are much faster, more intuitive and fluid, and the image quality is absolutely outstanding. While 2000 was really the turning point for digital, and 2003 perhaps when I stopped thinking about film... 2008 was really, for me, the first year that I stopped longing for film. Perhaps that's why more and more film makers have stopped producing film stock. Perhaps I should shoot some, before it's all gone...

One Year of Images in One Video

2008-12-30T12:52:00.000-08:00

Browsing the blogosphere, I found this really amazing video posted by Eirik Solheim:

One year in 40 seconds from Eirik Solheim on Vimeo.

Amazing and wonderful, enjoy!

Olympus E-420 Backfocus, it's on its way back to the Mothership

2008-12-29T09:54:00.000-08:00

I had an interesting discovery this weekend. When looking through some photos from Christmas Eve with my Olympus E-420, I realized that a few were sharper than I had ever noticed out of that camera. This was interesting, as I had just come off a weekend shooting at a Basketball game with the Olympus, and many of the photos were VERY soft. I just chalked it up to the strong AA filter.

When I looked at the Christmas shots that were sharp, though, I began to question the results I've been getting out of the E-420/Pancake combo. Although the sharp shots are nowhere the quality I get out of our new Canon 5D Mark II, they displayed levels of detail that I didn't think possible with the Olympus. What I realized was the sharp photos were all taken in LiveView mode...

I don't use LiveView much, preferring the shooting experience through an optical viewfinder, but I have used it on occasion to get shots that I otherwise would have had trouble with. Like this one, for instance:

Taken on the forest floor at Sharon Woods in Columbus, OH, I would have obviously had to lay on the floor of the forest to get that shot to compose through an optical viewfinder. LiveView made it easy, though. I have tended to use LiveView only when I had to, not all the time. But this weekend I decided to try out the Face detection in LiveView, as I had read that the Olympus would deal with challenging interior light in portrait situations in LiveView w/FD as it would optimize exposure for the detected human. Little did I know what else I would discover...

When I figured out what was ocurring, I setup a little test at home. I drew an "X" in the center of a piece of plain white paper, and taped it to a thick cookbook. I then stood up the cookbook on our kitchen island. Using a tape measure, I extended about 2 feet of measuring tape, and placed it perpendicular to the paper, using a candle to angle the tape at approximately a 45 degree angle next to the book/paper. I basically made a homemade version of the LensAlign:

I wouldn't use it for serious calibration, but it absolutely verified the backfocus problem on my E-420. I did two tests; in the first, I placed the E-420 about 5 feet from the target, and in the second, about 25 feet. I shot first using the viewfinder, and then LiveView with CDAF turned on. I repeated the test 3 times. In each case, CDAF nailed focus and TTL phase-detect backfocused. In fact, in the "long" test, it backfocused so much that the back wall of my kitchen was in perfect focus (about 7 feet behind the target).

In looking through "real" photos, I see evidence of backfocus, as well. Like this one, shot of my father and I after the basketball game; note we are in decent focus, but there is a few feet of backfocus as the only area of super sharp focus are the branches behind our heads:

And this photo, where I focused on the top of the knot and the focus plane is actually below that somewhere an inch or so from the top of the bouy (see the rope for evidence of this):

Really, perhaps the E-420 isn't as "soft" as I think it is. Certainly, the AA filter is stronger than many competing cameras. And when focus is generally decent (even a backfocusing body will give good photos at smaller apertures), the photos are softer than some would like. But it does have a very "natural" look to it when you have decent DoF/focus. I'm actually very HAPPY to figure out that my Olympus is behaving this way; as many times, especially when shooting people at wide apertures (something the E-420 should be good at), I've felt the results quite substandard. This has caused me to use the E-420 only in siutations where I want to "play" with photography a bit; develop ideas, shoot for fun, etc. In many of the very situations I thought the Olympus would excel over the G9, namely people photography, it hasn't. And that's been a disappointment. So now I know why, at least partially (the sensor will remain noisy...).

Here I will mention a little about the Olympus repair process. First, here is what I had to do with my dust-in-the-lens Canon G9:
1. Log on to Canon's site.
2. Fill out a web form indicating what the problem was, as well as date of purchase. This took all of 3 minutes.
3. Upon submission of the web form, Canon approved free warranty repair (even though I was out of warranty), and sent me an email within minutes detailing where to send the camera, as well as a repair # to log back in and track the repair.
4. I packed up the camera, a photocopy of the receipt, and a printout of the email they sent me.
5. Canon emailed me when the camera was received, when the repair begun, and when it shipped (with the tracking number). It was turned around in 3 business days, IIRC.

With the Olympus, it has already been a MUCH larger pain in the rear:
1. Log on to Olympus' site. No web submission form to be found, just a PDF describing where to send the camera.
2. Filled out the Olympus Repair order form, a downloadable PDF. I had to write my name, address, serial #s, and description of the problem on this paper. I declined to add my credit card preauthorization to the end of the form, as this shoudl be a warranty repair. Canon told me it was on the spot, no guessing.
3. As instructed, I copied the purchase receipts for my camera, lens, and all 3 warranty cards (camera, kit lens, and pancake lens).
4. Pack the camera, lenses, photocopies in #3, and the handwritten #2 in a box, and mailed it off.

So far, I have spent about 45 minutes getting all the materials together, handwriting the form, and photocopying the various purchase orders and warranty cards (for which I had to dig out the camera and lens boxes, in storage). I don't have a live repair # (obviously), and have no idea whether Olympus will repair my camera under warranty (as mentioned, Canon approved this on the spot when I submitted the web form). I have no idea how/when I will be contacted about the camera.

Definitely not as smooth of a process. Let's hope they nail the AF adjustment, though, because I absolutely adore the little E-420/Pancake combo (much less so with the kit lens, but...). I will update this blog with information as it arrives.

Lessons learned from Christmas shooting; and the G9 is back!

2008-12-27T07:20:00.000-08:00

Shooting family events with a new camera is always an interesting proposition. This is why pros always say: never take an unfamiliar camera on a job :) I shot both the Olympus E-420 and the Canon 5D Mark II this Christmas. The Olympus at my wife's aunt's house (her family doesn't like being in photos, so J thought it would be less obtrusive), and the 5D2 at my parents house.

First, the Olympus:
The Olympus is less "new" and unfamiliar to me, but unless you use your camera for shooting family, it's always a bit of a learning experience. I haven't been thrilled with ISO 1600 on the E-420, so I decided to shoot at ISO 400 and use the popup flash. I am NOT a fan of popup flashes, but the Olympus' flash isn't terrible if you dial it back a bit. Here's what I found when shooting photos of people with the built in flash:
Close Range: dial -1/3 to -2/3 (if very close) EV Flash Exposure Compensation so you don't blow out your subject, and use ISO 400 to balance the background.
Medium Range (~10-12 feet): -1/3 EV FEC, ISO 400
Longer Range (say, across the room): 0 EV FEC, ISO 400 or 800 if there is a deep background.

A few other tips:
1. Use LiveView with Face Detection for static people shots. It exposes well in that instance. Will not work for moving people: too slow.
2. Shoot RAW to allow for WB adjustments later.
3. Turn the "Noise Filter" OFF. Use 3rd party NR if you feel you need it. At ISO 400, you shouldn't have to. The Olympus NR gives poor P&S-like results, IMO.
4. turn off your AF assist lamp or it will REALLY annoy your subjects.

The E-420 performed fair-to-middling in this role. On one hand, it's super snappy and very very enjoyable to shoot with. AF generally was quick, surprisingly enough, but not always super accurate (I may have a bit of a backfocus problem with my Pancake; I need to test this). The 4/3 sensor has the benefit of allow decent DoF at f2.8, so you can get away with a lower ISO, which is good, because... There are some areas I wasn't happy with. ISO 800 is noisy, plain and simple. ISO 400 is fine, but still noisy. The Olympus RAW converter doesn't do it any favors, either; the grain size is a bit large, and the NR gives areas of very smooth NR couples with noise blobs, like a P&S camera. The popup flash is pretty weak, and you'll deal with WB inconsistencies; your subjects will need "Flash" WB setting, but the background will be rendered "yellow" from Tungsten lighting. Color inside is fairly good, but without an external flashgun isn't a strong suit. The strong AA filter means your portraits will be a little soft, which is usually a generally pleasing thing... But I couldn't help but be a little underwhelmed when looking at the Olympus' files. I kept thinking my Canon G9 wouldn't have been much worse, and certainly more flexible with its zoom. Of course, it would have been a lot slower, and perhaps I'm too hard on the Olympus' noise performance; it really is a lot better than the G9. Hmmn.

On the other hand, the files from the Canon 5D Mark II didn't disappoint. I usually bring along a flashgun to the Christmas celebration, but I didn't have time to sell/trade our Nikon SB800 for the Canon equivalent this year, and didn't want to fiddle with M settings... So I shot with available light, using the 24-105L IS "kit lens." Some things I learned about the Canon:
1. Do not shoot Auto ISO with moving people! Nikon's Auto ISO is better for moving subjects, as you can set the lower shutterspeed bounds. You cannot do this with the Canon; the Canon will essentially use 1/focal length as the lowest ISO. This sounds wonderful but in practice, you will see your shutter speeds dropping to 1/20-1/40 a lot, and therefore will get a LOT of motion blur. I wasn't paying attention to the shutter speed in the viewfinder, so later on I realized I was getting a lot of slow shutter speed shots at very low ISOs... Doesn't make sense to maximize S/N ratio at the expense of subject motion! Leave Auto-ISO for static photography. Lesson learned.
2. Remember the old "rule" of Full-Frame. When in doubt, f8. Although I knew full-frame would have a lot less DoF than APS-C, I conveniently "forgot" and selected f5.6 a lot (a common choice for my D200). DoF is too thin for more than one person at that aperture. Shoot f8-f11 for multiple people. This will take some "calibration" time for me to learn :)
3. Go ahead and pump the ISO. I still get that nervous feeling when I start cranking up the ISO... And I shouldn't. Everything up to and including ISO 3200 is simply fantastic on the 5D Mark II. The shot at the top of this page was an ISO 3200. ISO 1600 is absolutely superb, and I'll be using it a lot when I need higher shutter speeds; you really don't have to "compromise" anymore at high ISOs. This was an ISO 1600 shot:

4. I shot sRAW1 as I'm never going to be printing family photos at 3-4 feet wide. Saves space, is super sharp, and you get a bit less noise in them. I'm really enjoying this aspect of the 5D2; you get a high-megapixel wonder when you need it, and super clean and sharp 10mp when you really don't. I love the flexibility. That said, the file sizes are still pretty large, I assume due to the Large JPEG embedded in them...
5. I actually enjoy the Picture Styles options. I like Portrait for tight portrait shots, Normal for environmental/docstyle shots. Portrait seems a little pink-ish in skin tones to me, though, I may tune that a bit (you can do this). It's all about saving Post Processing time, of course.
6. The 24-105L is a great lens, I have no qualms about it. F4 is fine for close-ish people shots; f2.8 has too little DoF unless you're shooting wideangle, or just one person, so I didn't miss having f2.8 like I thought I would. Bokeh on the 24-105L is pleasing to me, even on pretty challenging subjects. This really surprised me, as I've read some not-so-great things about the bokeh of this lens. I see nothing wrong with it, here's a pretty tough test:

I also continue to be quite happy with the sharpness, contrast, and color of the lens. What a great all-around performer. I'll save f2.8 money for sports lenses.
7. I continue to love the color of the 5D2. It just has more subtlety than the APS-C cameras I'm used to; I'm certain this doesn't really come through in the heavily-compressed Flickr photos, so unless you have the opportunity to play with some full-res photos, you may not be "getting this." Trust me, it's really fun. It just seems more "real." And I think it is this subtle tonality that I so enjoy in B&W conversions, like this one of my lovely bride:

It just has more subtle tones than anything else I've owned. Here's one from a few years ago, the D200 and 85/1.4 combo:

Again, hard to see with Flickr posts; but it seems to have something to do with the colors and how they transition into each other. I really think there is something to this 14-bit color.
8. I can't wait until DxO supports this camera. The DPP workflow is a bit slow for large amounts of photos, and I'm pumped to see what DxO can do for extreme high-ISO noise which can be a little "blotchy" with DPP. And the Film Looks will be super fun.
9. Highlight Tone Priority won't save super-high DR photos from blown highlights, although it does help. If shooting available light in a harshly sunlight home interior, with harsh transitions between sunny and darker areas, you still have to remember the DR of digital won't capture it all. Use D+ or spotmeter and adjust later, and turn down contrast a bit.
10. The default tone curve is a bit punchy; I often dial it back a bit to get apparently higher DR.

I'm VERY happy with the performance of the 5D Mark II. The files are simply gorgeous to look at, if you do your part. I didn't always do mine :/ I learned some lessons and next time will be a bit better prepared for people in motion in available light...

--------------------------
Canon G9: my G9 came back from Canon on Dec. 26th; we missed the FedEx truck on the 24th. It came back in flawless shape, the lens is spotless. Noted on the order:

We have examined the product according to your request, and, it was found that the optical assembly was inoperative causing black spots. The optical assembly was replaced. Other electrical adjustments, inspection, and cleaning and parts replacement were carried out. 184.

Interesting. This matches a few posts I dug up on the web regarding the G9 dust issues, like this one on Flickr. Apparently, the A-series and the G9 share a lens assembly, and both have been afflicted with the problem. Canon just replaces the lens assembly and voila, clean camera. Sounds like I may have had some dust on my sensor, too.

It's back now, and spotless. I will try to get the F30 vs. G9 vs. E420 photos up soon.

Imaging Resource Posts 5D Mark II Image Tests

2008-12-24T05:23:00.000-08:00

Imaging Resource added a little more to their growing 5D Mark II review, posting the Image Quality, High ISO NR, Imatest, and RAW Comparison pages.

The 5D2 comes in with excellent color accuracy, slightly high contrast, excellent noise handling, and a bit softer images (vs. the 1Ds Mark III). On this page you can see a series of shots comparing NR levels on the 5D2. It's pretty clear to me that even with NR set to "Off" Canon is applying some Luminance, and especially Chrominance, NR at higher ISOs. I have seen this firsthand when developing RAWs with Canon Digital Photo Professional. Even with High ISO NR "Off" DPP will add some NR to >ISO 1600 files (a level that increases with the ISO). I'm certain this is what is occurring in the JPEGs here.

I postulate that this is also why the output is a bit softer than the 1Ds3, because on this page dcraw is used to decode the RAWs and the 5D2 and 1Ds3 look essentially the same. If anything, there is a slight softening of the 5D2 image and a slight improvement in noise at high ISOs. I'd say the improvement is certainly less than 1 stop, maybe 2/3 of a stop. The new filter stack of the 5D2 has obviously been optimized for high-ISO shooting, and does this with a slight decrease in ultimate resolution.

Also on this page the 5D2 is compared to the Sony A900 and Nikon D3x. It is quite clear from the images what has been repeated by Luminous Landscape and others on the web; at lower ISOs (up to 400, to my eyes), the Sony A900 files are a bit better than the 5D2's. From there, the 5D2 is better, and at 1600 and up the A900 really falls down on noise performance.

But the big shocker, to me, is the D3x. Whoah-Brother. This thing appears to take the best of the A900 (low ISO resolution and detail) with the best of the 5D2 (accurate color and high ISO performance) and combine the two. The initial images coming out of the D3x are stunning; it is a bit better than the Canon and Sony at low ISO, and to my eyes pretty comparable with the Canon at higher ISOs up to ISO 3200. At ISO 6400, the Canon 5D2 is noticeably better. But whatever Nikon is doing with the Sony sensor (obviously, the filter stack has been changed, and the EXPEED processing is far better than the Dual BIONZ in the Sony; perhaps better A to D as well) is fantastic. The 5D2 appears a bit sharper than the Nikon, and truthfully the performance between the two is quite close; but no longer will folks have to forfeit ISOs above 800 to achieve the ultimate in low-ISO resolution. The D3x is pricey, yes. But the initial images are indeed making it seem worth the money.

I'm glad to see this, because this means Canon needs to smack one out of the park with the 1Ds Mark IV. It will be interesting to see what kind of sensor Canon will come up with. Perhaps in the 30-megapixel range, using the 5D2's filter stack for high-ISO performance fairly equal to the D3x? We shall see. Interesting times.

The 5D2's performance is very good. It is obvious on the optics testing page that the 24-105L will benefit greatly from DxO corrections. I am very happy that DxO is quickly adding support for this combo (reportedly January, 2009). The LCA and distortion results, while slightly higher than ideal, will be essentially rectified by the use of DxO (DxO's distortion correction is well known; less known is its excellent correction of CA). Imaging Resource says:

We really liked this lens when we tested it for SLRgear a while back (see our Canon 24-105mm IS review for those details), and apart from the issues mentioned above, it seems to be up to the demands of a full-frame 21MP sensor, but just.

I tend to agree. What I've seen of the 24-105L on the 5D Mark II is excellent. But pushing the envelope a bit. Lens manufacture will have to get better as 35mm sensors start upping the pixel count. As is, those who predicted the 24-105L wasn't up to the task are dead wrong, IMO. It will never be as good as an excellent prime, but that's not the point... It's an all-in-one walkaround lens, and save distortion, vignetting, CA, and some softness in the corners (all mitigated if you use DxO), one that's more than up to the task of the 5D2.

Imatest Results of the 5D2 by Imaging Resource show the 5D2 to be near the top of the heap in Dynamic Range. It's about 2/3 of a stop worse than the Sony A900, IMO the current king of DR (no matter what the measurements say). The results also show the camera JPEGs to be much lower in DR than the RAW images; this seems to reflect Canon's choice of a somewhat higher than normal contrast curve in default processing. All in all, a good result. It is obvious that the Canons aren't quite up to the Sony in terms of DR; especially at the highlight end of the range. You can somewhat (visually, at least, you're really just shifting around available DR) compensate for this with the Highlight Tone Priority (D+) setting on the 5D2.

About image quality (print quality, specificially), Imaging Resource has to say:

We kept running out of superlatives to describe the Canon 5D Mark II's printed output: These are some of the cleanest, highest-quality images we've seen yet from a digital camera. Exceptional print quality, good color, sharp 24x36 inch prints from camera JPEGs (with a little unsharp masking). High-ISO shots are surprisingly clean, ISO 6,400 is good up to about 13x19, albeit with a little noise in the shadows and some loss of subtle subject detail.

