Finally, some meaningful comparisons. These shots, as well as the RAW files from the same source are very promising.
Mel Gross wrote:
I'm impressed. The noise us definitely better.
While the second image doesn't show much for detail, the first does. The mkII clearly is a good deal sharper at the same mag.
Not only that, but I can also see that apart from improved resolution and high ISO handling the images from two cameras are very very similar. And this is great news! Canon clearly managed to improve on 5D sensor. Coupled with a significant number of other updates this certainly makes it worth the extra $$ FOR ME. Now I only wish to get one before the end of November.
The only way I have been able to come close to even a 1 stop difference is with in-camera NR on and NR on in DPP for the 50D and NR turned off for the 40D.
I would disagree with your contention in practical terms. If the 50D takes noise reduction "with better grace," retaining more detail than the 40D, so that the end result is still superior, then for practical purposes, a superior image is a superior image. That it has to do so by means of initial RAW rather than better noise reduceability is not practically significant.
This is photography. The only thing that matters is what the picture looks like.
RDKirk wrote:
I would disagree with your contention in practical terms. If the 50D takes noise reduction "with better grace," retaining more detail than the 40D, so that the end result is still superior, then for practical purposes, a superior image is a superior image. That it has to do so by means of initial RAW rather than better noise reduceability is not practically significant.
This is photography. The only thing that matters is what the picture looks like.
That is the problem it doesn't retain more detail when NR is used. If it did then all would be rosy. It is just the opposite. NR on the 50D is causing a loss in detail vs none in the 40D. The only way to keep the detail is to turn off NR and then the noise levels are the same as the 40D.
Canon claimed the 1-1 1/2 stop better noise control than the 40D and it is just not true with both files processed without NR. Which means Canon's attempt to improve noise levels on the 50D by using NR is an improvement in software not hardware. The problem is the outcome results in less detail than not using NR.
Question about NR. If NR results in some loss of detail, does it mean that because you have lots of resolution and detail to begin with (with the 21-megapixel 5D II) you can afford to lose some of this resolution/detail and still end up with say 18 megapixels of resolution and detail in the final noise-reduced image? If so, the 21-megapixel is a real winner in my books i.e. if you can shoot at ISO 3200 with NR, and end up with say 18 megapixels worth of info/details.
Not if it results in smearing rather than just a little softening.
This is what my mention of the 50D claims were about. I didn't expect much from the 50D but if Canon is playing the same shell game with the 5DMKII I'm going to be pissed.
That is the problem it doesn't retain more detail when NR is used. If it did then all would be rosy. It is just the opposite. NR on the 50D is causing a loss in detail vs none in the 40D. The only way to keep the detail is to turn off NR and then the noise levels are the same as the 40D.
So you're saying Canon never produced a better image from the 50D than from the 40D?
Mel Gross wrote:
If that's true, then we would have to watch it. But from what I've been reading about the 50D, the IQ is said to be much better. I haven't compared the two yet myself. My plan is to replace my daughters 40D with the 50D in the next couple of months, and my 5D with the mkII shortly after.
the detail is much better, nothing else about the IQ is better.
noise is, at best, equal after downsizing it to the 40D's 10MP.
not that 15MP with similar noise is a bad thing though....
Kodak invented Kodachrome, not POP Photo. Besides, I think he was being facetious. Can you point to an actual POP Photo article on this?
i don't recall the issue off hand.
best i can do is pre-2008 issue and post 2004
i may come across it again eventually when i clean up some old piles of magazines eventually
all I recall is that they said that the best color print film was sharper than the best color slide film and that the 1DsMkIII revealed a little bit more detail than the best color print film they compared it to
of course they have also said stuff like D200 has better high ISO than 30D (i think)
RDKirk wrote:
So you're saying Canon never produced a better image from the 50D than from the 40D?
I guess your just not following what I have been saying. I'm talking about NR and its effect on the 50D files, which would be necessary to try and obtain Canon's claim of better noise levels. I never said anything about whether the IQ of the 50D was better or not than the 40D.
I was trying to point out the fallacy in Canon's statement about better noise levels in the 50D and that I would hope their similiar statement about the 5DMKII wasn't full of smoke too.
