Rob Galbraith's final findings on the Canon 1D MarkIII series AF

Rob Galbraith finally published his findings on the latest Canon 1D series AF performance, and it's a pretty neutral rating on the AF system. Ostensibly, it seems that the sport shooter ( and by extension the nature shooter) will have more OOF shots with the AI Servo mode.

What does this mean for the underwater shooter? Pretty much nothing as much of the shooting is done in one shot mode without having to track a highspeed subject swimming at you. Only highly dynamic action and specialized sequences (like the sequence of a diving gannet at 40km right at you) would it matter. The weakness of the 1DIII seems to be in the brightest conditions and underwater, we are pretty much down a few stops anyhow.

Rob does say the 1DIII files are awesome and I agree wholeheartedly. However I shoot highly dynamic subjects like diving birds and hunting dolphins which move very quickly. I found the 1D3 to be OOF more often than my older 1D2N, that is why I sent my 1D3 back... twice!

Then again, I know many people who are so stoked with the 1D3 files and don't shoot erratic moving high dynamic subjects. Caveat emptor!

http://robgalbraith.com/bins/multi_page.as...-8740-9068-9537

Sometimes when I read these guys they all start to sound the same, they all have biases ala Rockwell, and some type of position to defend. In this case I'm surprised how he gives Canon a pass on this one. If there were one thing critical in your flagship 'sportscamera' it has to be autofocus ability, irrespective of the awesome file you are able to achieve when the frame happens to be in focus. What good is the 'achievable' image quality if the shot at that instant that you want is out of focus. And then he exposes his bias even more by introducing the Nikon D3 and again justifying the choice of the 1DMkIII over the Nikon because of some marginal and questionable superior image quality of the Canon. When it comes to image quality in this range of cameras it seems quite clear that when one combines the file from the camera with very limited initial postprocessing in a RAW processor, all these cameras can produce similar image quality. There is even a thread here http://www.professionalphotographer.co.uk/...iii-vs-nikon-d3 that shows the Nikon D3 resolution/image appearance to be as good or better than the 1DSMKIII with simple interpolation of the D3 image.

So to sum up my rant, he should not be giving Canon a pass on this issue, and as a second issue he should just state that Nikon really nailed the sportscamera class with the D3.

Tell us how you really feel, Jeff.

I assumed Galbraith might come out with something like this - I think many pro shooters have learned to live with the intermittent issue, with the positive trade-off being the absolutely fantastic quality of the 1DIII/1DsIII images.

Hard to draw conclusions from the 1DsIII vs D3 article, seeing as how they only show a single 100% crop comparison of the two, no?

From Galbraith:

"The EOS-1D Mark III produces the best overall file of any digital SLR we've ever used. A CR2 coming from this camera contains a level of detail, dynamic range, tonality and overall quality that is hard to match."

Hard to believe that's true. Anyone thinking would realize that there are cameras that should be able to match based solely on those criteria. Medium format? Perhaps "they" ought to try more digital SLRs.

"The Nikon D3, for example, produces a better NEF at really high ISO settings (plus very good quality generally), but at lower ISOs it can't quite deliver the fineness of detail from each pixel that the EOS-1D Mark III does."

Hard to believe that as well. Perhaps "from each pixel" but the D3 has more pixels. It would be hard to believe that the D3 couldn't compare favorably to the 1D3 in overall resolving power. It's well known that Nikon has a stronger AA filter than Canon at this time. That has advantages and disadvantages to image quality.

I think Jeff is right, even the most capable reviewers seem to have brand bias. There are claims out there that the D3 produces resolution as good as the 1Ds3 does. That's remarkably hard to believe as well. I don't understand why everyone must be a member of a "team". Why so much insecurity?

Hard to draw conclusions from the 1DsIII vs D3 article, seeing as how they only show a single 100% crop comparison of the two, no?

I agree, I just brought that up to make the point that image quality is so good with all these cameras, that with fairly simple processing (not sophisticated image manipulation) it can be hard to tell the difference between cameras.

Pretty much nothing as much of the shooting is done in one shot mode.

To prevent people from switching to One Shot mode after reading this post, I want to point out that I pretty much never use One Shot mode. I use AI Servo, on the * button, and just let go when I'm done focusing.

Jeff

I'm not sure Rob gives it a pass. He likes the CR2 files and is a bad writer, forgetting to contextualize his statements as Craig noticed. Having tried the D3 in Africa, I'm sort unconvinced the D3 can do much better than the canon 1D series in chasing gannets taking off right at the camera. I had pretty much the same number of OOFs at 9fps in bring sunlight on a heavily contrasted subject (white gannet against dark blue sea). I will say the D3's exposure was slightly more accurate. Still I think it's splitting hairs on whichever one is better. Also having seen the 21mp file of the 1Ds, the resolution is just ridiculously good. I almost had resolution envy when I was watching Matt being able to frame not so accurately or have the wrong lens and still be able to crop and still have a file with twice my resolution.

