deepbluejh wrote:
I think film handles hard contrast a little more gracefully than many digital cameras. I don't have any science to back this up, but that seems to be the perception.
Is this due to their choice to keep the high iso better than on the same chip but with more accurate rendition in the 1Ds3? I know the Sony A900 people were saying that the high iso was weak specifically because they had optimised for incredible colour at the lower end of the iso spectrum.
Well the less sharply peaked filters do let more light through so you get better luminance SNR at the expense of a bit worse chrominance SNR and if they have tricky double main spikes and such then it can be hard to get the color as accurately even forgetting about SNR. Canon is pushing this filter trick farther than the other makers.
Hopefully they won't feel the need to go thin with the filtering on the new models, although they may.
I believe the 1Ds3 uses somewhat different filters than the 5D2.
The 5Dc definitely used truer filters than the 5D2, OTOH the microlenses and sensor were worse and not as efficient and with the filters cutting out more light as well it had worse SNR. The shadows on it were probably not better but the more noise meant you might not get to see the pattern banding as easily simply because you didn't feel like you could push the performance that far.
Note the 5D2 scores a full stop more DR than the 5Dc actually. Although note that DxO also doesn't get into pattern banding so that extra stop will definitely show the pattern banding so how much of the extra stop is comfortably usable is something to argue over (of course the 5Dc has some banding too). But despite the talk here the 5D2 has less random noise in the shadows than the 5Dc.
The 5D2 actually scores a bit higher for color tonal sensitivity at the lowest two ISOs (NOT when comparing both at 100%, only during normalized viewing, which granted is the way that makes sense, but don't get tricked by comparing both at 100% view) due to the higher SNR although with the shadow pattern banding and more uncomfortable peaks, if Suede is correct, the final story is not so simple and it can, all the same, have more accuracy trouble in some cases and muddle certain shadow colors more too and disagree with the eye's take on certain shades more than the 5Dc and 1Ds3. Complicated.
Beni wrote:
Is this due to their choice to keep the high iso better than on the same chip but with more accurate rendition in the 1Ds3? I know the Sony A900 people were saying that the high iso was weak specifically because they had optimised for incredible colour at the lower end of the iso spectrum.
Yes, the 1Ds3 is slightly different, in a way that I would say is better. It is a lot easier to coerce into giving "accurate" colour (than the 5D2).
About the A900/A850, the filters are only one small part of the equation. The main reason they don't do low-light shooting very well is.... Well, there's no other way of saying it - they suck at that part of sensor noise performance.
Colour filters can be tweaked to let more light through to the sensor, but the main reason for the not-so-brilliant high ISO behavior of the A900 is that it is quite a badly executed design. Not enough power feed stability, microlenses that aren't up to par with the stuff Canon and Nikon makes, and so on.
Sony has learned their lesson now though, the stuff they order for their current cameras are definitely top-of-the-class stuff. They skimp on some things that Nikon (and Pentax) emphasize a bit more to get the very last ounce of performance out of the base sensor. One does have to make the distinction between Sony Semiconductor and Sony Camera Division - they're part of the same company in the same way that Hitachi Kokusai (TV cameras) are part of the same company as Hitachi Construction machinery and Power tools.
Higher numbers mean that the colour saturation has to be amplified more. The number do in no way guarantee that the colours are accurate, it just states what you have to do to get the primary channels' saturation right. Higher numbers does hence mean very close to [number^2] more chroma noise.
But the hue accuracy and hue resolution isn't just about filter sharpness, it's also about how far apart you space the peaks. Peaks further apart: Better hue resolution, worse resistance to metamerism failure.
Personally I found it hard to tell the difference between 5dc and 5d2 files (shooting landscapes) apart from the obvious resolution difference. There is probably a discernable difference if you were to do detailed side by side comparisons but in the real world I doubt it really matters much. Buy the 5D2 for video, MUCH better LCD, better operating system, and higher resolution. I sold my 5dc as it was superflous.