Yikes. I haven't tried a 5D2 ISO 6400 file at 13x19, but it's clear we're really getting into the "phenomenally good" category with any of these cameras nowadays. Back on November 7th, I said:

I think we're very close to the mythical "film replacement" DSLR.

Now, more than ever, I find this to be true.

G9 on it's way back; Olympus E-420 and IS

2008-12-23T11:35:00.000-08:00

I just received an email from Canon, my G9 is on its way back to me. If you recall from my earlier post, I found that I had the "dreaded dust inside the lens" of my G9 when looking through some photos from a trip this Fall. Bummer. I opened a service ticket with Canon, dug out my proof of purchase, and was surprised when Canon offered to fix my G9 for free (my G9 was purchased 13 months ago). It took about a week for it to be fixed. Good stuff during the holiday rush. They are even sending it back air delivery on Christmas Eve... THANKS, CANON!! That's great service.

Anyway, now I need to decide: keep the G9 or sell for a G10, now that it's clean as a whistle?

------------------------------

Since I have been shooting a bit with the Olympus E-420 lately, I have been thinking quite a bit about it. I really should codify some thoughts into a mini-review of sorts; it is an interesting camera, full of potential and some compromise. One of the things that has baffled me is the fact that Olympus offers in-body Image Stabilization on the E-520 and E-3, but not the E-420. I always figured this was an arbitrary "feature differentiator" designed to push folks to part with the greater sum for the E-520 (and put up with the greater size of that camera)... However, it looks like I may be wrong.

Rriley on DPReview posted this photo of the Olympus sensor-shift IS and lens mount:

The cylindrical device to the right of the mirror box is the sensor-shifting motor. Then take a look at the E-520 body strip:

Notice the large capacitor where the camera's grip is. This is the capacitor for the flash unit. On the E-520, this large cap fits just where the grip bulge is. Of course, on the E-420, there is no grip bulge... So Olympus moved the flash cap to right where the sensor-shift motor is on the E-520. Here's the E-420 strip:

Ah, HA! That's actually quite clever of the Olympus engineers. And that explains why no in-body IS exists on the E-420. Either Olympus would have to design a new sensor-shift unit with a much smaller motor, or make the flash cap smaller, or... Who knows.

Some folks commented that they wished for an E-420 with in-body IS instead of a flash... That is an interesting idea, but in practice, I think I'd like to keep the flash. I'm not thrilled with the high-ISO in low-light performance of the Four Thirds sensors, so indoors for people I actually use the flash. This is shocking for me, as I am NOT a fan of flash photography, and cant really recall any time I've ever really enjoyed using an on-camera flash. But the Olympus flash isn't so bad, especially when you dial FEC back to -1/3 EV. It's pretty natural, really, and IMO far better than an ISO 1600 shot (which would likely be needed, and these little sensors really fall down at that ISO).

Anyway, I found this interesting, and thought I'd share it.

---------------------------

Speaking of the Olympus E-420, if any readers out there have this camera, and have DxO, PLEASE fill out the following form and try to convince DxO to support the E-420. IMO it could benefit HUGELY from DxO's prowess in:
1. RAW Conversion (the Olympus conversion doesn't give much of an advantage of JPEG, ACR Proves it can be done a bit better)
2. Noise Reduction! The fine-grained noise patterns produced by DxO's NR technology would surely benefit the somewhat noisy E-Series cameras
3. Distortion. The Pancake has a lot of this, and could really benefit from DxO's corrections. As could the consumer 14-42 and 40-150.
4. Color. I really enjoy the E-420's color, and it's even better with DxO Film Pack. Please!

What's Wrong with Mike Johnston's 5D2? The Photographer...

2008-12-20T06:29:00.001-08:00

As I mentioned earlier in the week, Mike Johnston (a self-proclaimed Nikon nut who believed Nikon would never, ever build a full-frame camera) posted a review of the Sony A900, Canon 5D Mark II, and Nikon D700. In that review, he claims:

And I haven't bought, and am not going to buy, any of these cameras, so anything I say is not brand partisanship on my part, or self-congratulatory pride of ownership, or anxiety that I might have made the wrong choice, or insecurity that somebody else might own something better than mine, or fear that having a "less than the best" camera might reflect on my bad photography, or any of the other motives that fuel all the passionate nattering.

I am sorry, Mike, but you're not being truthful with yourself. Come on: any journalist knows that it is impossible to be 100% "unbiased." We all have our biases, baked in through history and experience.

And you're a Nikon shooter. A passionate one, whether you care to admit it or not. Try to be more introspective and truthful with yourself in the future.

So although Mike claims to be unbiased, and offers the standard, "hey, I'm not plunking my own money down, how could I be biased??" statement... It doesn't fly. I don't care if he softens his disdain for the Canon with overly wordy and escaping-blame prose; he slams the 5D2, plain and simple. He claims :

my holistic overall impression is that its images just look more digital, in some fundamental way, rather than just "photographic."

Ouch, the kiss of death. It's what turntable fans say of CD players. How do you argue with that? Well, you can, and I will. But later. For now we'll stick to measureables (viewables, really).

I have to say, when I read Mike's review, I thought, "WTF is wrong with his 5D Mark II?" I'm not seeing these "artifacts" and problems with mine. And Mike states:

The 5D Mark II is better than the 14.6-megapixel K20D. But it's not that much better.

WHOA, NELLY. Come on, Mike, you're 100% full of it. You see, before I decided that this really WAS the time to move to full-frame (unlike Mike, I knew this was the route to take since my first DSLR in 2003, I just had to bide time until prices came down), I looked at ALL APS-C alternatives. At the time, I was holding two Nikons, with some nice Nikon glass: a D50 and a D200, with the 18-200, 17-55/2.8, and 85/1.4 Nikkors. A heck of a collection! The D90 was the logical alternative for me, and was fairly happy although not astounded with its image quality... But certainly not with the build. Disappointing. I wanted something more substantial, better. And the D300 turned me off, as the IQ looked actually lesser than the D90. I really wanted better high-ISO noise capability, too. So I expanded my search.

I have a propensity to like unusual things. Things out of the mainstream, off the beaten path, a little esoteric. Like my Olympus E-420. Such a sweet little camera, warts and all. I feel good using it, like the images within certain parameters, and it'd not your standard CaNikon. Definitely not mainstream. I kinda dig that. Just like my Triumph Daytona 675. Triple cylinders. Not your average UJM supersport.

I heavily researched the Pentax K20D (and well before its price drop... for $700, today's price, I still consider grabbing one for fun!), and I was totally geeked about the Pentax line of primes. I love primes. I love the discipline they force, and the subtle ways they draw more eloquently than most zooms. And the Pentax had in-body IS... A prime lovers' dream!

But something really smacked me upside the head: the image quality. It's not very good! Blame it on the lenses, the Samsung sensor, I don't know. Look for yourself here and check out lens performance here. You can actually downloaded the standard test photos on most of the lens tests page. Now compare these results with any CaNikon. Or Olympus, even. For kicks, compare the Pentax pancake with the Olympus pancake. Ugly.

So we know Mike is prone to speak in with hyperbole. Because the K20D's IQ is waaaaay below the 5D2. Way below.

In any case, you might ask: "but he show some clear evidence that the 5D Mark II isn't that great!" Like this shot (bad image copyright 2008, Mike Johnston):

Yep, that looks bad, alright. As bad as Sony high-ISO JPEGs (just a little humor for ya...). What's wrong? The photographer. Poor focus, poor exposure, and poor RAW converter choice. We all know Adobe/ACR is doing no favors for the new Canons (50D and 5D Mark II), yet DPReview and now Mike Johnston refuse to look at anything else, yet still proclaim their reviews "truthful." I present one crappy 100% crop as evidence:

ISO 12,800, DPP RAW converter, noise reduction sliders reduced from default.

What do you think? Which looks better? My dog's nose. And it was taken at a stop and a half higher ISO. Why does it look "better?" Because I actually put a little effort into 1) focus, 2) exposure, and 3) evaluating RAW converters.

Come on, Mike, I don't buy your excuses posted yesterday. While I agree that a "reviewer" can't evaluate EVERY RAW converter on the planet for a review, doesn't it make sense to at least try a few, when you first start using a camera? Or even before you buy a camera? I've had my 5D2 for about two weeks (only shot with it for a week and one day, b/c it's a Christmas present and I promised I'd put it back under the tree!) and I had no problems determining Adobe/ACR absolutely SUCKED for 5D2 photos.

Mike goes on to bash the 5D2 for CA:

I don't see much in the way of purple fringing at all from the 5D Mark II, but there's what Carl Weese calls "blue replacement," by which narrow objects imaged against a brighter background change from their own color into a darkish pastel hue.

I dunno, I don't see this one as much of an issue. It certainly isn't restricted to the 5D2, all DSLR's see this with various lens/sensor/exposure combinations. I suppose if you overexpose tree limbs against a bright sky using a zoom lens known for LCA, and then blow it up to 300%, it might give away that this was a digital photograph and not film. But good RAW converters like DxO automatically correct for this, and I've never, in the last 10 years of digital photography, had a print RUINED by this phenomenon... So I'm not worried.

The Black Dot issue? I still think its no biggie. Overexposed highlights? I haven't had a problem with that on the 5D2 yet. The A900 surely is ahead here, but the D700 is not. The Canon gives you D+ to play with to protect highlights, or perhaps you could try what photographers for a century have tried: protecting highlights on your own.

Finally, I present a few links for you. Look here at the 5D2 files, and here at the D700 files. View at 100%, convert a few RAWs for yourself. Which one looks more "digital" to you? Both cameras are amazing, for sure. Which one looks more or less "digital?" Either of them? Let me know what you think.

But, I suppose if you WANT to make the Canon look bad, or you're simply too lazy to "learn" a camera, or you like viewing at 300%... Anything can look digital. I'm being harsh on Mike because I don't buy his weak excuses. And very harsh because he claims to be "impartial." Baloney, Mike. Call it like it is.

What's my angle? I think Mike's deluding himself. What's wrong with his 5D2? Nothing, it's the photographer... The D700 is an also-ran. It's last-year, and prices are dropping precipitously for a reason. While the body and AF are awesome, the IQ is behind... The A900 IS a game changer. It forced Canon's hand; the A900's megapixel-bursting sensor forced Canon to drop the 1DS3's sensor in the cheap-ish 5D2. But the A900's high-ISO performance is disappointing, and the Zeiss lenses are few and expensive. The 5D2 IS a game-changer, too. Awesome base-ISO IQ, combined with D700-ish high-ISO performance and... HD Video. A true game-changer. You'll see the D700 and D3 disappear soon enough; the Sony has also forced Nikon's hand... And, in fact, Nikon will be using its sensor :)

Canon's Official Response to the Black Dot Issue

2008-12-19T12:28:00.000-08:00

I have known about this for a few days, but today I received a message from Canon about the "black dot issue." Also, a short mention of banding. (see message at the bottom of my post)

In my use so far of the Canon 5D Mark II, I have really searched and found very little evidence of "black dots." I have looked at a photo I took of a Christmas Tree at the Statehouse at 200%, and can't see any evidence of black dots. Last night I looked through some night photos I took, and I may have found one photo showing the black dots on very small overexposed point-source. At 200-300%, I could make out one or two lights with what I think might have been the "dreaded black dots." Nothing like some of the shots I have seen posted on the web. Since I had to spend 5 minutes looking at the photo to find these dots, I am maintaining that with my camera, this is really a non issue.

Really, in actual photography, I still agree with Ken Rockwell. It's a non-issue.

Banding, though, I would love to see this "fixed." I have noticed a bit of banding at ISO 12800 when shooting sRAW1, and you can see it in this photo:

ISO 25600 photos most definitely show banding; enough that the usefulness of that ISO is reduced. It would be most welcome if Canon would fix this issue, because my testing shows sRAW1 is really great to shoot at >ISO 6400; noise is less, sharpness is better, and you don't end up with a huge file mostly full of noise.

That said, I am finding this to be an amazing low-light camera. Really, my uses for ISO 6400 and up are slim; you can essentially shoot in complete darkness with the 5D2, and if you want, make it look like daylight! Coupled with the IS on the 24-105L, I am really enjoying "exploring the night:"

Canon really got the AA filter "right" on this camera, I think. I was pondering this last night as I looked at an image I captured on the way out of work:

Nothing truly special about this image, but it is part of my life. I pass this spot every day, passing by on my way from my parking space to my office, as I have for the past 10 years. I am so used to the this part of my life that it is almost difficult to take photographs around my office; I have seen everything so frequently that not much "catches my eye" anymore, especially those things which don't change. I started thinking about this the other day, though... And decided that I would make an "assignment" to myself to start documenting this part of my life a little. It has been hard, truthfully, but feels a little liberating when you nail it.

Anyway, I digress. I was looking through a few photos and was struck by the differences between the 5D2 shots and my Olympus E-420 shots. I shot the E-420, and settled on Manual Exposure, 1/15 sec, f2.8, ISO400. I chose these values because 1/15 is as low as I can handhold the E-420 with some degree of repeatability (again, I wish Olympus had given the E-420 sensor-shift IS!). When I shot the 5D2, I chose to mimic the Olympus by using ISO 1600 and f5.6 (at 50mm and f5.6 the 5D2's DoF equals the E-420 at effective 50mm and f2.8), and the Olympus shots were about a stop underexposed so I chose to use 1/8s shutter speed as I figured the IS could cope with this. I also shot sRAW1 as I like this for low-light work, and it provided a handy reference point to the Olympus.

The Canon exhibited less noise, which I knew would occur from my testing, but the real surprise was the difference in detail. NOt necessarily sharpness, per se; I'm not a sharpness freak. But it's clear to me that the Canon really exhibits much finer detail that the Olympus, even with both cameras shooting ~10mp photos. It is known the Olympus has a very strong AA filter, and I believe my eyes are telling me the Canon's is weaker. Much weaker. Check out the E-420 photo here, and compare to the 5D2 photo above (yeah, different photos, but I don't do much "testing" only capturing what I like :)

At night, the Olympus' strong AA filter is a weakness. A night photographer will usually open up his/her aperture much further than normal, to let in more light and prevent using the highest ISOs and the noise that accompanies them. One of the results of this is that there is much more potential to have parts of the scene that are not in focus; I have found anything out of focus in a night shot to be potentially problematic; showing noise much more prevalent than sharply detailed points of interest. The Olympus strong AA filter not only blurs out detail, but actually seems to contribute to this phenomenon and increase the apparent effects of noise. The Canon does better in that regard.

That said, I actually like the strong AA filter on the Olympus in good light. I think people focus way too much on "sharpness" nowadays, and super-sharp photos really tend to look "digital" to me and not "analog." Film never looked so sharp. Yes, it had sharpness... But not the whistle-clean icicle-through-the-heart biting sharpness that can come from digital. The Olympus takes a tack off the beaten path, and uses a very strong AA filter which does smudge detail, but also looks more natural to my eyes.

I'll leave the ultimate determination of the Canon 5D Mark II's AA "strength" up to the fiends at dpreview's forums... For me, though, this is proving a nice balance. The Canon is packed with detail, yet smooth and natural looking to me. I do not agree at all with Mike Johnson's assessment that the 5D2 files look "digital". Really, to me, what we are seeing nowadays is "flavors" of digital that are rather like types of film; you either like the emulsion or you don't. But as I look through the few hundred 5D2 shots I've taken so far, I have never been struck by something looking "digital" to me. Quite the contrary, in fact. I will share some thoughts about this in the coming days, and talk a little about how my experience has differed quite a bit from Mike's (primarily because although he claims to be unbiased, he's firmly a Nikon shooter, and being unfamiliar with the Canon has caused him to use it poorly). But one thing I can say for certain is that the AA filter that Canon has chosen is undeniably NATURAL in its presentation, and not at all digital. I think it's a bit like goldilocks, to me. "Just Right."

The 24-105L is performing well, too, by the way. Check out the night shot of the parking lot above ("greenlight"). Look into the corners of the photo. Very, very little softness at f5.6 and 24mm. Very impressive for such a flexible zoom.

Here's the email from Canon:

Dear Nathan ,

Thank you for using Canon products.

We have learned that some users of the Canon EOS 5D Mark II digital SLR camera have identified two types of image quality phenomena that appear under certain shooting conditions.

"Black dot" phenomenon (the right side of point light sources becomes black)
Vertical banding noise

We are currently investigating and analyzing the causes, and examining measures to reduce or eliminate these phenomena by providing correction firmware. An announcement will be made on the Canon Web site when such measures have been determined.

Details of the phenomena and shooting conditions under which they are likely to occur are as follows.

"Black dot" phenomenon (the right side of point light sources becomes black)
When shooting night scenes, the right side of point light sources (such as lights from building windows) may become black. The phenomenon may become visible if the images are enlarged to 100% or above on a monitor or, if large prints of the images are made.
Vertical banding noise
If the recording format is set to sRAW1, vertical banding noise may become visible depending on the camera settings, subject, and background.
Vertical banding noise is not noticeable if the recording format is set to sRAW2.
Vertical banding noise does not occur if the recording format is set to RAW or JPEG.
Noise can be reduced if C.Fn II-3: Highlight tone priority is set to 0:Disable.

Canon always strives to provide the highest quality products to its customers. We appreciate your kind patronage and support.

This information is for residents of the United States and Puerto Rico only. If you do not reside in the USA or Puerto Rico, please contact the Canon Customer Support Center in your region.

Contact Information for Inquiries:
Call Center
1-800-828-4040 (toll free)
8:00 a.m. - Midnight, EST (M-F)
10:00 a.m. - 8:00 p.m., EST (Saturday)
E-mail: carecenter@cits.canon.com

Sincerely,
Canon U.S.A., Inc.

Humorous Response to the Furor over the Black Dot Issue

2008-12-16T13:20:00.001-08:00

Found on dpreview. Love it!

Cameralabs Canon 5D Mark II Review: 5D2 is applying NR at high ISO

2008-12-16T05:26:00.000-08:00

Yesterday Cameralabs posted their review of the Canon 5D Mark II. I like Gordon's more practical review style, and his comparison of different cameras in real world situations as well as in a lab are always nice to see.