The 50Ds overall IQ is better than the 40D as long as your not noise challenged.
Hopefully the 5DMKII's noise control is better than the 5D without relying on NR to do so. In other words an actual hardware change resulting in better noise control instead of just software.
Jim Victory wrote:
That is the problem it doesn't retain more detail when NR is used. If it did then all would be rosy. It is just the opposite. NR on the 50D is causing a loss in detail vs none in the 40D. The only way to keep the detail is to turn off NR and then the noise levels are the same as the 40D.
Canon claimed the 1-1 1/2 stop better noise control than the 40D and it is just not true with both files processed without NR. Which means Canon's attempt to improve noise levels on the 50D by using NR is an improvement in software not hardware. The problem is the outcome results in less detail than not using NR.
skibum5 wrote:
i don't recall the issue off hand.
best i can do is pre-2008 issue and post 2004
i may come across it again eventually when i clean up some old piles of magazines eventually
all I recall is that they said that the best color print film was sharper than the best color slide film and that the 1DsMkIII revealed a little bit more detail than the best color print film they compared it to
of course they have also said stuff like D200 has better high ISO than 30D (i think)
I remember that article very well. K25 wasn't in the picture (no pun intended), K64 and other films were..
Mel Gross wrote:
I remember that article very well. K25 wasn't in the picture (no pun intended), K64 and other films were..
did they just not test it or had K25 already stopped production (it's not still made is it?)?
this reminds me i need to get on developing my last few rolls of film fast before the labs disappear.
skibum5 wrote:
did they just not test it or had K25 already stopped production (it's not still made is it?)?
this reminds me i need to get on developing my last few rolls of film fast before the labs disappear.
Sadly, it was already gone. The ret of it wil be gone soon. I still have a couple of K64 220 rolls I saved as mementos.
I just picked up two rolls of the new Ektar 100 at the Pro Photo expo the other day. Kodak claims it is the finest grained film in the world (I'm assuming they aren't including specialty films).
I haven't shot film for a while, but I'm interested in testing this.
A quote from the article:
"This tech-speak, says Chuck Westfall, Technical Advisor at Canon USA, adds up to the 50D performing better at higher ISO settings than the 40D, despite the new camera's smaller pixels (4.7µm for the 50D vs 5.7µm for the 40D). Purchasers of the 50D, says Westfall, can expect noise levels to be roughly 1 to 1.5 stops better in the high ISO range. For example, this means that a photo shot at ISO 3200 on the 50D is expected to have similar noise characteristics to one shot in the ISO 1100-1600 range on the 40D."
skibum5 wrote:
i don't recall the issue off hand.
best i can do is pre-2008 issue and post 2004
i may come across it again eventually when i clean up some old piles of magazines eventually
all I recall is that they said that the best color print film was sharper than the best color slide film and that the 1DsMkIII revealed a little bit more detail than the best color print film they compared it to
of course they have also said stuff like D200 has better high ISO than 30D (i think)
Well, I think I recall reading a similar article from whatever magazine saying the other way round, slides are better than film, that is why lots of photogs take slides and print from slides with the positive to negative conversion (taking the loss into consideration).
For the digital to film comparison, I would rather not get into because it can fuel a whole new thread. But after reading research reports done by Fuji, Kodak, Leitz, Zeiss and others, I beleive in those people that it will take a sensor with over 200MP using a Bayer GRGB color layout to reach the standard of present day 35mm film. For Foveon sensor with RGB stack will require less MP.
And taking the present day semicon industry standard we are not there yet to meet the photo site size and noise requirement.
By than we will not need to have an AA filter for it. One can almost forget about the moire effect.
Lens manufacturing standard will have to be improved because poor lenses does not meet resolution requirement.
As I said I will not get into a debate on this subject because it will need to review a lot of research that people are still working on, so count me wrong if you do not beleive in it.
Note: Color slides and color films of equivalent ISO. B&W should not be use for this comparison.
RDKirk wrote:
As I said before: This is photography. The only thing that counts is what the picture looks like.
And what in hell does that have to do with anything?
You could also say it is only in the eye of the beholder. What I think a picture should look like is not what you would think a picture should look like. You apparently don't have the brain matter to discuss something rationally without some inane comment.
