A Biased Evaluation of The Differences...

[Ray]

It would be nice if someone shot the exact same scene with the same lens and Av and TV values with all three cameras, but no one who isa ble to do so seems interested in doing so, and providing the RAWs.

This reminds me of the days when Canon's first FF DSLR, the 1Ds was released. Prior to that, arguments about the advantages and disadvantages of the cropped format had been frequent, yet those who then owned both cameras, a 1Ds and D60 or 10D, were strangely reluctant to shoot the same scene with the same lens to demonstrate either the resolution advantage of 6MP compared with the 4.2MP of the 1Ds image after cropping, or the superiority of 4.2M 1Ds pixels because they were bigger and perhaps therefore better.

One can only assume that the psychological element of discovering that under certain circumstances (with a telephoto lens that wasn't quite long enough) the el cheapo 6MP camera might produce better results, was just too much to bear.

In general, though, how does this compare to a 100% crop at ISO 51000 on the other cameras?

Well, earlier in the thread I showed a 5D ISO 64,000 crop, the Chiang Mai Mona Lisa, which looks pretty good to me, considering it's ISO 64,000   . However, I didn't use a specialised converter like DCRAW to drag every last spec of detail from the image.

The problem I have with your 1Ds3 crop above is that it seems to have some seriously blown highlights which might sort of disqualify it from the really high ISO stakes.

[John Sheehy]

The problem I have with your 1Ds3 crop above is that it seems to have some seriously blown highlights which might sort of disqualify it from the really high ISO stakes.
[a href=\"index.php?act=findpost&pid=169131\"][{POST_SNAPBACK}][/a]

Please explain.  The fact that there are much brighter parts in the full image have no effect on whether or not this is an exposure index of 51000.  The full range of tones is available there for ISO 51000.  This is a simple conversion; no extra highlights rolled in.  It clips harshly.

I don't have access to these cameras discussed, so I can only take what I can find RAW on the internet.

[Ray]

Please explain.  The fact that there are much brighter parts in the full image have no effect on whether or not this is an exposure index of 51000.  The full range of tones is available there for ISO 51000.  This is a simple conversion; no extra highlights rolled in.  It clips harshly.

I don't have access to these cameras discussed, so I can only take what I can find RAW on the internet.
[a href=\"index.php?act=findpost&pid=169133\"][{POST_SNAPBACK}][/a]

John,
All my estimates of ultra high ISO are based on the degree of +EC adjustment I need to make in ACR to simulate a correct ETTR at the camera's maximum ISO setting, with all other controls at zero, including highlight recovery.

The presence of bright highlights seems to confuse the issue when using this method. It's so easy to exaggerate the ISO figure by deliberately blowing highlights which are really larger than specral and perhaps should contain detail.

That's why I like to show a screen capture of the ACR window showing clearly the histogram and EC adjustments.

However, if you are using tools which analyze the RAW data at the pixel level in a more precise manner, then perhaps you don't have to rely upon such crude methods   .

[jjj]

Not to reopen old wounds, but I hope you aren't suggesting I misrepresented what you said, because I never did. I did make some comments that were intended as good natured needling (my "smart aleck" remarks), but I never misrepresented you. Also, I have to say that even tho you may have felt provoked by some of my comments, you strike me as someone who is looking to be provoked, or at least someone with a chip on his shoulder. Several of your posts seemed very combative right from the get-go, and you made several needling remarks of your own, which is why I responded to you the way I did. I really have nothing further to say on the matter. Hopefully this forum won't degenerate into the rude free-for-all we see on DPReview and other sites.
[a href=\"index.php?act=findpost&pid=169087\"][{POST_SNAPBACK}][/a]

