I want to see side-by-side comparisons of the same shots taken with the original 5D, with the same lens. I might very well be impressed at that point, but that's what it's going to take.
High ISO performance is certainly important, but so is absolute IQ at ISO 200.
fraga wrote:
iso 25600 seems pretty useless to me.
More of a gimmick than anything.
One has to consider though, that expecting good iso 25600 at present is, at least in my opinion, a bit unrealistic. Maybe in a few years.
On the other hand, having the possibility of shooting at that iso is better than not having that option, even if I think that I will never use it.
Why?
Because one day there might be a rare occasion where I might actually need it...
That is the way I feel about HD video mode.
back on topic
I think 12800 iso is quite usable for ME.
Andy Wilson wrote:
You are right I do shoot in RAW and so it is a bit comparing apples to oranges and also, as you say, this is a pre-production camera. However I'm certainly impressed with what I see so far. But maybe I'm too easily impressed. I would like to see some high iso in flourescent and tungsten lighting with WB corrected. That would give further clues.
The samples weren't exactly shot in challenging conditions, I agree.
The provided ISO 1600 and 3200 ISO samples look like the ones I get from my 1Ds3, only with more in-camera high ISO NR applied. My guess would be that differences between 5D2 and 1Ds3 RAW files are only marginal (if any at all). The benefit would be with the 5D2 when shooting JPEG, because the 5D2 has 3 or 4 levels of high ISO NR compared to 1 level on the 1Ds3. But from the looks of it, the higher levels of in-camera high ISO NR sacrifice a great deal of sharpness/details. But if you shoot RAW, none of that really matters, because you want to apply noise reduction in PP anyway.
So, IQ will no doubt be stunning. It opens up a lot of possibilities in the large printing department. But 21MP will also challenge some lenses, the 5D2 AF system and the shooters photographing skills. 21MP ask for very precise focusing. MF is going to be more challenging. The density of the pixels in combination with the 21MP ask for higher shutterspeeds. From the 5D2 files I have seen so far, a great deal seem to suffer from focus issues. Is this due to the shooter or the AF system itself (or both). Only time will tell. Once people are actually going to use the 5D2 we will certainly hear more about that
The DPReview samples show amazing noise performance.
Unfortunately, Phil has camera shake in a lot of the photos. About half the images were shot with non IS lenses at less than the 1/focal length shutter speed, so it's hard to tell sharpness on some of them. When he did get a proper shutter speed, the image quality is outstanding.
I really like what I see! Even its JPEG and no idea of the NR. But the ISO 1600 and 3200 look excellent to my. 6400 is great, too. Althoug the calculated High ISO 1 (12.800) appears to be nearly as good, as what my 30D does in 1600 and much better than it does in high ISO 3200. And not to forget its supporting 8MP instead of 21.
I did a quick test this am under tungsten light, custom WB with the 1ds3. First I tried to duplicate the in camera NR by shooting a jpeg w/ NR and adjusting the raw to match. It appears to me that the in camera uses around +5 or 6 chrom NR in DPP. I set the camera to neutral picture style because it appears these 5d2 samples are neutral, I may be wrong. The 5d2 seems really close to me to the 1ds3 as far as the amount of noise,,maybe a finer grain in the 1ds3?? My lighting really sucked here, but I would expect even better results under controlled lighting.
Whatever, Daan....if you're determined to hate on the camera regardless of what's shown, I can't help you. The fact is, the ISO 3200 shots are as good as 800 on my 30D, and that's at 100%. In a print it'd be even better. That's incredible. It maintains a very high level of detail at that level too, so if it happens to be excellent in camera NR, then so be it...it's the final image that matters to me, and these show (even with Phil's questionable settings in certain instances (200mm at 1/50s, come on?) that the 5DII has simply spectacular image quality.
Cableaddict wrote:
None of those images mean a thing to me. I want to see side-by-side comparisons of the same shots taken with the original 5D, with the same lens. I might very well be impressed at that point, but that's what it's going to take.
Well aren't we discerning!
Of course we all want to see more samples - ideally some we can take ourselves - but to say "the images don't mean a thing to me" implies that any camera would have done as well.
Considering that the photos weren't supplied by Canon - and in fact were taken by someone who supposedly hates Canon - I have to say that absent other evidence, the 5D2 looks like it handles 12,800 and 25,600 a lot better than my 5D does!
It does appear the in camera NR is pretty agressive. Its the first time I have sat down and compared a jpeg shot with NR since I usually shoot only raw.
ISO The camera has an ISO range of 100-6400, in 1/3 step or full step increments, plus ISO 50 (L), 12,800 (H1) and 25,600 (H2). The Auto ISO range is 100-3200. It's not possible to manually select a maximum ISO within the Auto ISO range, or a minimum shutter speed.
If you can't manually select a max ISO or min shutter speed, how the hell does it work?
Edit. From DPR.
Auto ISO in all modes except manual
The Mark II now features an automatic ISO option where the camera selects the sensitivity (in the range ISO 100 - 3200). In Auto, Program and Aperture Priority the camera attempts to maintain a minimum shutter speed of 1/focal length (hence with a 24 mm lens it will try to maintain 1/25 sec or faster). In manual mode Auto ISO is fixed at ISO 400.
That is pretty dumb, why not give the user the ability to select their own settings. Oh wait...that must be a pro feature to be implemented on the Mark IV
Ok, I suppose I should have phrased it better.
In my opinion, when producing sample images to show the image quality of a camera, it would be best to avoid camera shake, or use a tripod whenever possible. Shooting at shutter speeds that induce visible camera shake in the images is not the best way to showcase what a camera is capable of, in my opinion.
Jman13 wrote:
Ok, I suppose I should have phrased it better.
In my opinion, when producing sample images to show the image quality of a camera, it would be best to avoid camera shake, or use a tripod whenever possible. Shooting at shutter speeds that induce visible camera shake in the images is not the best way to showcase what a camera is capable of, in my opinion.
Phil's shots have always been dodgy, and he has admitted he is no photographer as such. I've never looked at his shots with more than a morbid curiosity as you just know they'll be disappointing if you are trying to assess sharpness or detail. Noise though does look good with the ISO 3200 watch shot quite amazing.
Cableaddict wrote:
None of those images mean a thing to me.
I want to see side-by-side comparisons of the same shots taken with the original 5D, with the same lens. I might very well be impressed at that point, but that's what it's going to take.
I'll say that the 5DII takes better quality pictures than my 5D at ISO 6400, 12800, and 25600. Happy now?
We saw with Nikon that high ISO performance can be significantly enhanced with software computation. However, the trade-off is significant loss of sharpness & detail.
I see that the 5DII has three levels of iso enhancement, so that implies there are similar trade-offs to the Nikon system
I have no doubt at all that the 5DII has TRULY improved high ISO performance. It certainly should. -but I want to know exactly how much better it is, WITHOUT the extra processing, in order to base my upgrading decisions. So, if the sample pics have a small amount of shake, it does not help me very much. I want to see high ISO shots, with the processing set to low (or off) and as much sharpness as possible. That obviously requires a solid tripod mount.
---------
Note: this is all really just academic for me. Again, whatever the actual ISO performance is, it is SURELY a significant improvement. -And I'm not about to switch to Nikon. I'm just raising a point that might be important to others, more on the fence than myself.
It's also always a good thing to try to spot flaws in any testing methodology.
he does note that the images have "stands nr". i'm not sure if thats incamera or a plugin. anyone? i'm still going to wait a month and see if canon announces a 1d3 price drop before ordering the 5d2.