The review is fairly predictable, allowing that Gordon basically shoots in 100% default settings (and JPEG, except where noted). I would have liked to see him do some sRAW1 vs. Nikon D700 shots, and have some more D700 comparisons.

The only thing of note, really, is the high-ISO test page. On this page, Gordon shoots the 5D2 at all available high-ISO NR settings (disable, low, standard, and high). It's clear to me that even "disabled" Canon is applying NR to the JPEGs. This shot is at ISO3200:

I can say with great certainty that this isn't what I see in DPP when developing a RAW ISO3200 image with NR truly off.

So that's that. I suppose DPReview will slam the 5D2 for "heavy noise reduction at High ISO because of the ridiculous amount of pixel packed on the sensor" or something like that.

My Review of 8GB CF EXTREME III MEMORY CARD

2008-12-13T16:11:00.001-08:00

Originally submitted at Abes of Maine

Extreme III is the exceptional, unbeatable best-in-class solution for high profile professional photographers. As the elite flash memory cards available on the market, Extreme III has extremely fast read/write speeds. This line provides the durability and high-speed quality demanded for serious pho...

Fast, reliable, well-priced.

By loplop from Columbus, OH on 12/13/2008

Pros: Reliable, Fast, inexpensive

Cons: Annoying blister pack

Best Uses: Digital Photography, HD Video

Fast, reliable, and not at all expensive. What more can you ask? Well, doing without the blister pack would be nice... Seriously, though, it's perfect for the Canon 5D Mark II. You don't need anything else.

(legalese)

Canon 5D Mark II black spots, a review, and Flickr

2008-12-13T06:29:00.000-08:00

A jumble of items to discuss today.

First, the "dreaded black spots" of the Canon 5D Mark II.

Are these black dots an issue? Not really. Ken Rockwell, called by some a "maverick" but I tend to think of as more of a quack, said:

I'd never notice this until some dweeb, the same sort with the Star Trek jokes, points it out. Now that I have seen it, do I care?

Nope!

I sheepishly admit agree with 'ol Ken here (this is actually the second time Ken's been right about something. Maybe I'll blog on the other someday). I have not yet seen the dreaded black dots in any real photographs. Before I knew of the black dot issue, I even took a photo of a christmas tree. With no black dots. After I learned of the issue, I tried to replicate the black dot issue. I was able to, but no worse than Ken, and only when I was trying. If it's fixable in firmware, great. If not, I'll still enjoy the camera.

Regardless, Chuck Westfall stated publicly that Canon will be commenting on this soon. This could only mean that Canon will publicly acknowledge the issue, and will likely not yet involve a fix. Whether Canon can or will fix the issue is still up in the air. Wouldn't it be great if a firmware fix for this came out by Christmas, so the internet forums would stop filling with bad photos of overexposed Christmas lights against a black background?

-------------------------------------
Pop Photo posted a "test" (first quick look, really, with some measurements) of the Canon 5D Mark II on Friday. Full of usual Popcorn Photo glowing praise, but in a surprise move, Pop Photo actually comments negatively about something: the autofocus. It's not as good as the D700:

The 5D-era AF, however, is no longer competitive with the blazing speeds we see in most DSLRs these days...Oddly, some older and newer Canon models have faster AF systems. Just as Nikon trickled down the D3's AF to the D700, we think Canon should have done this with the 1Ds Mark III and 5D Mark II.

That would have been nice. Some believe Canon purposely hobbles anything below the EOS 1-Series autofocus as a "line differentiator." Could be. Will this matter to me? Not sure yet. My Nikon D200 probably had better autofocus. It remains to be seen whether I will ever need anything more than the 5D2's system, though; so far, I haven't had any issues, even in low light. And truthfully it seemed to track a moving dog better than the D200 (hard to compare, though, different days with different dogs, only the yard was the same.). We'll see.

Otherwise, they loved it. Notably, the color, which I have also found was the surprise shocker of the 5D2. It's head and shoulders above any digital I've worked with before; the subtle tones and transitions are really intoxicating. Not a headline-grabbing number, for sure, but this is one of the huge wins for the 5D2.

---------------------------------
Flickr is the latest to suffer at the hands of the economy which is in full recession now. Sadly, Flickr laid off 3 employees this week, including a top designer largely credited with the clean google-like-look of Flickr. I don't know what this means for Flickr, but I'm hoping that they stick around. Although I find it a better place to post and contemplate small numbers of photos rather than, say, a vacation full of pictures... I still love Flickr, and would hate to see it go by the wayside or be absorbed into some soulless corporate giant.

----------------------------------
More Flickr: although no comments about the layoffs were made, nor was any hint of trouble at Flickr on the table, I found this interview at PopPhoto enjoyable to listen to. Enjoy.

Canon 5D Mark II Review, and even more on AF

2008-12-10T05:57:00.000-08:00

Phil Holland has just posted his review of the Canon 5D Mark II. As a pro in the motion picture industry, I thought his take on the video features of the new 5D2 was very interesting and informative. He is also a Canon 1Ds Mark III user, and as such, it was fasincating to see his comparison between the two cameras.

About the video capability, Phil says:

Possibly the most interesting and yet controversial features on the new Canon 5D Mark II is the ability to record full resolution 1080p movies...The downsides to having full HD on the Canon 5D Mark II is that Canon has limited photographers/filmmakers to just one shooting mode. And that my friends is the Devil born "fully automatic" mode.

I am no videographer, but I do think it's a shame Canon doesn't allow full manual control over exposure. Essentially, what they have given us, is a ridiculously high quality point and shoot video camera. But what photographers across the ages do with tools is learn how to get around their downsides... And the downside of not having full aperture control can somewhat be mitigated with Neutral Density filters. You can see this in action with the 50/1.2L in Phil's video.

The Canon 5D Mark II is a tool and if you play to it's strengths when it comes to video you can indeed produce some extremely lovely footage (<--we really need a digital term for footage nowadays).

So just what are those strengths? Here's a few...

- Full frame depth of field
- The ability to utilize a wide variety of lenses.
- High quality video codec that produces workable footage.
- Light and compact camera form factor. (this is good and bad)
- Relatively inexpensive entry level camera price considering the alternatives.
- Amazing low light performance that rivals cameras that are MUCH more expensive.

Interesting stuff. I have only owned 1 video camera in my life, and rarely used it. I haven't the foggiest notion of how to be a cutting edge videographer. Perhaps it is something I will grow into, over time, with the 5D2. I urge readers to read his full review, too, as Phil has some interesting things to say about history and why the 5D2 is quite revolutionary.

About the "stills" photos, Phil says:

Very quickly I was able to determine that the image resolution and sharpness was deadly similar between the 5D Mark II and the 1Ds Mark III.

He has many side by side photos that illustrate this quite well, both real-world and test. In this photo, you can see just how close the two cameras are:

Also, of interest is Phil's thoughts on the exposure/color difference:

I am noticing something that is just very slightly different between the two images. It would appear to me that the 5D Mark II seems to be perceptually brighter than the 1Ds Mark III and for whatever reason the 1Ds Mark III, even with the same lens, holds a slightly bluer look to it. Perhaps it's a new bayer filter tint coming into play.

I haven't seen anyone else on the web comment on this, but it is absolutely clear from Phil's photos that this slightly more sensitive, slightly warmer tone is something that the 5D2 will offer. Whether this is pleasing or not will depend upon the photo, and could certainly be compensated for in post processing... But the fact it is there is quite interesting. It is entirely possible that Phil is correct; the 5D Mark II does indeed seem to have slightly better high-ISO performance than the 1Ds Mark III (which Canon alluded to before the camera was released), and this could be why. Quite interesting!

Generally, Phil believes the image quality of the two cameras to be extraordinarily close, the difference in price justified by the far better AF and build of the 1Ds Mark III. This seems a very level assessment. I urge bloggers to read Phil's entire review, it is really quite good, and far more pragmatic than we will see from some of the "big name review sites."

--------------------------------------

More on the 5D2 AF Microadjustment

Last night I sat down to edit my Christmas Card '08 photos. We always take photos of our dogs in some Christmas finery, and then compile the photos on a Shutterfly Christmas Card. I've been pleased with the quality in the past, and it's cheaper to have Shutterfly print them than to burn through ink at home.

This year, I shot with the 5D Mark II. It was a bit of a difficult shoot for me; the camera is still highly unfamiliar, and pets in motion are not the easiest to photograph regardless. Throw in some snow, a black dog, and you have a worst-case photographic scenario :)

I had the AF Microadjustment dialed in for the shoot, and I'm sorry I did. I must have been floating between the close range where the microadjustment was really putting the AF in adjustment, and the mid/longer range where it was now OUT of adjustment. So, lesson learned. I will not be using AF Microadjustment on the 24-105L for anything but close range static subjects. Which is to say rarely; because truthfully, now that I know there's a little AF tweak needed in that situation, why not do it manually? The viewfinder is large, bright, and good enough for that (or Live View x10 is even better).

So what is AF Microadjustment for, then? I think it's for a mis-calibrated lens (or body). If your lens constantly needs +/- some AF microadjustment, then this will work beautifully for you. Canon wisely allowed this adjustment per body, across all lenses, or PER LENS. So that seems to be the most likely scenario, and actually makes me more apt to try a third party lens (like the very good Sigma 50/1.4) where body/lens adjustments could be more likely needed.

Other than the somewhat soft photos, how did the shoot go? Fairly well. The 5D Mark II dealt with the very high dynamic range of snow and black dog VERY well. With just a smidge of DPP tweaking, I was able to bring out detail in Tucker's fur, while maintaining the highlights. AF performance was generally excellent, as well; I did some AI Servo shots, and the 5D2 locked and maintained focus very consistently, even on a running dog. I'll be interested to try this out some more over time. I think it equaled my D200 in this regard.

I still have some work to do in learning best aperture/focal length combos on the full-frame. But that's to be expected. I need more time with the 5D2, but as of today, have to wait until Dec 25th to shoot some more... I promised my lovely wife that I'd wrap it back up for Christmas after a week of shooting, so I'll do that. I have a fair amount of photos I took over that week, though, so I will continue to develop/post interesting captures.

More on Canon 5D Mark II AF and a funny video link

2008-12-09T07:45:00.000-08:00

I often visit The Online Photographer when not much else is going on in the photography world. Nearly every day there is something fun or interesting there, and it's nice to get away from the technical minutia of some blogs.

Today, there was a great blog link to a hilarious YouTube video:

Simply great.

------------------------------------

Canon 5D2 AF Microadjustment: I played around with this some more yesterday evening, and found that my "adjusted" (to ~1.5m) settings do not give sharp results at ~15m. So it seems, at least for the 24-105L, that microadjustment might need to be different for close & far objects. I felt a bit gutted about this at first, but I then realized that Canon allows you to store ALL your settings on 3 separate "Custom" profiles, selectable right on the mode dial.

I use this frequently with my Canon G9; when photographing a sporting event, for instance, I set C1 for JPEG continuous, full telephoto, and open aperture. I use C1 to quickly fire up the camera for action shots. I set C2 to RAW, wide angle, and F4 (for optimum sharpness). I use C2 to quickly switch to high-quality "documentary" mode.

So why not do this with the 5D2? Set C1 for close range photography, optimum settings for portraits and the like. Set C2 for general photography and longer distances. I'll try it, and let readers know what I think...

Canon 5D Mark II AF Microadjustment

2008-12-07T10:55:00.000-08:00

I noticed some softness in shots taken at fairly close range with an open aperture, where the slim DoF reveals focus anomalies with aplomb. Nothing shocking, mind you, but the biting sharpness I have seen at longer distances at similar focal lengths led me to believe I might need a little adjustment for the 24-105L. Canon has recently started providing AF microadjustment on their new cameras (1D3, 1Ds3, 50D, 5D2), so I figured I'd give it a try.

There are multiple focus calibration methodologies to be found on the web, but one stuck out in my head as being really great. Keith Cooper from Northlight Images has posted a how-to here for the 1Ds3. One can follow the same method for any Canon with microadjustment. Essentially, you setup your laptop, pop your camera on a tripod, square it all up, and fire up Live View. Manually focus on the odd circular-patterned image provided in the article, and make it look like the one on the left (the right is slightly OOF):

I did this a few times, at a few focal lengths, and decided that +8 backward seemed to provide the closest match between what I thought was focused, and what the camera thought. At +8, the focus ring never moved when I shut off LiveView and half-pressed; it stayed perfectly still. I tested at 24mm, 50mm, and 105mm.

We'll see if this provides any increase in sharpness. It's such an easy process (and so easily reversible) that I figured "why not give it a shot."

5D Mark II Around Columbus, and Around the Web

2008-12-06T13:18:00.000-08:00

Yesterday I was able to get outside a bit with the 5D2 to take some photos. It was a chilly (27F, -3C) but sunny day, and save the need to bundle up, a good day for some photos. The lighting was fairly dramatic for a winter's afternoon.

I particularly liked the strong shadows cast by the winter light on the Statehouse, seen above rendered in B&W (RAW ISO 100 shot developed with DPP, monochrome, +3 CON, Red Filter). I continue to love the B&W output of this camera.

I'm quite surprised at the quality of the 24-105 F4 L lens, also. Before choosing the Kit (5D2 w/24-105), I did some research and felt pretty comfortable that the 24-105L would be a pretty good match for the 5D2. My initial thoughts confirm this; the lens is a fantastic all-around choice for the 5D Mark II. As long as you do your part, it is plenty sharp; you do have to be quite steady to get 21-megapixels of resolution from any lens, but that's not a knock on the 24-105L. Any lens on the 5D2 will require a steady hand for best results. I do see some blur in the extreme corner of the frame, the lower-left in my copy is probably a little worse than others. But this isn't of much consequence in most photography, so I will not worry about that. Across the vast majority of the frame, you are rewarded with great resolution, color, and contrast. On full-frame, the F4 aperture can give pleasingly blurred backgrounds at the longer focal lengths, such as in this shot in front of the Statehouse:

Or this shot, taken at ISO 3200 inside my office:

I also really love the color rendition from this sensor. It has a depth that I have not seen before in digital. Whether this is due to the sensor design, the 14-bit processing, or something else entirely... I don't know. But what I know is when developing a 5D2 file, I can easily get it to replicate what I see, and with great beauty. As readers know, I am a big fan of the Kodachrome look. Usually I use DxO to apply a Kodachrome film look, but since it does not yet support the 5D2, I have been using DPP solely. I have found something between +3 and +4 CON, and -1 to -2 SAT usually gives a nice Kodachrome "feel:"

---------------------------------------
Around the web, I see two bits of interest to 5D2 fans.

1: DxO released 5.3.1 yesterday, and there's a blurb about 5D Mark II support here:

Other recently announced DSLRs, including Canon EOS 5D Mark II, will be supported early next year.

2: Imaging Resource, as predicted, is hard at work on their 5D Mark II review. You can see their becy of test shots here. This means we'll be seeing a review in the next two weeks. I haven't looked through the test shots much, save comparing the 5D2 and 1DS3 ISO 100 shots. The EXIF says they were taken with the same lens, but to my eye, the distortion is quite different between the two. The 1DS3 is marginally sharper, with a tad more contrast. I wonder if the 5D2 shots were really taken with the 24-105L? I can't see how the distortion on the 5D2 shots would be so much greater if the lenses are the same, and this would explain the slight difference in sharpness and contrast. I should note: it's slight, and just a tad more in PP levels the files. Both are gorgeous. The 5D2 is shaping up to be a real bargain.

Two 5D Mark II Links

2008-12-05T12:28:00.000-08:00

I was able to get out and take some photos with my 5D Mark II this afternoon, so I will post a few more in the coming days. It has been a busy week, and I promised my lovely wife that I'd wrap the 5D2 up and put it under the tree this weekend... So these shots will have to tide me over for a few weeks :)

The first link is shocking: Tabb Firchau flying his RC helicopter with the 5D2 in a gimbal mount. Watch out for the fisheye shots! He's a heck of a pilot, and has way more guts than me. You won't see my 5D2 flying through the air anytime soon. At least I hope not.

The second link is the first 5D Mark II review that I have come across, at PhotographyBlog.com. Photography Blog always seems to do a good job explaining features and functions, but the actual testing isn't very thorough, IMO. It is, however, something to feed your 5D2 passion. I suspect we will be seeing many more reviews, quite soon. Perhaps cameralabs and imaging-resource will sneak them in before Christmas.

Until next time... I leave you with a shot from my Nikon D200, take at the Neue Pinokothek in Munich, Germany. I really like this photo. I developed it in DxO and applied my favorite Kodachrome look.

Can a Camera See in the Dark? The 5D Mark II at very high ISO

2008-12-03T17:11:00.000-08:00

Can a camera see in the dark? With the Canon 5D2, I'm not sure... But I do know that it can gather light rather like, and in some cases, better than the human eye. Take the photo I posted yesterday, in St. Joseph's Cathedral; that photo, shot at ISO 3200, it so closely matched what my eyes saw in the church that I was quite taken when I opened it. Not only is the tone very close to what I recall, but enlarging to 100% showed that the camera was capturing FAR more detail (at a distance) than my eye ever could.

Bolstered by the very pleasing ISO 3200 shot, I am very interested to see what ISO 6400+ can bring to the table. I had only a few minutes to shoot with the 5D2 today, and decided to "unlock" ISO 12,800 and 25,600 via the Custom Functions and take some shots. I also switched to sRAW2, after some positive testing of sRAW1 shots at ISO 3200. I made Henry "sit" and took a few...

Initial impressions of 25,600 are not wholly positive. If a strong point-source of light exists in the frame, banding seems to be quite prevalent. Also, in extremely low light, the dark areas of the image don't contain a lot of detail and do contain a lot of noise. Images taken with more light in the scene reveal quite a lot of detail, and light areas of the image look quite amazing, really. But 25,600 looks to be perhaps a step too far in low light, which is what the super high ISOs are for, anyway. I'll do some more testing, but for now, will stick with 12,800 and under. Speaking of 12,800...