I've never been a brand camp myself. In fact, I'm itching for a Nikon camera again. I still have a few lenses leftover and I really just want 2 lenses from Nikon right now... 14-24 and 200-400 ... oh and that SB900. I wish there was a universal charger that can charge all brands so I don't have to carry 4 different chargers.

Eric you are correct. I was a bad writer there. I meant one shot in the sense that there's no need for real tracking of a subject per se, except in macro. I doubt anyone really reads what I say anyhow LOL

LOL! Drew, just because it doesn't look exactly the way you want it to look doesn't mean it's framed "not so accurately." And you were the one trying to crop my photos!

Jeff

I'm not sure Rob gives it a pass. He likes the CR2 files and is a bad writer, forgetting to contextualize his statements as Craig noticed.

Yeah, as I said, these are all great cameras able to produce spectacular results. I am tickled with my D700; and the D3, DMKIII and DSMKIII are all pretty awesome. I just think that saying it's OK to accept a bunch more out of focus shots because the one that is in focus turns out to be arguably a little sharper is kind of dumb.

yeah but the cropping gave you a chance to get published... so you should thank me you ingrate! And I can't help it if you have crazy eye!

I've shot with all the cameras he compared (the D3 plus the 1D mk IIN, 1D mk III, & 1Ds mkIII) and can say that they are all fantastic cameras. The 1D mkIII did have problems - I was provided 2 by my employer, and one of them did not perform the way it ought to... but after the fix it certainly seemed to be better. I feel like it is different than the mkIIN - not necessarily better or worse, but different. I asked for a D3 for comparison and was provided one. For grab focus (surface stuff - breaching whales, etc) it felt like the mkIII responded faster. When looking at the final files, the D3 had less detail than the mkIII files (and you can see this if you look at 100% crops). It's not a sharpening issue - the detail just isn't there. I think the color out of camera looked a little better on the D3, but detail is a bigger concern for me. I think it goes beyond the anti-aliasing filter strength and has to do with sensor design. For smaller prints, the D3 would be fine, but for very large prints the 1D mkIII exceeded it for detail. The 1DsMKIII is a step beyond that, but it's in its own category.

For nature shooting, the mkIII served me better than the D3 - but they are both pretty incredible cameras.

I'm not the only one who thinks the AF on the mkIII is better either - http://tinyurl.com/4cwr3q

Dpreview compared the D3 directly against the 1D3 and 1Ds3 here. As would totally be expected, the D3 outresolved the 1D3 by a small amount and was beaten clearly by the 1Ds3. The 1D3 displayed visible moire in the tests because of its inadequate AA filter. The test images are pretty clear and dpreview has been performing these tests for a long time. I don't think they got it wrong.

I honestly don't see how anyone can believe that a company like Nikon (or Canon) would get something so fundamental as sensor resolution wrong. In any given test if there's a significant difference in resolution between the two cameras it will be because of technique, methodology, or lenses. The 1D3 sensor does not outresolve the D3 one.

As far as AF goes, it's easy to find believers on both sides of the argument. What's interesting is that lenses matter yet people like to argue bodies like there's a definitive answer. I fail to see how it would effect underwater shooting regardless. More than likely, the D3 will be for UW shooters just like the 1D3. It will be a body that can share a housing with its preferable, higher resolution sibling. Unlike the 1D3 it won't offer a difference in perspective than may be desired.

P.S. Regarding the AA filter, I think that a weaker one is preferable for underwater shooters. I wish I could have the option to remove it. I'd like to be able to remove the UVIR cut filter as well. If Nikon introduces a rumored modular system perhaps that will be easier. As it stands, the aftermarket for those mods is focused on Canon bodies.

Craig,

what I like most about that D3 review is this quote:

....anyone even close to considering a D3 will be more than capable of making an informed decision without a technical review such as this...

Basically he's saying reviews like that are for entertainment and gawkers only

However, Rob G's tests are pretty real world in terms of a specific task for the AF to do. Seeing how at various sporting events, I'm seeing more Nikon D3s and quite a few big names have switched due to the AF problems on the 1D3, I'd say it's pretty indicative of the frustration some have had over the 1D3 fiasco.

Yeah, I liked that quote too.

Galbraith is one of the best IMO. I believe these cameras are so good that people end up seeing what they want to see or expect to see. You have to credit Galbraith for his contributions in bringing the Mk3 AF issues to light.

Craig

I reckon you summed it up pretty well!