Hmm, I started off by being very positive about the D3 and Nikon, so hardly combatative from the get go. You seem oversensitive and yes you did misrepresent what I said. Obviously you cannot see that you did that. So you seem to be the one provoking and even admitted to it. So if you want to avoid 'DP Review style' rudeness [I rarely look there so I'll have to take your word for it], stop being a smart alec and read posts more carefully in future. Most online bickering occurs after someone fails to read a previous post/posts carefully and in context.
And if you think my description of Nikon cameras' failings for my needs , not someone else's, is needling, then you really need to chill out. Nikon have only just launched a FF camera, so if you wanted a FF camera you had to buy Canon. Hardly a difficult issue to grasp. Yet you accused me of bias for choosing a Canon 5D for that very reason. I wouldn't buy a 1dIII for the same reason, in fact I'd buy a D3 in preference, if that was the only choice.

[Ray]

Of course there could be an explanation for my test results showing a mere 1/3rd stop high ISO noise advantage of the D3 over the 5D. It's well known that Canon lenses have QC variability. I think this also applies to camera bodies. Whether to a greater or lesser extent I don't know.

However, I rejected my first 5D on the grounds it had to much banding in the shadows. The replacement was better. Perhaps I have an unusually good 5D. Perhaps the D3 demo model, the only D3 in the store that I was able to test, was below par. Who knows!

Experiments like this should be repeated.

[Quentin]

I know Ray does not trust tests, but every test I have read in UK reviews - every single one -  has been staggered by the high ISO performance of the D3 compared with previous generation dslrs of any brand.  

Quentin

[DarkPenguin]

That is true.  Wasn't Ray's point that the ISO on one doesn't equal the ISO on the other.  I know that when I compare the 20D to the 400D the ISO are not equal.

[Rob C]

That is true.  Wasn't Ray's point that the ISO on one doesn't equal the ISO on the other.  I know that when I compare the 20D to the 400D the ISO are not equal.
[a href=\"index.php?act=findpost&pid=169792\"][{POST_SNAPBACK}][/a]

No, I think Ray´s point was that he has grown weary of the tennis match! ( ;-) )

Rob C

[DarkPenguin]

That doesn't parse right.  I have to assume I meant "That is true.  But wasn't Ray's" yada yada.

[Ray]

I know Ray does not trust tests, but every test I have read in UK reviews - every single one -  has been staggered by the high ISO performance of the D3 compared with previous generation dslrs of any brand. 

Quentin
[a href=\"index.php?act=findpost&pid=169762\"][{POST_SNAPBACK}][/a]

Quentin,
I bet those reviews you refer to didn't show any 100% RAW crop comparisons.

The reason I went to the trouble of doing this test is that all the reviews I came across on the net simply didn't fully explore any comparisons. The attitude seemed to be, " We can't show you the 5D (1Ds3 whatever) at ISO 64,000 or 25,600 because these cameras don't do that.".... Well, really! What can one say.

Ken Rockwell's review was the only one which showed a comparison between the 5D and the D3 of the same subject at the same shutter speed and aperture. I was surprised to find that his noisier 5D shot cleaned up so easily in Noise Ninja and looked virtually as good as the D3 shot. I see Ken has now done further tests at simulated ISO's.

The other issue for me was to find out what I might expect in the 5D upgrade. If Nikon could provide as much as a 2 stop noise advantage at high ISO, then Canon would quite likely match that. If they could, then I wouldn't even be concerned if the pixel count remained the same.

It now seems clear to me that we can probably expect around a 1/2 stop noise improvement in the 5D upgrade, which might make it pixel-peeping better than the D3 regarding high ISO noise, but no big deal.

[Ray]

That is true.  Wasn't Ray's point that the ISO on one doesn't equal the ISO on the other.  I know that when I compare the 20D to the 400D the ISO are not equal.
[a href=\"index.php?act=findpost&pid=169792\"][{POST_SNAPBACK}][/a]

There is a slight difference in ISO accuracy between the 2 cameras, around 1/3rd of a stop, which is one reason why it's important to do comparisons at the same shutter speed and aperture. But that's really only a small part of the problem.