I took this shot of Henry sitting in our living room at ISO 12,800, sRAW2, 1/50s @ f4, 70mm on the 24-105L. It was developed in DPP, with +1 contrast, -1 saturation, 3 sharpness, 3 LNR, 10 CNR. It was after dark, and the living room had no light sources turned on... All of the light in the frame is coming from a trio of 13W Compact Flourescent lights in the kitchen; you can clearly see the source of light in the kitchen illuminating Henry's face. It almost looks like I had a strobe set up. The light on his face is quite striking! The 5D2 is really "seeing in the dark" here; it is revealing more light than my eyes were actually taking in.

Also of note is the color; for an extremely high ISO shot, the color is quite accurate. I actually set color saturation to -1 in this image, as it more closely matches reality; although set at zero, it does match the more "yellow" hue the CF lights give off.

Detail is maintained very well in this image. Here is a 100% crop from Henry's nose (this may be too wide for this blog, so you may wish to click for a full shot):

Note the very tip of the nose is in sharp focus, but at f4, the DoF is so shallow that traveling downwards you quickly lose sharp focus. Sharpness can actually be boosted a bit in DPP without making the noise objectionable. I am not a super-sharp freak, and this to me looks like what a typical film image would look like, so I didn't go any further.

I did turn Luminance Noise Reduction down to 3 (from 10), revealing more of the "grain" that a high ISO image "should" have. The grain is quite tightly patterned, and not at all (to my eyes) objectionable, somewhat increasing apparent acutance. You could play with LNR vs sharpness here and probably get a tradeoff that would satisfy anyone. I am VERY excited to see what DxO can do with these files; since DxO applies NR before RAW conversion, the grain will be even tighter... Chrominance noise is fairly prevalent, as is common with Canon's, but DPP does a very good job suppressing this (at the default CNR level of 10) without obliterating everything of interest in the photo. Yes, there's some noise. But really, this is similar to what I got at ISO 1600 in very challenging light out of my Nikon D200. This is astonishing to me. Not that many years ago, I thought ISO 400 speed film was about as "fast" as I'd care to shoot, and now I'm nitpicking an ISO 12,800 image.

There is a little bit of horizontal banding evident, but really I'm not concerned as it is quite minor. This would really not be evident in a 9x12 print, which is about as high as I'd take a shot like this anyway. So, for me, it's moot.

In any case, I am quite pleased with this performance. I will have to do some more investigation, but ISO 12,800 on this camera is quite usable at first glance.

Canon 5D Mark II, Day 1

2008-12-03T06:45:00.001-08:00

Taken at the St. Joseph Cathedral, within 24 hours of opening the box of our new Canon 5D Mark II. ISO 3200, 24-105L lens, handheld @ 1/6 sec, f4.0. RAW developed with Canon DPP, TIFF sent to DxO for Kodachrome film look.

Noise performance is stellar even at ISO 3200. I have turned high-ISO NR "off" in camera, and just use a smidge of NR in DPP on the ISO 3200 files (luminance 1 or 2 depending upon image, chrominance 2 or 3). I am extremely pleased with the high-ISO performance of this sensor; it handily beats our departed Nikon D200. ISO 3200 on this camera "feels" like somewhere around ISO 1000 on the D200; there is noise, but what is there is not at all objectionable, and the photos are quite useable. Also, color depth seems MUCH better at high ISO's; routinely the D200 would "blow" colors at high ISOs if you pushed the RAW at all in development... The 5D2 does not exhibit this trait nearly as much. I am finding the in-camera default saturation to be a tad high for high ISO work, so I occasionally bump it down a notch in DPP before sending to DxO if I want to apply a film look.

I intended to try sRAW in this church, as very brief initial low-light dog shots shows me two things:
1. sRaw is a little less noisy than RAW, maybe usefully so?
2. high-ISO RAW files are HUGE (normal ISO RAW is about 25mb, high iso can be 35+!)
3. there isn't 21 megapixels of detail in an ISO 3200 shot, anyway.

The last point I will investigate some more. If more detail exists in the 21MP RAW than the 10MP sRAW, it may still be worth shooting full RAW files at very high ISOs. I definitely suspect ISO6400+ files will likely be better at sRAW, ISO 3200 may be a tossup.

But I don't live on high ISO alone. So what is the 5D2 like at "normal" ISOs?? Detail at low ISOs is astonishing. I really wasn't prepared for what a 21-megapixel low-ISO image would really look like once I opened it at home; initially I suspected this camera would "equal" the best of 35mm film cameras, but now I feel it is well beyond that level, approaching medium format. Looking at 100% pixel-level detail on my monitor reveals an astonishing ability to reveal the most minute detail, even at a distance. Back when imaging-resource reviewed the Sony A900, Mike Tomkins said:

First, the Sony A900's image quality is truly awe-inspiring. Opening a file in Photoshop and hitting Command + to get to 100% reminds me of that scene in Blade Runner when Deckard uses the Esper photoanalysis machine. If you know the film, you'll remember that he finds tremendous detail in the print photograph from the year 2019: "Enhance 57 to 19. Track 45 left. Stop. Enhance 15 to 23. Give me a hard copy right there." In the end, he's found a reflection in the mirror that gives him an important clue.

This is exactly the feeling I get when opening the Canon 5D Mark II images at 100%. Holy cow. I nearly fell off my chair when I opened the first outdoor ISO 100 shot. I am telling you, the detail in these photos is astonishing; and even better, it looks natural. I will have some examples to show soon.

I am absolutely gorging on B&W with this camera. Somewhat paradoxically, the color performance of this camera left me speechless when I did a few test B&W conversions in DxO (with Tri-X, natch). I don't know if the 14-bit color of the 5D2 is responsible, but the subtle shifts in color tonality seem very natural to me. When rendered as B&W, this leaves the viewer with an image that has all the subtle tones of a fabulous B&W print. I am very excited about the color and B&W performance of this camera. Bear with me for a bit as I gorge on B&W. I do promise to show some color.

There are some negatives to this camera. Like the shutter release. In short: it is awful. I don't remember any previous Canon having this mushy, indistinct shutter, but we have been Nikon'rs for the last few years, so perhaps the distinct click-click of the two-stage Nikon shutter release has me trained? The 5D2's half-press is initiated by a whisper of a movement... Very good, then. But now you want to take the photo and you press... And press... And keep pressing down. Eventually the shutter trips. But it is such a movement that, with gloves on (in 24 degree weather) I actually got some camera shake outdoors. Hmmph. I will have to get used to this seemingly glaringly poor choice of shutter, I suppose. What was Canon thinking? Perhaps in time I will learn to appreciate this? Right now, I find it annoying, counterproductive to capturing the "decisive moment," and apt to cause camera shake. OK, so I'm speaking in hyperbole here: it isn't that bad. But you get the picture. (or not, because you missed it!)

I'm not totally used to the UI yet. Somehow I twirled in -1 1/3 EV, and couldn't figure out for the life of me how to twirl it out. The top LCD read +/- OEV, so I thought I was shooting without EC. But the rear LCD gave it away (dark image where I didn't want one), and upon popping up the Quick Menu on the rear LCD I saw the - EC dialed in. I tried to dial it out with the button on the top of the camera, to no avail. I had to use the joystick to adjust it on the Quick Menu. WTF? OK, Canon, you have completely lost me on this one. I know, I need to read the manual. But its clear there will be some growing pains with respect to UI. Hopefully these will disappear quickly, as it really can get in the way of vision.

For now, I am reveling in the fact that I can comfortably carry an astonishingly smooth, colorful, and detailed almost medium-format-like camera for normal light work, and a camera that practically sees in the dark for low-ISO captures. Even with a F4 lens. Isn't technology wonderful??

Nikon D3x images hitting the web

2008-12-02T12:35:00.000-08:00

Barely a day after the Nikon was officially announced, some images are starting to hit the web. Marcus Bell blogged the D3x and posted some images this morning taken with the D3x and the new Nikkor 50/1.4 AF-S.

Of note, Marcus describes the D3x as having a:

Full frame 24.5 megapixel sensor delivering a whopping 6048px by 4032px file size.

So it is confirmed, this is exactly the same sensor the Sony A900 uses.

The new 16bit expeed processor gives added dynamic range over the Nikon D3, which I have already found to be incredible

As should the A900-derived sensor, which according to DPReview and imaging-resource has some of the highest DR of any current sensor.

The specifications advise it is 5fps

So same spec as the Sony, then, with its Dual-BIONZ processors. Nikon has labeled the D3x to have their "EXPEED" processors, so this gave some hope that high-ISO noise would be improved... But the specs are suggesting it could be the same sensor and pipeline, which could mean little difference beyond better JPEG processing. (It MUST have better JPEG processing than the Sony to satisfy Nikon's current owners!)

Curiously, there is no mention of noise performance, and all sample images were taken at ISO 200. This is quite curious because there was a time almost every single Canon 5D Mark II image posted pre-release was a high-ISO image. One could opine that Nikon doesn't want the high-ISO images out yet (they did prevent release of the RAW files, as did Canon with the 5D2). Or it could simply be Marcus isn't focused on that type of photography, which is fine, of course.

So the big question remains... Is the image quality any better than the A900? More news to come... and fast, apparently.

Canon G10 rebuttal, Canon G9 lens dust, and a teaser...

2008-12-02T06:19:00.001-08:00

Browsing the blogosphere, I found this blog post which is essentially a rebuttal to the DPReview's ridiculous Canon G10 review that I blogged on previously. In this post, Bill Lockhart says:

I was especially shocked by the last sentence which implies that the G10 has “image quality shortfalls.” Perhaps at higher ISOs this may be true, but frankly the images DPReview has displayed for the Panasonic LX3, which they highly recommend, show similar issues at higher ISOs.

Right on! All photographers know that point-and-shoot tiny-sensor cameras fall down at high ISO. Arguing over which P&S camera is "best at ISO 1600" is like arguing over the best way to drink a cheap whiskey. No matter what you do, cheap whiskey tastes... Crappy. So for most photographers, we care about P&S cameras at ISO 400 and under. And for this, it is clear to my eyes that the G10 is the current king.

Bill goes on to say:

Seems to me that Mr. Wan started with a conclusion, then did his review.

I agree, Bill. The "conclusion" was DPReview's blind belief in the "Megapixel Myth." I've been blogging on this lately, and I agree with Bill here. DPReview decided the outcome of the review before ever picking up the Camera; rather like the equally ridiculous Canon 50D review which I blogged here.

Bill also says:

Mr. Wan’s poorly written review, contradictory statements, incorrect statements about prints, and misleading conclusions aren’t the sort of stuff I thought DPReview was about.

Yes, Bill, this is what DPReview has been about for some time now. Phil has determined that Megapixels are Evil, and has set out to Change the World to fit his world view. I was duped by this oft-repeated DPReview mantra, until after some real-world photography (and prints!!) made me realize I was a fool.

DPReview is sadly losing relevance in this changing world. The forums are entertaining, though.

---------------------

In other news...

My Canon G9 has a lens dust problem. I have been seriously considering the purchase of a Canon G10 due to the excellent Live.com rebate program, but I really have no need to keep the G9 if I go down that path (save the DxO support... I hope DxO announces support for the G10 soon!). However, I can't in good concience sell my G9 with dust in the lens, so it's time for it to go back to Canon for some cleaning. The initial warranty case I opened with Canon read "free of charge" even though I'm a nearly a month over my year's warranty coverage. Fingers crossed on that.

---------------------

A quick teaser: Recently, I was able to borrow a charger for my charger-less Fuji F30 (thanks, Tom!). So with a freshly-charged F30, I started thinking: is this thing all it is hyped up to be? It is widely espoused on the Internet that the Fuji F30 is the "compact to measure compacts by." And as DPReview and others have insinuated or outright stated that they'd like to see manufacturers "roll back the megapixels" I am left wondering: are our glasses a little rose-tited here?

Well fear not: I will be testing my F30 vs. my Canon G9 before it goes back to Canon for a cleaning. Let's see if this F30 lives up to its reputation. I loved the F30 when I shot with it, although I never missed it (at all) after buying the G9... I never did a head-to-head, so this could be enlightening. And I'm even going to make some prints.

Canon 5D Mark II 5 Minute Impressions

2008-12-01T19:39:00.000-08:00

Santa's sleigh has arrived! I received a call from Midwest Photo Exchange that our Canon 5D Mark II arrived and was ready for pickup. Excellent! With all the anticipation of a 7 year old, I swung by on the way home from work to pick it up.

Now that I've charged the battery and had a few minutes to sit down with the new camera, I thought I'd offer a few thoughts. I honestly haven't had more than 15 minutes with the camera, and it feels very unfamiliar at this point... But hey, why not.

I like the user interface. This bodes well for a nice photographic relationship, as the UI can really make or break the camera. I really like that Canon snagged a page from Olympus' and Sony's books and allowed major settings to be changed right on the screen. I really got used to this on my Olympus E-420, and like the quick set options.

The camera is quite large and heavy, but feels very balanced in the hand. I will need to get used to the size since it has been some months since I've carried a full-size SLR (my last was my Nikon D200), and my Canon G9 or Olympus E-420 are MUCH smaller. I'll get used to it. FYI, here's a size comparison with the Olympus:

My first shots were taken with completely out-of-the-box settings. I believe Canon set it to Auto ISO, and Program AE. Sadly, my best dog friend (and most excellent model) Jake died earlier this year, and I tried to make my Lab and Chessie sit for some photos. Tucker, our Lab, always looks kinda sad and pathetic when you take his photo. Henry, the Chessie, doesn't cooperate at all. Period. In any case, here's a few samples:

I'm pleased with the initial results. The first two were shot at ISO 3200, and are quite clean. I have done a lot of presupposing of the high-ISO performance of this camera, and now that it is in my hands, shooting in poor compact-flourescent kitchen light, I'm very happy with the results. I need to tweak NR, as the default NR applied by Canon Digital Photo Professional is too high (in luminance NR, at least) for my tastes; the photos aren't as detailed as I'd like, and I don't mind some grain. But generally, noise performance much better than my D200, and the ISO 3200 shots are plenty useable. sRAW1 (10 megapixels) seems slightly better than full RAW, but I need to play with this some more. Certainly, 10MP is plenty for in-home dog snaps.

The shutter is VERY quiet.

Color is nice and muted, but increments up nicely with some saturation adjustments, and I predict will work really well with the DxO film pack. Speaking of which...

DPP: this is the first time I've used DPP. I need to play with DPP a bit to get what I want out of it... I can say that I really REALLY miss DxO already. I will be waiting with baited breath until DxO releases support for the 5D2. I'm not guessing this will happen until next spring at the earliest, though, so I have some learning to do.

Full Frame sensor: I'm so used to sub-frame cameras that I need to step back into the world of film cameras... F4 isn't going to cut it to get dog noses and eyes in focus anymore. I will need to get used to this. One positive thing I noticed immediately, though: the viewfinder image is huge and bright! Very nice, especially after the Olympus E-420 and the Nikon D50, both of which I have used a lot lately.

Nikon D3x Offical: Nikon enters the Megapixel Race

2008-12-01T06:21:00.000-08:00

Last Friday, Engadget reported that Nikon "outed" the D3x a little early in the newest Pro magazine:

Today, we have official verification from Nikon.

This has been a little-kept secret fore awhile now, and ever since Sony's A900 arrived, observers have opined that the 24-ish megapixel sensor would be arriving in Nikon form soon enough. Now that we have confirmation of this, what does this mean?

1. Nikon feel pressured by Canon to "up the megapixel stakes."
2. High-ISO performance will be scrutinized greatly; Sony's JPEG engine does a poor job at high ISO, and Nikon folks will be wanting MUCH better performance after getting used to the fabulous 12-megapixel FX sensors.
3. What does this new OLPF (Optical Low Pass Filter) mean? All AA filters are "OLPF"'s, so this one must be unique vis-a-vis the Sony. It's anyone's guess at this point whether the D3x will have a weaker or a stronger AA filter than the Sony.
4. Nikon's EXPEED better be good at high ISO, or their current position in the market will be threatened.

I also wonder what this means for the current D3. I could see Nikon keeping it around; after all, Canon's long sold the 1D series (less MP, higher frame rate) and the 1Ds series (high MP, lower frame rate) side by side. Nikon could also do this, to satisfy multiple groups of photogs.

What about the D700? I think likely this one will disappear. I can't see Nikon keeping around a D700x AND a D700, as at the "enthusiast" end of the market that these cameras are marketed to most folks will choose high-MP over high frame rate. So sadly, I bet the fantastic low-light D700 will fade into oblivion.

We shall see, though. I think this makes for interesting times, and am glad Nikon is moving down the path of more megapixels and Fx sensors.

UPDATE: we don't have our Canon 5D II in hand yet. The position of Santa's sleigh relative to our local camera shop is unknown at this time; reports say we may know by EOB today what the timing will be. So although we didn't get to play with it over the long Holiday weekend, it is coming soon enough.

Canon G10 Review @ DPReview

2008-11-25T07:15:00.000-08:00

DPReview posted their review of the Canon Powershot G10 this morning. It was a fairly predictable review, with a smattering of derision about the increase in megapixels mixed in with some comparison shots. As is typical, DPReview did not do comaprisons shooting RAW; which, to me, is the point of a camera like the G10. At least this saves us from criticizing their choice of RAW converter :)

This page shows the G10 vs. its rivals, the Panasonic LX3 and the Nikon Coolpix P6000. The Nikon annoys me, as it has in other comparisons; at base ISO, the output is oversharpened and there's annoying JPEG artifacts. What inquiring minds really want to know, though, is "does the Panasonic LX3 take better photos than the G10?" and this series of test photos seems to indicate that the G10 produces photos with better detail and more natural color than the LX3. This finding closely correlates with what I see from the Imaging-Resource test photos; here is the G10 ISO 80 studio shot, and here is the LX3 shot. Open them both in a new window and flip back and forth; it is clear the G10 outresolves the LX3.