The whole "banding problem" is way, way overdone in photography forums. The usual pattern is that someone brings it up, a few passionate participants rattle on about it, some number of photographers who use the camera point out that it virtually never affects actual photographs, the "passionate participants" become indignant and share (highly post processed) "example" files and rationalizations of why this is a Terrible And Awful Problem, names are called, insults are made, and those who think the whole brouhaha is quite silly give up and go away to make photographs with their 5D2...
I shoot extensively with the 5D2. I frequently push the dynamic range of the camera and I often have to lift shadows by any of several means in post. On occasion, when I'm doing post-processing work and staring very closely at the 100% magnification images on the screen, I too can locate what sometimes appears to be a faint banding effect in the normal noise.
This has turned out to be an actual problem in a real photograph - e.g. a large print - exactly zero times in my work.
I write this fully aware that the noise-banding enthusiasts are quite likely to jump all over me for this. Been there, done that, and I'm not about to engage in a lengthy forum battle over this. But when it comes to actually making photographs, as opposed to pixel peeping and fretting over anything that is less than God's Own Divine Camera Perfection, this is pretty much a complete non issue.
One additional point on the 5D v. 5D2 image quality issue. I have tens of thousands of image files from both cameras. I cannot ever tell which type of file I am opening just by looking at the "image quality" - in every case, unless I know by the date of the shot, I have to look at the image dimensions in my raw convertor to tell which camera was used. And in the end, I can produce equally good prints from files from either camera, with the exception of the effect of having more pixels to work with when the source file came from the 5D2. In the latter case, if everything was done with care at the time the shot was made, I can likely produce a somewhat larger print.
Gyroscope wrote:
Personally I found it hard to tell the difference between 5dc and 5d2 files (shooting landscapes) apart from the obvious resolution difference. There is probably a discernable difference if you were to do detailed side by side comparisons but in the real world I doubt it really matters much. Buy the 5D2 for video, MUCH better LCD, better operating system, and higher resolution. I sold my 5dc as it was superflous.
+1
Almost exactly mirrors my experience, with two additional comments.
1. I kept my 5D as my backup body.
2. You can also probably guess which camera a file came from by the larger number of dust spots on the 5D file. :-)
Yes I would have done the same and did for a while but I got the 7D as I shoot wildlife as well. While the 7D has been disapointing for noise reasons at high ISO I think it would serve me okay as a backup landscape camera in an emergency if used only at ISO 100.
The biggest bonus for me with the 5D2 which I hadn't expected was shooting landscapes in liveview which has really been great.
Gyroscope wrote:
Yes I would have done the same and did for a while but I got the 7D as I shoot wildlife as well. While the 7D has been disapointing for noise reasons at high ISO I think it would serve me okay as a backup landscape camera in an emergency if used only at ISO 100.
The biggest bonus for me with the 5D2 which I hadn't expected was shooting landscapes in liveview which has really been great.
yeah liveview is quite awesome
like a superior giant ground glass view camera view
chez wrote:
The subtle differences in colours are very hard to tell in a landscape shot. These differences on human faces are more apparent.
+1, I've always found human faces are the best subject to verify cam AF and it's rendering colors. why? because low contrast subject is the worse camera enemy
If your comment is based on viewing my landscape work, please be aware that you do not know my full body of work. I shoot human subjects as well and have done so using both of these cameras.
Dan
chez wrote:
The subtle differences in colours are very hard to tell in a landscape shot. These differences on human faces are more apparent.
Gyroscope wrote:
Yes I would have done the same and did for a while but I got the 7D as I shoot wildlife as well. While the 7D has been disapointing for noise reasons at high ISO I think it would serve me okay as a backup landscape camera in an emergency if used only at ISO 100.
The biggest bonus for me with the 5D2 which I hadn't expected was shooting landscapes in liveview which has really been great.
Regarding the 7D as backup to a 5D2, there are a number of photographers who agree with you! The two cameras have complementary strengths (and weaknesses) and depending upon the situation their positions in the primary/backup relationship might reverse.