We have access to photographic tools beyond anything we even imagined a few years ago. In the real world all the mdels discussed here are pretty amazing and on a stand alone basis all will yield results beyond reproach to almost every viewer. Issues such as this one (which I seem to remember was actually brought to light by Galbraith) affect some relatively specialised applications - again Galbraith seems to be into news sport and sometimes working in the precise conditions where the issue arose. Whilst I can see that they might justs, very occasionally show up underwater I'd say that this will be very rare and certainly isn't something over which I'd suggest one brand over another.

Now if Nikon would produce a 24mm f/1.4 or 20mm f/2, then I'd be far more interested in looking at a Nikon rather than because of the issue highlihted on the 1DMkIII!

Paul, don't be a brand snob, Sigma has a 20 and 24 f1.8.

Yes the Sigmas! Personally I've never got on well with them - if they were only for underwater use I wouldn't have a problem with them, but given that I'd be carrying them as well as the 24/1.4 which I use a great deal above water too, I don't want yet another lens to cart around. I find Canon, Nikon and Leica offer well made, and quickly and economically repairable lenses. My experience with other brand lenses has been that they quickly become uneconomic to repair and this I have found to be frustrating, when a lens has been discontinued or upgraded. So I suppose I am a brand snob but for very specific reasons! Apertures 1.8 to 1.4 may not seem much but it can be!

I honestly don't see how anyone can believe that a company like Nikon (or Canon) would get something so fundamental as sensor resolution wrong. In any given test if there's a significant difference in resolution between the two cameras it will be because of technique, methodology, or lenses. The 1D3 sensor does not outresolve the D3 one.

I wasn't suggesting that the 1D3 sensor outresolves the D3 sensor - the D3 has more megapixels, and a higher resolution. What I had said, and what I think is apparent if you look at the files from those cameras when shot under the same conditions, is that the 1D3 preserves more detail than the D3 does. It may be a combination of things - anti-aliasing filter, pixel density, in-camera noise reduction, etc but it is the case. Post-processing narrows the gap, but the detail preservation in the 1D3 files definitely is higher than in the D3 files.

The flip side of that is that the 1D3 files appear slightly noisier than the D3 files. The appearance to me is that even at its default settings the D3 has some in-camera noise reduction that is causing some detail loss.

The AF issue has been a problem. I hung on to my 1D2N bodies much longer than I thought I would (I still have a few) because of the problems with the 1D3 AF. As I said before, some of the bodies have the problems, and some don't. It's inexcusable in a camera of this caliber.

I'm not proposing that dpreview, rob galbraith, prophoto, etc got anything wrong. As one of the only people on this board who has shot both cameras in comparison side-by-side, I can offer a real world experience, not just spouting what your favorite blogger has to say about the cameras.

Brian, are you talking about RAW files? It's generally accepted that there are differences between Canon and Nikon JPGs. For RAW files, no in-camera noise reduction is done. I don't believe data supports the assertion that the 1D3 offers more resolution when RAW files are used.

Pixel pitch is tricky. Finer pitch over a smaller sensor does not mean more overall resolution but it could offer more depending on the lens used. That's the age old argument Nikon used when justifying their lack of FX offerings. It's conceivable that the 1D3 does better in some tests and worse in others based solely on the tests themselves and the lenses being used. That's why I say that even the most experienced reviewers / pros can end up seeing what they are predisposed to see.

I wouldn't discount any personal experience that says one camera is better resolving than another, I'd just chalk it up to the particular method used. Thing is, there are well informed opinions that are entirely contradictory on the subject. I think resolution differences between the two are negligible personally.

You're confusing resolution with detail. I know it sounds odd to make a distinction, but it does exist, looking at either RAW/NEF files or jpegs. I'm not talking about resolution - you can't argue the fact that the D3 is higher resolution. That's simple. I'm talking about the detail in the photos taken. The 1D3 seems to have greater detail than the D3 files when shot under the same circumstances. I shot a 1D3 with a 400/2.8IS next to a D3 with a 400/2.8VR literally side by side and with the same settings. I had more DETAIL in the shots from the 1D3.

I'm not trying to defend Canon - I've seriously considered shifting my personal gear to Nikon for several reasons, and my gear provided by my employer is a mix of both (why doesn't Canon make a 14-24? Why doesn't Nikon make a 14/2.8 or a 24/1.4 or 35/1.4? A mixed kit solves some of these problems).

Galbraith agrees with this point as well (for "well informed opinions") - he says "The EOS-1D Mark III produces the best overall file of any digital SLR we've ever used. A CR2 coming from this camera contains a level of detail, dynamic range, tonality and overall quality that is hard to match. The Nikon D3, for example, produces a better NEF at really high ISO settings (plus very good quality generally), but at lower ISOs it can't quite deliver the fineness of detail from each pixel that the EOS-1D Mark III does."

So you need to understand the difference between resolution and detail. Higher resolution creates a larger dimensioned file, but that doesn't mean it's one with more detail. That has been my experience comparing the D3 to the 1D3.