DPReview also sinks to a new low in posting an entire page diatribe about how the G10 at base ISO is actually worse than the LX3 at base ISO, and why this backs up their continual push against the rise of megapixels. I find a few problems with this:
1. They didn't shoot RAW. Yes, we know Canon applies NR to base ISO JPEGs. No one shoots JPEG (save Ken Rockwell) with a G10, so shoot RAW and then compare.
2. Exposure was not matched between the two cameras. The G10 was slightly overexposed.
3. Optimal aperture was not used for the G10. A little nitpicky, but as we saw in a prior blog post, f4 is no longer optimal on this small sensor. I am not sure about the LX3 (f4 was used there, as well).
4. Their findings don't gel with their own studio test shots: the G10 should be resolving more detail than the LX3. As mentioned above, the shot used to illustrate DPReview's "findings" is flawed.
5. Their findings don't gel with shots found at imaging-resource (LX3 here and G10 here) and cameralabs. Imaging-Resource test photos clearly show the G10 far superior in detail retrieval, and the Cameralabs shot shows the LX3 is significantly behind the G9 in terms of real-world detail retrieval.
6. Open the full DPReview test images for yourself and compare; it is obvious Don Wan cherry-picked portions of the image to support his "findings." To my eyes, the G10 has significantly more detail (not to mention more accurate color) in many portions of the image.
7. Don says:

Here in on the brickwork of Tower Bridge again you can see the LX3 producing more fine detail than the G10.

7 (cont). Open the images yourself and compare--the G10 is significantly ahead here. I don't know what Don is looking at... Except maybe a double-recompressed JPEG??

DPReview, in pushing their flawed agenda, comes up with another doozy of a stinker. You can tell that their delivery was a little softened in the G10 review; the internet furor over the ridiculous Canon 50D Review "conclusions" must be making the Editorial staff a little gunshy on strong "conclusions." However, their anti-megapixel bias comes through, and obviously they worked hard to support their "Megapixel Myth."

That said, there are some excellent portions of this review. DPReview included, for the first time that I can recall, a comparison between a Point and Shoot and a DSLR. This is appropriate for many users, and is something Cameralabs has been doing for awhile. I like the idea. On following pages, it is very clear why DSLR's are the choice for high-ISO shooting. The G10 suffers at ISO 400, and suffers greatly at ISO1600. Of course, the Panasonic LX3 and all small-sensored compacts would "fail" this test, but DPReview doesn't tell you that :)

In trying to keep ahead of the megapixel race, Canon has produced a camera that in the real world can't deliver on the promise of the styling and control layout. In the studio it produced some incredible results at base ISO, but out in the real world and as ISO settings increased, the loss of fine detail and increase in noise really let it down. A camera is ultimately about taking pictures, and that is why we put so much emphasis on the image quality output.

I call BS. If DPReview was really concerned about image quality, they would 1) shoot RAW, 2) investigate the performance of RAW converters for each camera and not stick to "one level playing field" which isn't level, and 3) stop clouding their vision with the flawed perception that more megapixels==worse images. Open your eyes, DPReview! Even your test images prove the G10 to have better image quality than the LX3.

They are the only site that I know of that has crowned the LX3 king over the G10; all other reputable sites have clearly shown that the G10 resolves more detail at base ISO, the color is better, the UI is more fluid, and the lens is far more flexible. The LX3 certainly moves ahead in poorer lighting conditions; if you need to shoot in low-ish light, and absolutely cannot carry a DSLR, then the LX3 is obviously your best choice. It is a flawed choice, as DPReview makes clear in the G10 review, but it is the best choice.

For all other point-and-shoot uses; travel photography in good light, and carrying a small camera to a sporting event... The Canon G seems to remain the "king."

Canon 5D Mark II Ships Tomorrow

2008-11-24T12:50:00.000-08:00

Rob Galbraith posted today that shipments of the Canon 5D Mark II begin tomorrow, November 25, 2008. Canon has historically shipped overnight to dealers, so this could mean the camera will be hitting the streets just in time for the Thanksgiving Holiday. Bad family photos will immediately commence :)

Hopefully, our dealer will be ringing us on Wednesday...

DxO Mark: Exposing the Fallacy of the Megapixel Myth

2008-11-21T06:02:00.000-08:00

High ISO D200 (equiv. ISO 4800)

Since the launch of the DxO Mark site, I have been reading and contemplating the data presented. Not only is the volume of data high, but philosophy behind what the data means, measuring the "quality" of the RAW image at the sensor, causes a paradigm shift in one's consideration of camera performance. Until now, we have all judged camera output by looking at images. Indeed, this seems like the logical thing to do; after all, cameras make images. However, DxO is essentially proposing something new: to consider the performance of the camera at the sensor level, without taking into account the final image. If I understand their philosophy correctly, DxO believes better images come from better sensors. Indeed, this makes sense when you consider it; although MUCH more than this goes into what your eye sees as a good image, if you consider it from a RAW Converter's perspective: if a RAW converter is given a high quality RAW image, it will be easier to create a high quality finished image.

Ponder that for a bit. We all get so focused on resolution, noise charts, etc... But DxO is saying: why not measure this at the RAW level, BEFORE image processing is applied? Then we can know what the sensor is really capable of.

So what does this mean? From my initial research through the DxO Mark site, it does mean that some of the current "internet wisdom" might be suspect; specifically, perhaps the Sony A900 isn't as bad at High ISO as we all think (and we'd have their JPEG engine to thank for that). DxO Mark also clearly shows us that the "Megapixel Myth" in itself is a myth. I've blogged on this in the past here and here, but DxO has some measurements which might even help convince Phil Askey that he's wrong. Here's DxO's measurements of Signal-to-Noise Ratio (SNR) at the RAW level, over the past 6 years of camera technology:

This graph clearly shows that at the RAW level, sensors are noisier today (lower SNR) than they were 6 years ago. Taken at face value, this seems to completely validate Phil's claims. However, when you take into account increasing resolution, things change. As resolution increases, the perceived "size" of noise decreases. DxO graphs the perceived SNR when taking into account resolution here:

What is clear by this graph is that although SNR is decreasing, when you take into account the increase in resolution, we will perceive camera noise has in fact decreased.

I do believe Phil and the Megapixel Pundits indeed believe themselves and their "Megapixel Myth." Phil was quoted in his New York Times interview as stating:

And you’re cramming so many pixels in such a small sensor that noise is becoming a real issue.

I believe that he believes this. I used to believe this. But a funny thing happened. I stopped looking at 100% pixel-level images on my screen and started printing them. When I first started using my Nikon D200, I felt it was "noisier" than my Canon 300D. And, at the pixel level, it was. DxO Mark makes this very clear from this graph:

The first thing you notice about the graph is that the 300D is more "sensitive" than the D200. At marked ISOs, the 300D is about 1/3 stop more sensitive than the D200. This confuses real-world testing a bit (and is why testing two cameras at the "same" ISO on-camera can confuse matters), but the graph makes it easy to see what a truly comparable ISO SNR result is for each camera. Look at the vertical ISO1600 line, for instance. The Canon has about 2-2.5dB better SNR. Remember, this is at 100% pixel level, viewing on a computer. So the D200 is clearly "noisier."

But look what happens when you normalize the data to the same printed size:

As I worked more and more with my D200, I noticed that in physical prints, I really wasn't as much bothered by noise at ISO1600. The noise character was a little different than my old 300D, Nikon preferring to leave luminance "grain" and clean up chroma vs. the Canon's more "clean" output; but in the end, I shot at ISO1600 without reservation. DxO Mark makes it clear why: SNR on the two cameras was essentially identical when normalized to the same print size. And the positive of more megapixels is more resolution; so the image will also carry with it more detail and nuance.

There's a lot to dig into with this new DxO Mark site, but one thing is clear to me: the data seems quite analagous with my perception of how past cameras have behaved. It's a gold mine!

Sony A900 better for art photgraphy than the Nikon D3?

2008-11-20T08:32:00.000-08:00

I stumbled upon this "mini review" of the A900 at The Online Photographer, and it has me pondering the possibility that "I was wrong" about the Sony A900.. Although I felt very negative about the Sony A900 after reading the DPReview test, I have still followed some photographers online who are using the A900. Sometimes, first impressions can be wrong, and all that.

Back in October, I blogged about the A900 reviews which first hit the web. The thing that struck me (like a 2x4 upside the head) was the apparently awful high-ISO performance. I'm used to dealing with "high ISO fear" on my Canon G9, and one og the HUGE advantages of moving to full-frame digital is the excellent high-ISO performance of the large sensor cameras, so at first pass, I thought: why spend $3000 on a full-frame camera that I need to tread lightly with, like my Canon G9?

Well, Mike Johnston believes that the A900 he has used might just be better for "art photography" than the Nikon D3. Essentially, he feels the D3 is a better tool... But the A900 produces prettier photographs. So if you have the time (art), maybe the A900 is worth a look.

I can say that some A900 photos I've seen on the web are very intriguing. The color and contrast response, especially, is interesting to me. I noticed the Zeiss optics on my Nikons to be particularly nice with a very interesting color and contrast response, but this Sony is much better. HMMN.

Canon G10 wins over a pundit

2008-11-18T09:04:00.000-08:00

Ken Tanaka of The Online Photographer posted a fairly negative response to Edward Taylor's initial G10 review on TOP. For the most part, Ken's thoughts seemed to echo my concerns about the upgrade to the G10.

However, Ken has changed his tune, and is now whistling that he "was wrong" about the G10:

Image Quality: Despite Canon cramming more pixels on that tiny CCD (which Canon claims is all-new) the G10's image quality is actually markedly better than the G9's. Raw images are noticeably sharper right out of the camera. Highlights seem less prone to blowing out. Colors are more lively. Noise is well sublimated, more so than the G9.

and

Operation: The G10 is certainly no slower than the G9. Its shot-to-shot cycle time (my personal bugaboo) may even be just a bit zippier. Moving the exposure compensation control from a screen-based control to a knurled mechanical dial was actually an excellent design change. I've already begun to wonder how I ever lived without it.

He goes on to say that it is a bit bigger than the G9, but feels better built. I'm not sure if I could fit a G10 in my jeans pocket... Where it resided through our entire Wyoming vacation. But this is not the first G9 owner who has been won over by the G10, so it is something to ponder. I'm still not sure where I shake out...

Canon 5D Mark II: "the best camera Canon has ever built"

2008-11-18T07:56:00.000-08:00

Perusing the blogsphere, I came upon this entry from Richard Bui. Apparently, he met with Jim Rose from Canon (Senior Professional Market Specialist for Field Market Support, what a ridiculously long title) to discuss the new 5D Mark II. Jim's feeling?

This is the best camera Canon has ever built.

And that it will...

re-energize Canon

Who has been a bit battered by Nikon as of late. High ISO performance, a potential question mark with the pixel-packed sensor, was discussed:

Jim Rose was discussing how low noise is at ISO 6400 and when the staff photographers at San Francisco Chronicle (they have two on order BTW) saw the images at ISO 6400, their jaws dropped. Jim was discussing how even George Lepp (of the famous Lepp Institute of Digital Imaging) was very impressed and said that ISO 6400 on the 5D Mark II is very usable. That’s a very nice seal of approval. At ISO 12,800 things start getting a bit noisy, but not terribly bad. By ISO 25,600, things are quite noticeable noisy, but if this is the only way to get a shot, it is still usable. According to Jim (not Canon’s official stance), ISO 3200 on the 5D Mark II is comparable to ISO 800 on the 5D.

When will we see the 5D Mark II? It seems like information is solidifying, and that the cameras may actually be on the ground now on the way to retailers. Canonrumors says here that:

It looks like allocations have been set for the 5D Mark II for at least a few retailers. Expect shipments to begin next week.

I’m told most mom & pop stores should get the allocation they asked for.

Canonrumors prior post here seems to have been verified on various net forums, so this is probably as close as we'll get to a solid date. Note that they discuss Canada... No word on USA delivery, but it may be the same. I have also heard late this week... Friday seems possible. I will hopefully have some test shots soon.

DxO Mark: a new, free camera image quality database

2008-11-17T09:28:00.000-08:00

As regular readers will know, I have become a fan of DxO Optics Pro. It has often become my RAW converter and workflow product of choice due to excellent RAW conversion quality, automatic distortion correction, an excellent "DxO Lighting" module, and the FilmPack "film looks" which I love. I have also started using DxO Noise in some situations, and DxO Lens Softness for my D200/17-55 combo where it really excels over straight USM.
I was so impressed with the RAW conversion quality for my Canon G9, and the new improvements in DxO Noise in version 5.3 for my Nikon D200, that I just last week upgraded from 4.x to 5.3. I also upgraded from the standard product to the Elite version, in preparation for our (hopefully) forthcoming Canon 5D Mark II.

This morning I found this post on the imaging resource site which indicated DxO Labs has introduced a new, free camera image quality database called DxO Mark. Further, Dave Etchells' posted his thoughts here. He said,

As part of the testing needed to create their excellent DxO Optics Pro software, DxO Labs collects extraordinarily in-depth data on a wide range of digital SLR cameras. It occurred to them that - with a little additional analysis, graphing, and some effort by a web designer - the data they were collecting anyway could be made available to the photo community as a whole. The result is the new DxOMark website, just unveiled today (November 17, 2008).

The DxO Mark website is very cool, indeed. There you can view things such as true base ISO, Signal-to-Noise Ratio, Dynamic Range, Tonal Range, and Color Sensitivity. Some of this data is provided both on a 1-to-1 pixel level, and referenced to a physical print! It is going to take me a bit of time to digest what all this data really means, and present my view on it; but it is clear that we, as of today, have an unprecedented tool at our hands to more objectively measure digital camera sensor performance. DxO Mark also has comparison tools to help an end user determine if an upgrade is "worth it" both in terms of 1-to-1 and print performance. This is absolutely astonishing, made even better since it is FREE!

Here's an example of what the site can give you, provided by Dave Etchells:

EDIT: I should also point out that the Canon 50D, Nikon D90, and Canon G10 are already in the DxO Mark database, but not in DxO Optics Pro. Obviously, DxO Labs has analyzed these cameras already; so I have high hopes that DxO Modules should be available for these cameras soon (especially the G10, since it has an integrated lens). (No Canon 5D Mark II, unfortunately!)

Panasonic G1 Micro Four Thirds review on Imaging Resource

2008-11-15T06:45:00.000-08:00

Yesterday, Imaging Resource posted their review of the world's first Micro Four Thirds camera, the Panasonic DMC-G1. I was not at all excited when the G1 was announced; to me, it is a waste of the promise of Micro Four Thirds. Yes, it's a little smaller than the world's smallest DSLR (which I own) the Olympus E-420... But not small enough to make a whit of difference in a pocket. And you need a pretty big pocket to carry these things. Now what made me very excited about Micro Four Thirds was the Olympus mockup:

Sometimes, its annoying to schlep around a DSLR. They're big, bulky, and very obtrusive. When I travel, I usually carry one, anyway; they are intuitive to use, extremely flexible, and produce great photographs. I carried my D200 all over Europe, and although it was sometimes a bit of a bother, was rewarded with many wonderful photographs. But sometimes you want to "travel light." Mostly, when I feel like traveling unencumbered, I take along the Canon G9, as it fits just fine in my pocket. But there are compromises you have to live with when carrying a small-sensor point and shoot, as I have blogged previously. Micro Four Thirds, in eliminating the traditional SLR mirror box, promises a small(ish) camera with far less compromises. The Olympus above is what I'd really like to see develop in this arena as I find that my Olympus E-420 is really only a useful tool with its "pancake" lens installed. With the kit zoom (which is small, for sure, as is the G1's) I can't fit it in a pocket and I might as well carry a larger DSLR.

So because the Olympus is still at least 9 months off, and the Panasonic G1 isn't that exciting, I haven't blogged much about Micro Four Thirds (yet). But this review bears some discussion, as the image quality of the G1 is probably a good indication of what is coming down the pike.

Compared to its closest competition, the Panasonic G1 comes in with better image quality than the Olympus E-420, but slightly less than the Rebel XSi...I was very concerned that it might have the E-420's odd tone curve, the only major drawback we saw to that camera. According to our Imatest results, the Panasonic G1's dynamic range performance is quite a bit better, and the tone curve looks more natural than the E-420's, with more detail in the shadows and better performance in the mids.

This is probably a good thing for the average consumer. I have actually really enjoyed the sensor "feel" and tone curve of the E-420. I love black and white imagery, and the B&W look I get off a RAW converted E-420 image with DxO's Tri-X applied is, to me, beautiful:

The tone curve and "Olympus Color" really lends itself to a Kodachrome interpretation, also, which was one of my favorite color films ever:

I don't have a lot of E-420 images up, I promise more soon, but I digress... The tone curve of the Olympus E-420 is sharp, and prone to blowing out highlights and blocking up shadows. Certainly, you can get beautiful results... But the average consumer is probably better off with a Canon XSi or Nikon D60, as the results will be very good with less effort. Imatest results show the Dynamic Range to be about equal to my old Nikon D200 at about 7-8.5 useable stops, which is good enough albeit behind the newest APS-C sub-frame sensors.

If I'm truly honest with myself, although I do love the E-420's punchy tone curve and quirky "feel" ... It's not exactly the camera I'd want to be dragging around Europe, as I'd have to really work it to get outstanding images in the varied light one encounters while traveling. The Dynamic Range is not great, and the photographer really has to work the light to get good results in all circumstances. For me, the E-420 remains a bit of a "fun" camera, and it's nice to see Panasonic moving a bit more towards the norm here.

The Panasonic G1 does pretty well at ISO 1,600 considering the smaller sensor. While its color is slightly more washed out, the G1 retains a decent amount of detail. The Canon Rebel XSi, for its part, has greater contrast, a little more detail, and greater color saturation, but it retains more chroma (color) noise... The Panasonic G1's sensor has more pixels, so it naturally captures more detail than the Olympus E-420 at ISO 1,600, but key is that it holds onto that detail despite the anti-noise processing necessary with its smaller pixels.

This is the bane of the Olympus Four Thirds cameras. Test shots on Imaging Resource and other sites, in static conditions, show the E-420 is "behind" other current cameras in high-ISO performance. Real-world shots show me that it's well behind. I really don't like shooting it in poor light over ISO 400; even ISO 800 degrades noticeably with forced NR. ISO 1600 is very well behind the XSi and its rivals. It is very nice to see Panasonic improving in this area, all the while moving to a more tightly-packed 12-megapixel sensor. You can check out Dave's test shots at all ISOs on this page.

I like the fact that imaging resource actually prints and views photos from test cameras, instead of just 100% pixel-peeping. This is where the rubber meets the road for me, as I like printed photos and not just staring at a computer screen to view my photographs. On this page, Dave says:

20x30 inch prints would be entirely acceptable for display at any normal viewing distance...ISO 800 shots made really great looking 13x19 inch prints...At ISO 1,600, we felt that 8.5x11 was really as large as most people would like to print.