I'd say the major problem is that reviewers simply don't take the time to compare RAW images. I wouldn't be surprised if many of them don't even take any shots in RAW mode. It's my experience than most photographers in general (if one includes the amateurs) shoot jpeg. Anyone who's in the habit of shooting in jpeg mode would quite likely get the impression the D3 is about 2 stops better regarding noise at high ISO.

[Quentin]

Quentin,
I bet those reviews you refer to didn't show any 100% RAW crop comparisons.

The reason I went to the trouble of doing this test is that all the reviews I came across on the net simply didn't fully explore any comparisons. The attitude seemed to be, " We can't show you the 5D (1Ds3 whatever) at ISO 64,000 or 25,600 because these cameras don't do that.".... Well, really! What can one say.

Ken Rockwell's review was the only one which showed a comparison between the 5D and the D3 of the same subject at the same shutter speed and aperture. I was surprised to find that his noisier 5D shot cleaned up so easily in Noise Ninja and looked virtually as good as the D3 shot. I see Ken has now done further tests at simulated ISO's.

The other issue for me was to find out what I might expect in the 5D upgrade. If Nikon could provide as much as a 2 stop noise advantage at high ISO, then Canon would quite likely match that. If they could, then I wouldn't even be concerned if the pixel count remained the same.

It now seems clear to me that we can probably expect around a 1/2 stop noise improvement in the 5D upgrade, which might make it pixel-peeping better than the D3 regarding high ISO noise, but no big deal.
[a href=\"index.php?act=findpost&pid=169882\"][{POST_SNAPBACK}][/a]

Ray

1.  Several showed crops from the D3, one a double page spread shot at some ridiculously high ISO which it claimed would have been impossible before the D3

2. Nikon already had a 2 stop advantage; its Canon that has been left behind on this occasion.  I realize that's not your reality, but it is just about everyone else's.

Quentin

To be entirely open here, I have to say I dislike Canon dslr's and I don't now and likely never will own one of them (I do though have the excellent G9).   Nothing would make me happier than to seen Canon's backside kicked     So am I impartial?  Not a chance.

Here is a other thought.  Suppose for the moment you were right and the D3 high ISO advantage was nominal. While I don't for one moment believe that is the case, if it was, then its a brilliant marketing coup for Nikon; they now have the backing of just about all major independent reviewers and journals who have tested it.

[Streetshooter]

Ray
Here is a other thought.  Suppose for the moment you were right and the D3 high ISO advantage was nominal. While I don't for one moment believe that is the case, if it was, then its a brilliant marketing coup for Nikon; they now have the backing of just about all major independent reviewers and journals who have tested it.
[a href=\"index.php?act=findpost&pid=169926\"][{POST_SNAPBACK}][/a]

Well, according to Ray, they don't know what they're talking about !

Pete

[michael]

Ray, hate to say it, but though you seem to otherwise be a reasonable sort of fellow, on this subject you're simply wrong. This comment is based on my intensive experience comparing the D3 and the 5D and 1Ds MKIII, and that of quite a few other photographers who have done detailed side-by-side comparisons of the D3 at high ISO against Canon's best.

You may decide that we are all wrong (and this includes probably half a dozen reputable magazine and web reviewers as well, many of whom I know to be level headed and knowledgeable people) but then again maybe, just maybe, it's you that isn't seeing things for what they are.

You can push process Canon files till the cows come home, and you can claim that the D3 smoothes detail, but neither of these changes the fact that knowledgeable observers seem to be in agreement that there is at least a two stop advantage to the D3, and that it's real, quantifiable, verifiable and there for anyone with an experienced eye to see for themselves.

But, you are of course free to believe whatever you wish. It seems pointless though to continue to argue otherwise here since all that we have from you is your opinion on the matter, while the other side of the argument has experienced observers on three continents claiming otherwise.

Michael

Ps: I would add that "cleaning up a 5D image in Noise Ninja" to make it competitive is a red herring, simply because one can then do the same thing to a D3 and still have the two stop advantage.