That's really quite good. 20x30 is a huge print for most people, and it is nice to see the 12-megapixel G1 doing well here. If you grab that stellar vacation landscape you want printed to a very large size for wall display, you can do it with this camera. The ISO 800 result is also very encouraging; this means the G1 is holding detail well, and not simply smoothing it over with NR as the E-420 would. 13x19 is still a very large print for most folks! It looks like by ISO 1600 the party's over... So essentially you get 1-2 stops more from this camera than a small-sensor point and shoot. That's useful and bodes very well for the Micro Four Thirds future.

I'm excited about the future... How can you not be, with something as small and flexible as the Olympus mockup below??

Pixels are like cupcakes" according to the New York Times

2008-11-14T11:13:00.001-08:00

Yesterday, the New York Times printed an article titled "Pixels are Like Cupcakes. Let Me Explain." Apparently Phil Askey, of DPReview, is now an accepted "expert" on the "Megapixel Myth" and deemed a credible source by the New York Times, as he is quoted as saying:

Once you get beyond seven or eight megapixels in a compact point-and-shoot camera, the small lenses are struggling to keep up

Really, Phil? That doesn't seem to be the case with the Canon G10, which is resolving ~2600 lpph (shooting RAW) according to Cameralabs' tests here, which is 15% more than the Canon Digital Rebel XSi (itself a much larger-sensored DSLR). The G10 has 17% more resolution (on paper) than the XSi, so its clear nearly all that "paper resolution" is "real resolution" in actual photographs.

Clearly, some other photographers don't agree with Phil Askey, either. As I've blogged before, Michael Reichmann from the Luminous Landscape site posted this (only slightly) tongue-in-cheek G10 vs Medium Format article.

Additionally, Phil says:

And you’re cramming so many pixels in such a small sensor that noise is becoming a real issue.

It can be, but that's not always the case. My (now departed) 10-megapixel Nikon D200 displays more noise at ISO 1600 than the 12-megapixel Canon Rebel XSi at ISO 1600; and the Canon has a smaller sensor. Technology CAN overcome the "noise limitation."

And it's not all in the sensor... But the RAW converter, also. As long as DPReview continues to use Adobe Camera Raw as a RAW converter to test "sensor noise" Phil won't really know how good images can be. The link I posted earlier today to the DxO interview is illuminating; it is not explicitly stated in the interview, but you can see with your own eyes that Adobe Camera Raw literally multiplies noise in the D700 files that DxO uses; check out the wormy-looking blue channel noise that DxO and Capture NX don't magnify. If you start with less (and finer) noise in an image, you won't need as much noise reduction, and the final image will be "better."

It's not all about the RAW converter, though. DxO is no magic on my G9's ISO 1600 files... Which clearly are past the limits of current technology, and fit very well into Phil's "Megapixel Myth" box.

But it's not that clear, folks. If I maintained the "Megapixel Myth" mantra, I wouldn't have a Canon G9, I'd still be using my "big pixel" Fuji F30. News flash, Phil and Mr. Juskalian of the New York Times: my Canon G9 produces images that are far superior to my Fuji F30, all the way through ISO 400. ISO 800 and 1600 were better on the Fuji, but that's like saying "this cigarette is better for you than that one because it has less tar;" all the point-and-shoot images will kill you at high ISOs. Similarly, my eyes tell me the new Canon G10 with is packed sensor has superior image quality at all ISOs save 800 and up to the new "sensible" Panasonic LX3.

Isn't it about the image quality, in the end? Sensor technology is improving, RAW converters are improving, and more megapixels gives us more choices (both in cropping, and downrezzing which also ameloriates noise)... So doesn't it make sense to open your mind and not blindly stick to the long-held belief that "the earth is flat" and "pixels are like cupcakes?"

DxO Optics Pro 5.3 Noise Reduction

2008-11-14T04:52:00.000-08:00

Back when Photokina 2008 was in full swing, I watched this video interview with Cyrille de La Chesnais on imaging resource. Although Cyrille touted DxO's new Noise Reduction strategy in v5.3, I passed that over and got VERY excited about the Canon G9 support. Regular readers of my blog have already had a glimpse of what DxO can do for the G9's images, but I haven't talked about DxO Noise Reduction at all.

That's because, as a rule, I don't use noise reduction. I don't like it.

In the past, noise reduction, to me, has indeed scrubbed the noise... But also scrubs the detail. I don't mind luminance noise too much, as I grew up with film, and film had plenty of "noise." In fact, the grain of certain films actually helped boost visual acuity of a photo. Luminance noise in the digital world can have a similar effect.

The real problem is chroma noise. The blotchy color-like noise is quite objectionable, and noise reduction algorithms typically scrub the heck out of chroma noise because of that fact. The problem is that it affects so many pixels on your image that detail is scrubbed out, and images often look "plastic" and "processed."

I tried DxO Noise Reduction on my G9 images and to tell you the truth: I was not impressed. ISO 1600 images were still unuseable, and ISO 400-800 images were scrubbed of the fine detail that makes the G9 so neat. So I just went on thinking it was another failed attempt at NR.

That is, until I started working with my Nikon D200 files from our Europe trip. As always, I start off after a DxO import by turning off DxO NR on all images. I happened upon a night image I took, whereupon I completely mucked up the exposure. It was at least 2.5 stops underexposed at ISO 800. I recalled the portion of the video where Cyrille explained that he shot the D700 not at the highest ISO, but at a lower ISO and pushed 3 stops. His results were remarkable, and suspect. I thought I'd give it a whirl. The results were astonishing! The D200 shot at an equivalent ISO 4500 was quite good. DxO Noise in v5.3 really does leave a lot of detail, obliterate nasty chroma blotches without affecting neighboring pixels (and obliterating detail), and leave a very small, film-like grain. I really never liked my D200 at ISO 1600; to see a shot processed 1.5 stops above that look so good astonished me.

So maybe there's something to this "noise reduction before RAW conversion," then. If you haven't already, check out the video at this link. Here's what DxO has to say about it. I'll blog more on DxO Noise eventually, with some examples, after I get a little more experience with it.

Travel Photography with the Canon G9, Part III--What Mode?

2008-11-13T05:16:00.001-08:00

Part III in my series on travel photography with the Canon G9 will conclude my blogging on the G9 for a short time, most likely. I have finally worked through my backlog of G9 photos, and now it's time to address that elephant standing in the corner of my room: my photos from our Europe trip, in Spring of 2007. Yep, 2007. I took a bazillion photos, and its time to get them off the hard drive and onto my wall and into the websphere. It'll take awhile, and I need to get it done before (fingers crossed!) our Canon 5D Mark II arrives.

I will be blogging on DxO and the G9 in the near future. DxO is now, by far, my favorite RAW converter/developer for the G9... But there are some tips and caveats I can share for those condsidering it. Hopefully I'll work that in soon, although I will most likely be blogging mostly on news in the coming days.

So... when traveling with your G9, what mode do you use? "Green Box Auto?" Program? Aperture Priority? Shutter Priority? Scene Modes?

Well, you should know by now, NO to the Scene Mode spot on the dial. It's not that scene modes are bad in and of themselves; they can certainly help folks who don't understand how to set a camera for good photos. But you can't shoot RAW in Scene Modes on the Canon G9, and that's a dealbreaker. (I think the G10 may allow this?)

In any case, there's really only one mode I use on the G9, and that's Aperture Priority mode (Av). I know, I know... With the small sensor, the G9's depth of field is severely limited. I won't blog on for ages about the reason why this is, but if you aren't familiar with the curse of small-sensor cameras, here's a pertinent quote from the Wikipedia article:

...a point-and-shoot digital camera with a 1/1.8″ sensor (7.18 mm × 5.32 mm) at a normal focal length and f/2.8 has the same DOF as a 35 mm camera with a normal lens at f/13.

When I was a kid, I shot with full-manual or partially-automated film cameras. I quickly became enamored with Aperture Priority mode on my father's SLR, and have used that mode as a default on every camera I could since. Back in the day, and still today on DSLR's, I use Av to decide the amount of Depth of Field; often, I simply float between two apertures on a camera. The first is a very wide aperture I know will bring with it very small DoF, and the second is a small aperture for when I want everything in focus. Experience and testing will tell me what those apertures are with each camera/lens combo.

But if a small-sensor Point and Shoot digital camera effectively cannot "do" the very small DoF I am used to, and your control over DoF is solely limited to how close you photograph your subject (the closer you are, the more blurred the background can be) and what focal length you choose (longer == less DoF), why is Aperture Priority mode useful at all? Because of one very important factor: diffraction. Click this link to see what a blogger has to say about diffraction on the Canon G9. The essential point: what you know abotu 35mm film photography will ruin your G9 photos. Back in the film days, I'd shoot a minimum of f8, and sometimes up to f16 or so, to get max DoF for landscape shots. Unfortunately, if you do this with your G9, your images will be hopelessly blurry due to the effect of diffraction on a small sensor.

You can see an excellent representation of this in the Canon G10 Review from Photo Review Australia. They posted a most excellent chart showing resolution vs. aperture:

What you're looking at is aperture on the X axis, and resolution on the Y axis. Higher on the Y axis is "better." As you can clearly see, stopping the G10's lens down at all starts to limit resolution. So while you will get more DoF, you will also get a blurrier image. And since the small-sensor point and shoot cameras have so much DoF anyway...

Put your G9/G10 in Aperture Priority Mode, choose f4, and leave it there!

Just forget about it, and shoot away. You can choose any aperture you feel is best on your camera; on my G9, f4 is fairly optimal. On the G10, maybe f3.5 would be a better choice. f4 will give you excellent DoF, lots of resolution, and a decent shutter speed in most travel photographs. If you "lock in" the Aperture like this in Av mode, you'll prevent the G9 from deciding what aperture to use--which will likely be closer to f8 in bright sunlight, and you'll be losing a lot of resolution! That G10 at f8 is barely resolving more than a 10 megapixel camera, and the image would be SOFT SOFT SOFT.

OK, so if you need a shorter shutter speed, open the lens up... No worries. Also, since the G9 has a variable-aperture lens, the aperture will get smaller as you zoom. But the G9 is smart enough to open that back up when you zoom back to wide, so you don't have to ride the control all day.

So that's why I almost always use Av on my small-sensor cameras. Even when I'm trying to "blur motion" a bit, I will often keep the cam in Av mode and simply turn on the Neutral Density filter if needed. I despise letting this camera choose the aperture... And you should, too.

Before the G9 Travel photos are lost in a sea of Nikon photos from Europe, here's some links.
1. G9: Our Trip to Vermont
2. G9: Family Vacation to Wyoming
3. G9: Weekend trip to The Wilds

Canon 5D Mark II hands-on blogged by David Ziser

2008-11-12T11:39:00.000-08:00

Surfing the blogdom I found a post on David Ziser's blog describing his whirlwind experiences with the new Canon 5D Mark II. Real news has been hard to come by lately; there was a fury of data emerging just around the announcement from Canon, and now the calm before the storm (release date) has photographers around the world gnashing their teeth. Maybe the storm is getting close, though, as it seems more and more pros are getting their hands on the 5D2, even for a short time.

David was mostly testing high-ISO available light photography. He said:

Everybody knows I'm a speed junkie so for me the big question is how much can I "goose" that ISO and still capture a good picture? So how far could I go with the 5D Mk II? OK, folks - here it is - how about 3 stops higher than 800 ISO!!! Let me re-phrase, as Law and Order Jack McCoy would say, at 6400 ISO, I'd say I see the same noise characteristics as I see currently see on my 40D at 800 ISO.

Nice. That was JPEG (w/Digic 4 NR, obviously), staight out of the camera. Sounds like ISO 6400 will be quite useable, which is maybe even a bit better than I had hoped.

I continued to shoot some available light images at the Bar Mitzvah celebration and found that I could still get extremely usable results at, are you ready, at 12,800 ISO!!!

The "useable results" were with some post-processing NR, but that is also quite encouraging. I think with some downsampling the images could be great, as I surmised earlier here and here.

My main lens was Canon's 24-105mm IS lens - yep, you heard me right - even with the smaller aperture, the super fast ISO still let me get the shot - AMAZING!!!.

Nice. Glad to hear the 24-105L seems to be up to the task--it's such a cool focal range (with IS!).

OK, keep in mind, I'm using the 24-105mm IS for most of the images I was shooting. I found the 5D Mark II did a fine job finding and locking focus - MUCH better that the first 5D.

So maybe the pundits on DPReview saying the "5D Mark II Autofocus SUCKS!!!" are wrong? This is the second pro photog to comment that, hands-on, the 5D2 works just fine for weddings. Jeff Ascough blogged about this earlier, although he specifically said he preferred centre point only.

I'm getting very jazzed about getting my hands on a 5D2... If Santa comes through! :)

And speaking of Jeff Ascough... His most recent series (with a 1DSMK3) is astoundingly good.

And speaking of Wedding photography... I really like Tim Larsen's most recent engagement session. Absolutely fantastic! It really tells a story about the character and heart of his clients.

Canon G10 Review Posted at Cameralabs

2008-11-12T04:50:00.000-08:00

Cameralabs just posted its Canon G10 Review. As all Cameralabs reviews, the empirical data is somewhat thinner than Imaging Resource or DPReview, but what is there is interesting.

I won't comment much on the review's discussion of design and features, as I've covered that before. But the photo of the new exposure compensation knob is an excellent one. As I mentioned a few days ago, using Exposure Compensation is quite important with the small-sensor Canon G9, and all small Point and Shoot cameras; so it's nice to see the physical dial added to the G10:

The Cameralabs review won't be the one to do heavy Canon G9 vs. G10 comaprisons; Gordon chooses to compare current-gen cameras in almost all circumstances. Although this can be disappointing, he often tests P&S cameras vs. low-end DSLR's; he does so in the G10 review (comparing it vs. the Canon Rebel XSi). Other review sites don't mention comparisons like this much more than in passing, feeling that most folks will carry a G9/G10 when they don't want to carry their bulky DSLR, or just don't care as much about image quality. This is probably quite true, in many cases; but I'm interested in understanding "what am I really giving up by carrying the P&S?" so a direct comparison is quite helpful. Additionally, I'm sure there are folks who have about G10 money to spend, and can either buy a low-end DSLR, or a G10... And don't know which way to go. For those folks, the Cameralabs testing philosophy is especially helpful.

I've long believed that lenses make as much of a difference as the camera itself. You're often better off buying a lower-end camera with a high-end lens than the opposite. For years, I shot the original Canon Digital Rebel (300D) with Canon L lenses, and got fantastic results. Since the sensors are smaller on point and shoot cameras (see my prior blog entry here for a visual), manufacturers can make their lenses on those cameras smaller. Generally, smaller == cheaper, and its easier to make a "better" small lens than one for a large-sensor camera. What does this mean? Canon can put a very good lens in the G10, and still sell it for $500. Gordon compared it to the Canon Rebel XSi with its "kit" lens on this page. At base ISO, with the kit lens, I think its hard to argue for the XSi. The G10 simply outresolves the XSi (the SD880IS does well in this comparo, too).

What does this mean? Well, if you're not prepared to spend some $$$ on a better lens for the XSi, you're better off (at base ISO) with a Canon G10. That's remarkable, and is a feat that the Canon G9 also pulls off, IMO. It's the same thing Michael Reichmann was talking about here, and I blogged here. It also means that, at base ISO, you don't have to feel too badly about taking your G10 vs. your DSLR. If you want to be less encumbered on that vacation, go for it. You'll still get poster-printable images if you do your part.

But it's not all roses. Due to the aforementioned small sensor, image quality degrades very quickly as you raise the ISO. Check out Gordon's High ISO Noise page here for an example. My comments on the G10's noise mirror the G9. Stick to ISO 80 if at all possible... 200 will reduce your post processing latitude, 400 even further, anything above and you might as well use the nasty flash. You can also see very clearly why a DSLR might just fight for the mythical $500... Their much larger sensors allow for far cleaner high-ISO shots. If you like available light photography, the DSLR is your ticket.

The Gallery pages highlight some of the G10's strengths and weaknesses well. Here's how the G9 "saw" the boat at its widest angle:

And here's the G10:

The G10's wider 28mm lens is hugely useful, and this pair of photos really demonstrates what that can mean. However, what Canon giveth, Canon also taketh. At the long end of 140mm, the G10 can't blur a background well even at max aperture:

What do I think, after reading Gordon's review? Well, nothing much new was revealed. It's hard to compare the G9 and G10 with his images, as the test images are not as "controlled" as those on other sites. But I like his practical review methodology from a photography standpoint, and for that, this G10 review seems to indicate the G10 will be a great choice of a camera for anyone interested in DSLR-like (or better, in some cases) image quality that can truly fit in your pocket. No real surprises vs. the G9 in this review. It looks a little better in some areas, and a little worse in others. I'm still undecided as to whether to replace my G9 with a G10. If you're still on the fence, or just considering a G10, Gordon has a nice review summary/verdict here which might help you make up your mind.

Subaru STI Fire Sale

2008-11-11T04:55:00.000-08:00

If anyone's in the market for a slightly cantankerous, ridiculously fast way to transport family and dogs, look no further than a Subaru STI. Right now, probably due to the abysmal economy, you can buy a Subaru STI for $7500 off MSRP with 0% financing for 63 months. That's as insane as the STI itself.

I'm sticking with my '07 STI, but I know a lot of folks buying these '08s. If you want one, go to your local dealer NOW and be ready to buy--one buyer has told me stock is disappearing by the minute. The deal's good through the end of the year, but you won't get one if you wait.

Both Buyers' dealers in Columbus are participating; I'm not sure about Hatfield et. al.

Have fun. It's the 4-season sports car :)

Travel Photography with the Canon G9--Part II, the importance of base ISO

2008-11-10T06:02:00.000-08:00

All current point-and-shoot cameras save the odd-duckling Sigma DP1 currently have very small digital sensors. Above you can see what this means; the Canon G9's sensor is roughly the size of the "Coolpix" camera above. While larger than some other point-and-shoot sensors (the "Xi" size is roughly the same size as most P&S sensors), it's still MUCH smaller than a sub-frame APS-C DSLR or full-frame DSLR. What this means, essentially, is the sensor is receiving far less light when looking at the same scene as a DSLR. Even at base ISO, most P&S cameras have "noise" or speckly grain, due to the fact the manufacturer has to essentially turn up the gain in order to compensate for less light.

Obviously, for the least amount of noise, you want to shoot at "base ISO" (ISO 80) with your Canon G9, whenever possible. But noise isn't the only reason; as you crank up the ISO, not only will the noise increase, but the color fidelity will decrease. In practice, this isn't much of a problem, and even ISO 400 images can come out very colorful and nice, like this one:

I was able to make this one look quite like what my eye saw using DxO. I can't remember the exact settings, but I believe I used a little highlight recovery, slight DxO lighting, and applied the velvia film look. I believe I had to reduce contrast a bit, and stop DxO lighting from preserving shadows lest the trees fall too deeply into shadow. The end result was fairly pleasing, although I feel the grassy area is too high key and detracts severely from the composition I "saw" in my head.

The reason for this is the fact I shot it at ISO 400. I've found, when working with the G9 RAW files, that your latitude for enhancement in post-processing decreases a little at ISO 100, a lot at ISO 200, and very very much at ISO 400. Some ISO 400 images, especially those shot in very high-contrast environments, just can't be "saved" especially if you didn't protect the highlights. The scene above wasn't high in contrast, but at ISO 400 the G9's sensor really was pushed to its limits.

When you shoot at base ISO, which on the G9 is ISO 80, you have a lot more latitude in post-processing. Colors don't "blow" so quickly, you can subtly bring down the highlights and bring up the shadows to match more what the eye saw at the time of capture, and overall fidelity is much higher. Of course, noise is also lower, which helps if you decide to make very large prints. Here's an example of an image shot at base ISO that really "felt good" when I was post-processing it; DxO was easily able to get the image I imagined in my mind's eye, without a lot of fiddling:

To summarize:
Shoot ISO 80 if at all possible. ISO 100 really doesn't give you much more speed, so I use it quite infrequently. Most of the time, I'll just capture 2x at ISO 80 and trust the Image Stabilization to steady the shot. So likely, if you can't get the shutter speed you need, you'll be bumping to ISO 200. At this point, you will really want to be shooting RAW, as the JPEGs are very bad at ISO 200. RAW ISO 200 files are usually pretty useable and even still "feel" a little like film in post processing. Bumping to ISO 400 is much more risky. Usually, your images will still be quite useable and enjoyable; but the "feel" in post-processing is like a delicate and not well-exposed piece of slide film... You just don't have a lot of latitude.

Skip ISO 800 and 1600 altogether, IMO. Look for something to rest the camera on, shoot multiples, and hope you get a clear shot.

Travel Photography with the Canon G9 -- Part I, Exposure Compensation

2008-11-08T17:37:00.000-08:00

Now that I've taken a few trips with only the Canon G9 along, I've really started to get comfortable with how the camera will react in a variety of circumstances. I decided to start a series of tips about how to get the most from your G9 when traveling, instead of typing one hugely long braindump of all I've learned. I'm thinking it should be more a stream of consciousness than an organized effort or review, so if you don't agree, drop me a line.

Part I: Exposure Compensation. This is hugely important with the G9! When I first started using the G9, I assumed that I could trust its meter like my DSLR's. Nope, no can do. Every digital camera reacts differently; once you use a particular camera long enough, you start to get a feel for how its "film" (sensor) reacts. Most DSLR's nowadays can get away with no Exposure Compensation; to maximize the dynamic range of the sensor, it makes sense to tweak sometimes, or use the spot meter... But, in general, if you shoot RAW (and you should), you can get away with Matrix metering on a Nikon without any undue trouble. This is nice when traveling; if you're like me, you like to capture photos as your eye sees them, and fiddling with metering, histograms, etc is distracting to the joy of travel.

With the G9, though, you will find the sensor to react much like slide film. Exposure latitude and dynamic range is not wide, and the meter tends to allow the highlights to blow readily. I didn't realize this when I took the camera on our Wyoming trip this year; I noticed a bunch of blown highlights from our Vermont trip, but I thought this was due to shooting JPEG (I'll discuss why in another blog entry)... Nope. I didn't realize this until I got home. Here's an example:

The wide dynamic range posed by this scene is tough for a digital camera, ANY digital camera. The deep shadows in the pines on the left, "middle gray" in steamy haze in the center of the frame, and the bright sky above... That's a lot of f-stops of dynamic range. Exposed at +/- 0 EV, the G9 allows the sky to "blow" and properly expose the middle range. This isn't a terrible idea, but if you shoot RAW and at base ISO, you can afford to expose the sky better and bring the shadows up later. When I went to The Wilds, I shot all outdoor high-DR images at -2/3 EV. Later, when processing the RAW files in DxO, I found that choosing "Slight" for highlight recovery and "Slight" for DxO lighting protected the sky, and brought up the middle tones to proper exposure. An example of this is here:

Notice how the sky looks to be a nice deep blue, as your eye would see it in "real life."

You could use the histogram overlay feature of the G9 and adjust the Exposure Compensation of each image... And I'll do this when really trying to nail a Landscape, for instance. But for general travel photography, if you're outside in bright sunlight, simply dial in -2/3 EV and be done with it. Just remember to dial back to +/- 0 EV if you walk inside a building :)

Nikon D700 or the high-megapixel Canon 5D Mark II, and what lens?

2008-11-07T07:24:00.000-08:00

Not much seemed to be hitting the web last night/this morning in the photography world, so I went hunting. I found this blog entry on why one photographer has chosen the Nikon D700 over the Canon 5D Mark II. He has some very good thoughts about why one might want the Nikon over the new Canon.

I especially liked his discussion of megapixels--do you need them? This is something I have been pondering as of late. Working with my Canon G9 at ISO 80, shooting RAW, I feel sharp per-pixel output I'm getting is pretty close to the compact film cameras I've had in the past. I also think it's pretty close to the bare-bones film SLR's my wife and I have owned over the years, with their so-so kit lenses. Digital is getting very good, and the G9's RAW files have enough in them to allow me some push/pull latitude. I kinda feel like I'm darkrooming again. I started wondering, what would truly approximate the really good film output I've handled over the years? Like a fine-grained color or B&W film with a sharp prime lens on my (sadly STOLEN) Minolta XD-11 or Nikkormat?

I think closely approximating really good film would require the move from sub-frame APS-C DSLR's to a full-frame DSLR. The larger sensor allows for larger photosites which both improves high-ISO performance, as well as lens performance; if you take DPReview's suggestion that the pixel pitch of the new Canon 50D perhaps starting to push the boundaries of what current lenses can resolve, full-frame sensors have some time before they hit that same pixel pitch (the Canon 50D is 4.5 mp/cm2 and the 5D Mark II is 2.4 mp/cm2), so current lenses should have a better chance of showing actual resolution bumps from higher-megapixel sensors.

But where is this line? Where would a sensor start to be seriously let down by current lenses? The article above makes a good argument for 20 megapixels being "the line." If you read down into the article, the author discusses his feelings on MTF vs a 20 megapixel sensor:

With a 20M sensor, you need a lens that has an MTF50 with 70 lp/mm or an equivalent of 1700 lp/picture height for full frame. This is already surpassing the optimal resolution (center of the frame with a stopped down aperture) of even the best prime lenses.

I don't entirely agree with the latter statement, as there are quite a few lenses that exceed 1700 lp/picture height on APS-C; they should be even better in the center on full-frame, although worse on the edges, but current lenses should out-resolve the 1700 figure (see here, here, and here for some examples. For reference, here's a much poorer-performing lens). But I do agree it makes lens choice more difficult. Obviously, tossing a sub-par lens on a 20+ megapixel Canon 5D II is a bad idea; you might as well have a cheaper camera. But what is "good enough?"

With the original 5D, many folks found the 24-105 F4L (bundled as a kit or sold separately) to be an excellent compliment to the 5D. Resolution and contrast seemed good at almost all focal lengths, with a slight degradation at the tele end. Wide open performance was excellent. The IS was useful, and good. And the focal range made it the perfect all-around high quality lens. Bokeh wasn't the best, but not awful in most circumstances.

Will this be so for the 5D II? Canon is again bundling the 24-105 F4L in the 5D Mark II "kit," and it's likely what we'll end up with, so I'll have some serious thoughts on this in the coming months. My conjecture, at this point, is that the 24-105L will do the 21-megapixel sensor justice. It is a very high (for a zoom) MTF lens, with unusually high wide-open performance. Canon's MTF chart is here:

And for tele, here:

If you're not familiar with MTF charts, see this article on the Luminous Landscape site for a good intro.

Photozone tested the 24-105L which was linked above (here if you missed it). It looks to me like the 24-105L does very well in the center and border at all apertures from 24-70, and falls down a bit at 105mm when wide open. On full-frame the border numbers should drop (as the sensor will be using more of the image circle than on APS-C). SLRGear.com, which publishes absolutely excellent lens tests, doesn't publish lp/picture height numbers, instead using "Blur Index" numbers which are not directly transferable. But you can see in their 24-105L review here that the corners will suffer somewhat. Their copy of the lens visibly degraded at 70mm, but everything up to that point looks great, wide open or stopped down.

The only other data point I am aware of is the Cameralabs Sony A900 test. In that test, a Canon 1Ds Mark III image shot with the 24-105L snuck in on the outdoor resolution page here... Performance looks quite good compared to the Zeiss and Nikkor optics, and it doesn't look to be letting down the 1DS III too badly.

Since the 5D Mark II should show similar (if not the same) sensor performance as the 1Ds Mark III, and taking into account all the data points above, I think the 24-105 F4L lens will be a good partner for the 5D II... Albeit I do think that the sensor will be pushing it to its limits, and the lens may hold the camera back at the tele end and certainly in the corners at some apertures. I don't think it will be quite as limiting as the 18-200 appears to be on the Canon 50D (see the DPReview lens test for a good visual example of this), but its clear the megapixel race is probably pushing the boundaries of current technology here.

And more to the point: I think we're very close to the mythical "film replacement" DSLR. I didn't feel the Nikon D200 was quite there; perhaps the D700 is much closer. But I think the Canon 5D Mark II is closer still... Which is why we're (not so patiently) waiting for it.

Phil Askey's final word on the 50D review and RAW converters

2008-11-06T11:25:00.000-08:00

Since DPReview isn't allowing comments on their "blog" some hot conversations have been going on in their forums. Phil stepped in and made a "final" statement:

The hilarious thing for me is I designed the tests, I set the criteria, I provided the examples for all these review formats. We use ACR today as a consistent platform with which to compare cameras *from various manufacturers* (hence why saying 'oh use DPP' is pointless), that was a decision I took many years ago and nothing (no amount of whining) makes me believe that decision needs to be changed.

All this "Phil should do it" rubbish is just that, my team work to the same standards I do and I proof read these reviews, I was involved with Lars in comparing and re-testing the box-shots with the various lenses and I edited the conclusion. We as a team decided on the final rating and the overall mood of the conclusion text. (These reviews may carry one or two people's names as the primary writers but they are all collaborative efforts).

I have nothing further to say on this subject or this review.

Well, I guess that wraps it up. It's too bad Phil's so closed minded about the subject, but it reflects that which has infested DPReview as their success has grown. If you like their methodology, great... If not, stuff it. The forums are full of trolls and pixel-peepers, and the tests have not kept pace with technology.

And if you don't like the DPReview methodology, I suggest you visit some of the other excellent sites on the web. Try Imaging Resource to add some pepper with your salt. An upcoming site with promise is Cameralabs. If you're looking for more about compacts or P&S cameras check out the Digital Camera Resource Page.

DPReview is still a great site, and has good information on it, so I suggest readers continue to visit. Just put your flameproofs on and wear some supicious sunglasses as you read.

DPReview responds to ACR Fiasco

2008-11-06T08:43:00.000-08:00

DPReview's new blog had an entry added yesterday from Lars Rehm to explain why DPR uses Adobe Camera Raw for all RAW conversions.

Firstly we aim to create a level playing field and to make results - as much as possible - comparable between cameras. For this reason we simply have to use a third party converter that can handle most of the many different RAW formats that are out there rather than a proprietary converter.

Here's the problem with that: it's well known that Adobe "tweaks" ACR (sometimes frequently) to improve RAW conversions as time goes on, because initial results usually aren't that great. Here's an old Fred Miranda post which details some challenges which ACR has had in the past, and triumphs as time goes on (even beating some proprietary manufacturer converters). The same thing still occurs today. It makes sense; Adobe certainly has access to Canon/Nikon/etc RAW formats well before public release, but not as much or as extensive access as the camera makers themselves.

So while ACR may eventually be similar to, or even much better than, Canon's DPP... It's not now, and that's clear. I, as well as most, really appreciate DPReview getting their reviews out ASAP after a camer'a release. But statements like:

in terms of detail the 50D is not quite the step up from the 40D that we would have expected... There is only a very small amount of extra detail in the 50D output ... it appears that Canon has reached the limit of what is sensible, in terms of megapixels, on an APS-C sensor (using current technology)... Even the sharpest primes at optimal apertures cannot (at least on the edges of the frame) satisfy the 15.1 megapixel sensor's hunger for resolution. The result is images that look comparatively soft at a pixel level and only show marginally more detail than images from a good ten or twelve megapixel DSLR...

...are quite definitive, and to make such statements, I think DPReview should reconsider their "ACR for all" stance. Personally, I think a "level playing field" will never be achieveable, unless all camera makers move to one RAW format, color space, and sensor. So, to make the best of what we've got, doesn't it make sense to use the camera maker's RAW converter, if it does a better job than ACR?

Canon 50D DPP vs ACR

2008-11-05T06:24:00.000-08:00

As posted earlier, I took a bit of issue with DPReview's findings that the Canon 50D wasn't really resolving much more detail than the 10mp 40D. It just didn't jive with samples I've seen at other sites. However, DPReview correctly pointed out that maybe we're getting close to max pixel-count for APS-C sub-frame sensors, at least with current lens technology.

I surmised that maybe the Canon 50/1.4, not their best design, was maybe the limiting factor. Well, I missed something obvious. The RAW converter! DPReview used Adobe Camera Raw to convert the RAW images (initially a beta version, then replaced that with a shipped version), and it appears to be doing a far worse job than Canon's own DPP RAW conversion tool. This unscientific but illuminating sample on DPReview from a forum member seems to suggest maybe the RAW conversion tool is holding back the 50D images.

There's a lot to think about, and balance, when choosing your digital workflow. The converter is yet another piece of the puzzle, and can radically affect how images are perceived after processing. With the Canon G9, for instance, using Canon's own RAW conversion tools is IMO the worst way to process your G9 images. Noise Reduction is applied even at base ISO whether you like it or not, and the fantastic fine detail the G9 is capable of resolving just isn't pulled out of the RAW files by Canon's software (hopefully the G10, now supported by DPP, will do better with Canon's own software). RawTherapee is a fantastic free tool for seeing what your G9 images can do, Adobe ACR is good, and DxO is proving quite good as well. And there are others.

DPReview can't test them all, to be certain. But I do think when testing a brand-new camera, at the cutting edge of technology, they should have actually tested Canon's shipped RAW conversion tool vs. their own chosen tool. Perhaps it would have changed the result a bit? I don't know, but I did notice this morning that DPReview's statement that there was "only .4EV of headroom in the RAW files" was gone, now that they used non-beta ACR software. So there's some suggestion their review might have been a little hasty in some conclusions.

Hey, it's clear that all-in-one zooms like the new Canon 18-200 aren't up to the 15 megapixel APS-C sensor of the 50D. But perhaps some of Canon's lenses are. And perhaps you need to use DPP, for now, to get to the newfound resolution.

Canon G10 image sneaks into LX3 Review

2008-11-05T05:29:00.001-08:00

DPReview posted their Panasonic LX3 review yesterday afternoon. Keen readers will note Canon G10 sample images snuck into the mix here (at ISO 80), here (ISO 400), and here (ISO 1600).

Besides the obvious comparison to the LX3 (in which the G10, to my eyes, proves superior at base ISO, about the same at ISO 400, and behind at ISO 1600), one can compare the G10 test shot (click here) vs. the G9's shot (click here) at DPReview.

I believe my earlier thoughts as pulled from the shots on Imaging Resource still ring true; essentially the G10 does seem to offer more resolution than the G9 (check out the viewfinder on the camera and the bailey's label for instance), and the G9 may have slightly better per-pixel sharpness. It is evident to me that the G10 will surpass the G9 over most of the frame at base ISO. Enough to merit the upgrade? I'm not sure, I'm still pondering that. It's a little better, yes. Not an astounding difference.

It's also clear that the LX3 is well below both the G9 and G10 at base ISO, about the same at ISO 400, and well ahead at ISO 1600. I still don't think ISO 1600 is all that great, though, certainly nowhere near what a DSLR can do. So the LX3 is a little bit of a failed beauty in my eyes; it's setup to be the best low light camera in a class of cameras which aren't that great at low light... Although the lens looks great, I'm glad I didn't preorder the LX3. The vast majority of the shots I took at The Wilds, for instance, have better IQ with the G9 than I could have gotten with the LX3. All were taken at base ISO. Of course, I would have had a wider lens, which would have been nice for shots like this:

But the poorer base-ISO image quality is a big turn off.

I wonder why DPReview doesn't test RAW vs. RAW on these cameras, as they do with DSLR's. I wish they would. It's all I shoot, unless I have to...

Canon 50D Review at Imaging Resource

2008-11-04T13:22:00.001-08:00

Imaging Resource has posted a thorough review of the Canon 50D. Their findings are a little different from DPReview; believing the 50D to be a better (resolution, high ISO performance) than the 40D. Check out the review here. Most interesting to me is that imaging resource actually prints photos to view; supposedly, at even very large print sizes, the 50D does very well.

Since the end result of all high-megapixel images should be print (on-screen viewing necessitates only a megapixel or so), I think the IR reviews are quite interesting. See the bottom of this page to read their thoughts on image quality, ISO, and print sizes.

Dynamic range as measured by IMATEST is a bit of a tossup, with slightly better "high quality" numbers but perhaps slightly lower total DR. A good showing, I think. What do readers think?

Abandoned Oil Tank in Kodachrome (Canon G9)

2008-11-03T09:14:00.001-08:00

IMG_1363_print_raw, originally uploaded by ndjedinak.