As for quoting ken Rockwell on the matter, well, what can I say? I should simply leave it at that.

[Ray]

Ray, hate to say it, but though you seem to otherwise be a reasonable sort of fellow, on this subject you're simply wrong. This comment is based on my intensive experience comparing the D3 and the 5D and 1Ds MKIII, and that of quite a few other photographers who have done detailed side-by-side comparisons of the D3 at high ISO against Canon's best.

I am a reasonable fellow, Michael, but what can I do? I take a number of RAW shots with both cameras covering a range of F stops and shutter speeds, 5D set on ISO 3200 and D3 set on various ISOs from ISO 3200 to 25,600. I have dozens of RAW files that I can compare at the same exposure, 1/3rd stop difference, 2/3rds stop difference, 1 stop, 2 stops and so on.

I would have liked to have done more tests. I tried to hire the demo D3 for a day but the shop wouldn't let it go. I did make a few errors in not precisely matching FoV. The D3 shots are also tilted up slightly because the single focussing square in the viewfinder was not actually in the centre. (I didn't realise one could change the position of that square on the D3.)

I would add that "cleaning up a 5D image in Noise Ninja" to make it competitive is a red herring, simply because one can then do the same thing to a D3 and still have the two stop advantage.

I've not found that to be the case. The 5D image is improved by reducing chroma noise (keeping luminance NR at zero) and without noticeable loss of resolution. The chroma noise in the D3 image seems to have already been removed in-camera. Attempting further noise reduction with the D3 image tends to reduce resolution.

As for quoting ken Rockwell on the matter, well, what can I say? I should simply leave it at that.

It was because Ken Rockwell's tests comparing the D3 with the 5D were the only comparisons I could find on the net that I decided to do some tests myself.

If people who own both a D3 and 5D don't show their test results, what can I do? In the final analysis, I have to trust my own eyes. As you know, I've spent years on this forum debating such issues as noise and resolution. I think I can recognise noise when I see it.

If you or anyone reading this post who owns both a D3 and 5D care to demonstrate the 2 stops noise improvement you claim, using RAW images, and comparing 100% crops of exposures that really differ by two stops rather than ISO settings that differ by 2 stops, I'd be very interested.

[Ray]

1.  Several showed crops from the D3, one a double page spread shot at some ridiculously high ISO which it claimed would have been impossible before the D3

Quentin,
But did they show shots of the 5D or 1Ds3 downsampled to the D3 size for the purpose of comparison. And if they did show such crops, did they take them at the same exposure and simulate the high ISO of the D3 by underexposing?

2. Nikon already had a 2 stop advantage; its Canon that has been left behind on this occasion.  I realize that's not your reality, but it is just about everyone else's.

Everyone? Could this be another bit of hyperbole   .

Here is a other thought.  Suppose for the moment you were right and the D3 high ISO advantage was nominal. While I don't for one moment believe that is the case, if it was, then its a brilliant marketing coup for Nikon; they now have the backing of just about all major independent reviewers and journals who have tested it.

I'm sure you realise that there are relatively few reviewers that take the time and trouble to do thorough, detailed and comprehensive reviews. Most reviewers take the conclusions from the experienced few and dress up the results to make it appear they have done the work themselves.

I'm not into conspiracy theories, but I do think there's something odd going on here.

Below are a couple of jpeg images which show the 5D shot, at one stop more exposure than the D3 shot, still looks better, although both are about equally noisy, the implication being that even in jpeg mode the D3 in terms of general image quality has less than a one stop advantage.

For these shots I used tripods, but the 5D tripod was a bit lower. However the FoVs are very close. On the D3 I had the Nikon 50/1.8 prime and on the 5D the 24-105 zoom set at 50mm.

The 5D shot looks sharper and more detailed to me, and generally better. How does it look to you?

[attachment=4929:attachment]  [attachment=4930:attachment]  [attachment=4931:attachment]