I am very much enjoying post-processing the G9 RAW files with DxO. I captured this one over the weekend, on a trail maintenance day with my Land Rover club (Green Oval Guild). Shot at -2/3 EV in Av at f4.0, and post-processed with the Kodachome film look, medium+ contrast, DxO Geometry Correction, no LensSoftness applied, slight highlight recovery and DxO Lighting set to slight. A touch of USM and non DxO Noise applied.

I found most of the files from the day, when processed in this manner, responded wonderfully. I applied the Astia film look to most, as I wasn't looking for the Kodachrome punch on all of the photos... But the ones that I do want it on turned out fantastic, IMO.

Hard to believe you can get this from a little point and shoot.

You can see more shots at my photostream or in this Flickr Set. The photos were taken at The Wilds, the largest animal conservation site in the US. Through the graciousness of The Wilds, our 4wd club has been allowed to maintain a series of 4wd trails on property. It's a gorgeous place.

Canon 50D Review @ DPReview, and some thoughts on the "Megapixel Race"

2008-10-30T06:19:00.001-07:00

DPReview has posted its new Canon 50D review here. I have just a few comments, mostly centered around the following quote:

Having said that, in terms of detail the 50D is not quite the step up from the 40D that we would have expected. After all the new model's nominal resolution has increased by approximately 22% in both dimensions. There is only a very small amount of extra detail in the 50D output though (in fact even at 100%, if you scale the 40D's output up to match the 50D the results are almost indistinguishable). While the new sensor makes the 50D the highest megapixel APS-C DSLR currently on the market it also makes it the one with the highest pixel density and it appears that Canon has reached the limit of what is sensible, in terms of megapixels, on an APS-C sensor (using current technology). At a pixel density of 4.5 MP/cm² (40D: 3.1 MP/cm², 1Ds MkIII: 2.4 MP/cm²) the lens becomes the limiting factor. Even the sharpest primes at optimal apertures cannot (at least on the edges of the frame) satisfy the 15.1 megapixel sensor's hunger for resolution. The result is images that look comparatively soft at a pixel level and only show marginally more detail than images from a good ten or twelve megapixel DSLR. If all you end up with is a larger image (and file) one starts to wonder what the whole point of pushing the resolution up to these dizzying heights is.

Finally let's take a look a the raw output of the EOS50D next to the ten megapixel 40D. Removing any in-camera noise reduction and processing the images using Adobe Camera Raw (without NR) gives us the nearest thing to a 'level playing field' for assessing the relative noise levels of the two cameras' sensors. Despite the fact that the 50D is the newer camera it shows visibly more chroma and luminance noise than the 40D. Considering the 50D's much more tightly packed sensor (4.5 MP/cm² vs 3.1 MP/cm² on the 40D) this comes hardly as a surprise. It would have been unreasonable to expect Canon's engineers to overcome the laws of physics.

Taken at face value, with their sample crops, I tend to agree. Its obvious that the increase from 10 megapixels (40D) to 15 megapixels (50D) hasn't delivered a big jump in resolution. And the noise performance does appear to be a little worse than the 40D. But, things aren't as simple as DPReview makes them out, and I think they've missed the boat a bit here.

Megapixels: DPReview has been a huge detractor to the rapid increase in megapixels for many years. They felt that the rapid rise in megapixels was completely marketing-driven. Up until recently, I would have completely agreed; it just didn't seem like the huge jumps in megapixels were netting us anything but more noise, especially on small sensor P&S cameras. However, the Canon G9 has completely changed my viewpoint on this. The G9 packs 12 megapixels onto a sensor far smaller than the APS-C DSLR's mentioned above... And the image quality is fantastic. For reference, the G9 has 28 mp/cm2. The 50D has 4.5 mp/cm2. The 50D would have 93 megapixels if Canon filled the sensor at the same density as the G9's.

Yes, on the G9, diffraction sets in early. Yes, it's noisy at high-ish ISOs (and even at base 80 there's some noise). But the resolution is there, and the files at base ISO, shot RAW, are absolutely astonishing (if you "develop" them outside of Canon's software). It is clear to me that more resolution was gained by moving to the 12 megapixel sensor in the G9.

So what gives? Why doesn't the new 50D trump the other APS-C cameras in resolution? I'm not totally certain, but my guess is "lenses." Sure, the Canon 50/1.4 is sharp... But couldn't Canon do better with a newer design? If Canon can put a lens onto the $500 Canon G9 that delivers the resolution that sensor is capable of, then Canon can certainly build a lens that delivers what the 50D is capable of (ok, so the small lenses on the G9 are easier/cheaper to design and build... but it's possible). We just haven't seen that yet. I think DPreview is blind to that simple fact.

Noise: Again, I think Dpreview has missed the boat. Yes, the more tightly-packed sensor shows more noise. But the 50D has more resolution; so at the same print sizes, those noise speckles will appear smaller than the 40d's. This often means that, when printed, noise appears to be less in the higher-megapixel image. Would this be the case? Frankly, we don't know, as DPReview didn't see fit to consider that.

Additionally, Canon is providing the sRAW file option. sRAW on the 50D gives you a 7.1 megapixel file. Preliminary shots I have seen on the web have shown that noise decreases substantially in the sRAW files. From what I have seen, the noise decreases to the point where it's nearly 2 stops better than the 40D. So would ISO 6400 sRAW equal the 40D's noise at ISO 1600? Maybe. We won't know, because DPReview doesn't test this, even though many on the forums have requested it.

Megapixel Race: From my thoughts above, it's clear that the "answer" to the "Megapixel Race" isn't clear at all. Technology has a way of changing long-held perceptions over time, sometimes using the brute force of Moore's Law. In this case, I'm ready to start opening my mind to the possibility that we may be seeing some real gains with these higher megapixel sensors... I have certainly witnessed this with the tightly-packed G9.

Canon 5D Mark II (1DS Mark III) vs Nikon D700 Part II

2008-10-30T06:06:00.001-07:00

I didn't mention base ISO files yesterday (oops).

So I essentially find the stand-in for the 5D II to be equal to the D700 at ISO 3200. Great. For me, the high ISO performance of the 5D II should be great.

What about base ISO? This is hugely important, as the preponderance of my photos are outdoors and at or near base ISO. To me, this is where the Canon's trump the Nikons; and not by a small margin. Look at the base iso outdoor test image for the Canon here vs the Nikon here. And the static test shot for the Canon here and the Nikon here. Download them, load them into your favorite tabbed browser, and flip back and forth.

To me, the Canon's base ISO images pratically glow with life. The fine, smooth detail rendered just looks less digital to my eyes. I do think we're reaching the point where megapixel increases aren't delivering a 1-for-1 increase in image quality (DPReview is suggesting the new Canon 50D actually drops back compared to its predeccesor because of "too many megapixels", see the bottom of this page), but to my eyes the Canon's images soundly trump the Nikon's. The Canon 1DsIII/5D II has 75% more pixels packed onto its sensor than the D3/D700, and the images certainly aren't 75% better, but they are visibly "better" to my eyes.

So there you have it. I think the Canons have moved ahead of the Nikons. It's nice to see, as Nikon really did throw Canon for a loop with their new FX sensors. And more competition is better for consumers :)

Canon 5D Mark II (1DsM3) vs Nikon D700, more high ISO thoughts

2008-10-29T09:56:00.000-07:00

So like everyone else I've been chomping at the bit to get a hold of some good 5D Mark II info. The 1600-6400 shots that I mentioned a few days ago have been pretty impressive; at 6400 it is possible things start falling apart, but the detail retained should allow for some pretty good options in post-processing.

Imaging Resource has a great library of test shots that I like to look at to see what a camera is capable of as the ISOs rise. In particular, I really like the "Still Life" series that is always done; it is pretty comparable (for the most part) between cameras. Since it is pretty well documented that the Canon 5D Mark II sensor is VERY similar to the 1Ds Mark III, I decided to play with some Canon 1Ds Mark III shots to see what I got. And why not compare them with the current low-light champion, the Nikon? So I downloaded the 1Ds Mark III and D700 profiles for my favorite RAW editor, DxO, and set to work.

The JPEGs and RAW-to-JPEG conversions can be directly compared on Imaging Resource's excellent site, so I won't repeat that here. What I did was more like my "real" workflow: downloaded the RAW files, loaded them into DxO, equalized the color between the cameras, sharpened, allowed DxO Noise to work its magic, and had a look. In the end, I had to boost the Canon's exposure by .4 EV to match the Nikon, so that would boost noise for the Canon a bit. But that's the breaks. I did have to fiddle with the sharpening a bit to get the files to look similar, the Nikon needing a bit more sharpening than the Canon. In the end, the two files did indeed look similar, so I went with that. My thought is this: the workflow above is exactly what I would do with both cameras as the ISO rises. With the Canon, which is natively more noisy than the Nikon, I would eventually want to downres the image to around 10-12 megapixels to lessen the impact of noise.

I made two sets of test files and cropped bits of them to compare: one set was ISO 3200, with the Canon down-rezzed to 12mp to equal the Nikon, and the other at base ISO (100 for Canon, 200 for Nikon) at original resolutions. First, the High-ISO series:

In that image, you have in the left column, the Canon. In the right, the Nikon.

Top Row: Canon downres with Paint on the left, Nikon straight out of DxO on right. Equalized JPEG compression.

Bottom Row: Canon downres with Dxo on left, Nikon straight out of DxO on right, equalized JPEG compression.

My thoughts: there's very little in it at ISO 3200. Obviously, I need a better downres program than Paint or DxO; both had their weak points here. Paint brought down some of the noise speckles that downres should get rid of. DxO did a better job there, but left jaggies. So if I end up with a Canon I'll need to find a better way to downres the images keeping the detail and losing the noise.

But in general, I can make some assumptions from this test:

1. If the 5D II sensor proves to be the same as or very similar to the 1Ds Mark III, which it should, ISO 3200 between the Canon and Nikon will be very very similar when post processed to the same resolution.
2. Do I need to downres 3200? No, probably not. I have uploaded a full-res DxO-processed ISO 3200 sample for the Canon here. It looks great to me.
3. If I do need to downres to minimize noise, I should be able to "toss away" some noise by resampling the Canon 5D II images to 12mp. Done properly, this does look to reduce noise.
4. DxO leaves in more luminance noise in the Canon ISO 3200 shots. This is not unpleasant, and is very fine-grained. It should print well. Here is a full-res ISO 3200 DxO'd test shot to show what I mean:

The unknown? How will the 5D II sensor react at 6400 and above. There's some scuttlebutt on DPReview that suggests Canon is "faking" ISO 6400+... Which could mean a rapid loss of quality and tonality as compared to the Nikon. All digital sensors seem to "fall apart" at some ISO; on my Canon G9, this happens at ISO 1600 which is utterly hopeless. The new G10 holds up pretty well at ISO 1600 from what I see. I hope to see the 5D II's output still quite useable at 12,800. If it falls apart there, though, the Nikons will still "win" the low light wars; downres or not.

We shall see!

Canon G10 vs. G9

2008-10-28T05:22:00.000-07:00

As I've mentioned in previous blogs, I have really enjoyed owning my Canon G9. It's the Point-and-Shoot that "acts like a real camera." Using the G9 is mostly an organic experience for a photographer, which is great because P&S cameras don't always have an interface that gets out of the way of making good photographs. I really don't want to be thinking about the technical aspects of my cameras, I just want to capture photos as if they are second nature. The G9 isn't quite the organic experience of using a good DSLR, but it's close enough.

I have been a Canon G-series fan since I bought the original G1, and I've wondered if I should upgrade my G9 to the new G10 now that it's hit the streets. To make matters more intriguing, the recent Live.com cashback offers (the amount goes up and down, currently it is 25%) and the instant 10% off ebay coupon that many people (including myself) received which is good through October 31, I could get a G10 for a little over $300. I should be able to sell my G9 for more than that, so should I get paid to try a G10? I have been trying to figure this one out for awhile now, and since some good information and test images have hit the web recently one can start to form opinions without owning both cameras.

For me, the pros of the G10 are:
1. Wideangle 28mm. This would be more useful than 35mm of the G9 for landscape, indoor, and creative shots)
2. More megapixels allow for more cropping choices. I'm not a huge cropper, preferring to "get things right" the first time in-camera, but sometimes you simply can't do that due to environmental or other factors. More megapixels give you more cropping choices. HOWEVER, the 28-140mm lens of the G10 is nowehere near as "long" as the 35-210mm of the G10, so some of the cropping benefit is lost as you can't optically get as close. I have cropped on some G9 images to "simulate" a longer lens when I couldn't get any closer to my subject, especially when shooting sports, so this would be a negative. See this photo on my Flickr stream for an example of this technique. Here's where I was sitting!
3. Digic 4. Normally, I could care less about the processor in a camera, as I always prefer to shoot RAW over only JPEG for post-exposure processing latitude and control over noise reduction. Especially on P&S cameras, noise reduction on-chip applied to JPEGs can really eat up the detail in your image, and you often don't have any control over the amount of noise reduction applied like you would with a DSLR. However, I do shoot JPEG with the G9 in a few circumstances: sports and very low light long-lens work. For sports, I shoot JPEG because even though the G9 does pretty well shooting RAW images, it is much more responsive when shooting JPEG. I'll blog on this in a future post, but suffice to say I find RAW worthless for sports shooting. For low-light long-lens work, such as sitting in the pew of a friend's wedding, I shoot JPEG. The reason is: I won't use anything higher than ISO400 on the G9 due to noise, and shutter speeds drop very low to the realm of extreme difficulty in hand-holding, and very little ability to "stop motion." If you shoot continuous burst mode, out of 3 shots, usually at least 1 will be useable. You can get shots like this with this method, 1/5 second handheld at 210mm! That's 6 stops below what you "should" be able to hold, made possible by IS and continuous drive. So... what can Digic 4 do here? Well, my preliminary investigation of web shots shows the noise reduction on-chip is better with Digic 4, so when I'm forced to shoot JPEG, the results should be better.
4. Slightly better high-ISO photos. ISO 400 and 800 seem very close, but ISO 1600 is better on the G10. I never use 800/1600 on the G9 because the files just don't clean up well, especially the ISO 1600 shots which are awful. I have applied noise reduction to some of the ISO 1600 test shots on the web, and the G10 files clean up remarkably well. Useable? Not sure, more investigation is needed.
5. Added Exposure Compensation dial. Not a huge deal, as pushing a button and twiddling the rear dial is an easy habit to get into, but since EC is often needed in high-contrast situations this could make it more intuitive. I'm always a fan of dials over menus.

Cons of the G10:
1. 140mm max telephoto. I'm not usually a long-lens guy, but as mentioned previously some of the stuff I use my G9 for calls for as much telephoto as possible, and then some. This would be a negative. The increase in megapixels should somewhat make up for this, although I'm unclear as to how it would shake out in the end.
2. Slower burst performance. CNet's Review didn't specifically address JPEG shooting, but RAW performance in continuous drive mode has dropped nearly 20%. This would drop frames per second in JPEG from ~3fps to ~2.5fps, which is really significant for the stuff I use continuous drive for. This may simply be a RAW issue, though, as the rest of the performance numbers look good. Simply a risk.
3. It's bigger. I do pocket my G9 (it fits great in a jeans pocket), and the G10 is bigger, and lumpier. It would be harder to carry, which is the entire purpose of the P&S for me, as I'm leaving the DSLR at home.
4. DxO doesn't support it yet. Just a little spoiler here: although my initial response to DxO and the G9 was sorta lukewarm... I have changed my mind. It's awesome! I suspect DxO will eventually support the G10, but it may be awhile, so in the meantime the G9's files will look better and (more crucially, for me) workflow speed is much greater. DxO has been slow (glacial, really) with P&S cameras, so I'm not sure how much of a risk this is. I have emailed them.

Image Quality's the thing, though, and does the G10 put out in that regard? My feelings are mixed. Imaging Resource has their most excellent test shots up, and at ISO 80, to my eyes, there's not much in it. See the G9 photo here, and the G10 here. Same story with the outdoor ISO 80 shots. G9 here, G10 here. The G10 does have a little more detail due to having a megapixel advantage, but it's slight, and to me the images look softer. Perhaps the G10 has a slightly stronger anti-aliasing filter in front of the sensor; at least, that's what it looks like to me. Both cameras look good at ISO 80, which hopefully is the most-used ISO on your G-series camera... There's just not much in it, really. I'm not thinking this is a reason to upgrade.

At higher ISOs, things do change in the G10's favor. Using the Imaging Resource shots again, ISO 400 is where things begin to sway towards the G10. The G9 images lose a little more detail, and are a little less saturated, with the G10 being a little better with both. ISO 800 the G10's a little better again. At ISO 1600, the G9 totally falls apart, with a blotchy, difficult-to-remove noise pattern that is less like grain and more like impressionism. The G10 holds together much better, as mentioned above. And the trump card may be more megapixels: as I've mentioned previously with the Canon 5D Mark II, we're getting such high-megapixel sensors that we can "afford" to throw away megapixels in post-processing (or sRaw in camera like the 5D II, which I don't think the G10 supports) in the name of less noise. Perhaps a G10's ISO 1600 downrezzed to 10MP image could equal the G9 at ISO 800? Or ISO 400 downrezzed to 7MP? 7MP is still enough for a decent sized print. This could be intriguing.

Perhaps I would be able to utilize higher-ISOs on the G10, but my hunch is probably not unless I down-rezzed. I still think ISO 400 is about where things really start going downhill, and if I'm really in that much of a pickle I should have brought the DSLR, or I could take my chances with ridiculously slow shutter speeds and continuous drive JPEGs. So at first glance, the cameras appear to be a bit of a wash, and the G10 doesn't seem to be the significant upgrade many hoped for. There are some refinements, to be sure, but nothing earth-shattering. The unknown (which could be earth shattering), of course, is RAW. What will the G10's RAW images have to show for themselves? As I mentioned a few days ago, shooting RAW, MR from Luminous Landscape at least feels that at 13x19 print sizes, the G10 can equal Medium Format. I'm guessing the ISO 80 RAW images must be fabulous.

Hmmn. I'm really not sure about this one yet. I'd love to hear what other folks are thinking.

I'll leave you with some G9 Kodachrome Porsche awesomeness:

And a bit of G9 Kodachrome loneliness: