Can we stop now with the photos that were hugely underexposed than pushed in post... and the presentation of these photos (such as that of the boy above) as if they were typical of 5D2 results? This is really, really, really getting old now.
I've shot many thousands of frames on both cameras - OK, many tens of thousands of frames - and this whole linear noise in the shadows thing is really starting to look like an obsession for you brainiac. The _only_ time I think about it is when I see another one of these posts.
I've pushed underexposed images on both cameras and both will look awful if you push them enough.
Live view at 10X for precision focusing was alone worth the upgrade from 5D to 5D2 for me. I am not aware of giving up anything in making the transition (except for a bunch of $).
DocsPics wrote:
Live view at 10X for precision focusing was alone worth the upgrade from 5D to 5D2 for me.
This is for sure true. If anyone is using a 5D2 on a tripod and has not yet discovered Live View, you are missing a definite treat.
By the way, in the images I posted above, the 5D image was focused automatically--and it missed! The 5D2 image was focused manually via Live View and the 10x loupe.
I've pushed underexposed images on both cameras and both will look awful if you push them enough.
I agree with you.
Back in the film days, the truth of "pushing" was that no matter how long you extended development, you could never gain detail in shadows where the exposure was simply below the level of silver halide crystals being struck by enough photons. Where there was no detail, there was no shadow detail--pushing only raised density in the negative where there was sufficient detail.
But when you extended development, you did get something called "base fog"--silver halide crystals that had never met a photon would be forced into development anyway. That gave a false density where there should have been no density and actually swallowed up very delicate detail that could not have been printed anyway.
If a photographer tried to print the base fog as though it were actual pictoral density, all he got was ugly, detail-lacking gray...something the eye never sees naturally. So the lesson was: If a shadow lacks detail far enough above the base fog, don't try to save it--print it down to black. The eye sees, understands, and accepts black shadows without detail--those are natural. The eye does not accept light shadows without detail.
In digital imaging, noise is analogous to base fog. You don't see noise in even a flat tone that is exposed enough to raise it above the noise floor. The basic rule is the same as with film. If the shadow lacks detail far enough above the noise, don't try to save it--print it down to black. The eye understands black shadows, the eye accepts black shadows.
If you can't put light into the shadows, print them down to black.
Dan and RDKirk, are you saying that the huge dynamic range available in Digital Medium Format and Nikon D3x is a waste, and should never be utilized by anyone?
Does this look underexposed to you. Camera was on AV plus +1/3 stops in day light at the time. I don't see many other cam has this funny noise at low iso. There was plenty plenty light here gentlement. Unless, my 2nd 5D2 is also a bad one (sold the first). Note that, i've always shot with NR disabled
Breitling65 wrote:
Besides mpx, video, dust, micro-adjustments do you think overall 5D2 better in IQ camera than classic 5D?
The MPs, dust and micro adjustments that you mentioned make more detailed, cleaner and sharper images, making the overall image quality better
The images from 5D2 look different than 5D. But I don't really know if this is function of the camera or ACR, so I can't answer the question. The 5D2 is incrementally better in almost every way. Better battery life, better AF, manual live view focusing at 10x, resolution, ability to dial very high ISO, better display, AF micro adjustments (gave my 35/1.4 new life), sensor cleaning and many, many others. I even like the IR remote functionality sometimes. It is incrementally better and so is the chance that I end up with usable files.
alundeb wrote:
Dan and RDKirk, are you saying that the huge dynamic range available in Digital Medium Format and Nikon D3x is a waste, and should never be utilized by anyone?
No, didn't say that anywhere. The dynamic range at the bottom of a digital image stops somewhere within the noise--it gets lost in the background.
As has been written numerous times, the lower limit of dynamic range is often a matter of how much noise a particular photographer can tolerate. The pictorial answer is always "print the offending noise in shadows down to black," and in doing so, the practical dynamic range of the sensor is revealed by the noise tolerance level of the photographer.
alundeb wrote:
Dan and RDKirk, are you saying that the huge dynamic range available in Digital Medium Format and Nikon D3x is a waste, and should never be utilized by anyone?
Not sure if I'm the Dan you refer to, but...
The DX3 has a slightly larger dynamic range, but those who test and analyze these things recognize that a) the difference is trivial, and b) you would be hard pressed to see it in most if not all photographs. Some folks make a big deal about such small differences because a) they focus more on test parameters than on photographic results, and b) they are looking for ANY difference between different pieces of gear so that they have something to write about.
Basically the Nikon and Canon (and Sony) FF bodies all provide a very good dynamic range. The way that the image fails at the extremes is different than the way it failed with film, but once you figure that out and work with it the results are great.
MF is another story, but perhaps not in the way you might think. (Do you use MF digital?) The real advantage here is going to be resolution for very large prints primarily. Certainly the MF digital backs - at least operating at full resolution - don't handle high ISO as well as the DSLRs do, they have some serious focusing issues that you must deal with both while shooting and while setting up the camera, and there are other downsides such as shorter battery life, size, range of lenses available... oh, and price.
MF digital is a great solution for certain types of photographers - I don't shoot it but I have friends who do.
gdanmitchell wrote:
Can we stop now with the photos that were hugely underexposed than pushed in post... and the presentation of these photos (such as that of the boy above) as if they were typical of 5D2 results? This is really, really, really getting old now.
I've shot many thousands of frames on both cameras - OK, many tens of thousands of frames - and this whole linear noise in the shadows thing is really starting to look like an obsession for you brainiac. The _only_ time I think about it is when I see another one of these posts.
I've pushed underexposed images on both cameras and both will look awful if you push them enough. ...Show more →
The peculiarities of the 5D2 shadow noise at low isos is a perfectly reasonable thing to talk about and illustrate on a thread about the differences between the 5D and 5D2. While you may not have come across this limitation of the 5D2, others have, not just myself.
The picture of the boy reading the speech is perilously close to overexposure in parts of his white shirt. The 5D2's noise pattern can become intrusive when you need to maximise DR in coping with highly contrasty subjects. When you look at the detail in the boy's hair, it is clear that one would have to throw away plenty of recorded information in darkening the background to black. Anyway, the background was not black, and I didn't want my picture to pretend that it was.
I use two 5D2's and they are my camera of choice at the moment, but this is a real limitation of the camera, and I really don't see any need to keep quiet about it. We come here to share. It's a great camera, but it is far from perfect, and it has some issues that the 5D doesn't, while generally being much better.
The picture of the boy reading the speech is perilously close to overexposure in parts of his white shirt. The 5D2's noise pattern can become intrusive when you need to maximise DR in coping with highly contrasty subjects. When you look at the detail in the boy's hair, it is clear that one would have to throw away plenty of recorded information in darkening the background to black. Anyway, the background was not black, and I didn't want my picture to pretend that it was.
Few shadows are truly black--that is to say, there aren't very many true "black bodies" in nature. But a heck of a lot of scenes are beyond the dynamic range of the sensors, and that scene is one of them. Look at the edges of the shadows cast by the main light. That's a stage light, for pete's sake, and you're trying to capture detail in shadows off-stage.
You show that crop a good three stops pushed over the full image. In the full image, the shadow in her hair behind her neck is about Zone 2, while you've got lightened close to Zone 5 in the crop.
In the full image, you've done precisely what I said should be done--you've printed the shadow down to near black--in the crop, you've raised it enough to reveal the noise.
gdanmitchell wrote:
Not sure if I'm the Dan you refer to, but...
The DX3 has a slightly larger dynamic range, but those who test and analyze these things recognize that a) the difference is trivial, and b) you would be hard pressed to see it in most if not all photographs. Some folks make a big deal about such small differences because a) they focus more on test parameters than on photographic results, and b) they are looking for ANY difference between different pieces of gear so that they have something to write about.
Basically the Nikon and Canon (and Sony) FF bodies all provide a very good dynamic range. The way that the image fails at the extremes is different than the way it failed with film, but once you figure that out and work with it the results are great.
MF is another story, but perhaps not in the way you might think. (Do you use MF digital?) The real advantage here is going to be resolution for very large prints primarily. Certainly the MF digital backs - at least operating at full resolution - don't handle high ISO as well as the DSLRs do, they have some serious focusing issues that you must deal with both while shooting and while setting up the camera, and there are other downsides such as shorter battery life, size, range of lenses available... oh, and price.
MF digital is a great solution for certain types of photographers - I don't shoot it but I have friends who do.
What you say makes good sense, but I don't get this: While you say that there are limitations in DR you have to deal with, why don't you then accept that with better DR, in some sitautions the results can be better (more shadow detail can be made visible), and some people want just that? Who can say that 11 stops is all the DR photographers ever want to capture?
I don't agree that a 1-2 stop difference in DR is trivial when working with shadow detail. It is what it is. 1-2 stops.
My experience limits to 350D, 40D, 5D and 500D. The 350D is horrible for shadow recovery. The 5D is sometimes good, sometimes banding shows up just at 1 stop. With the 40D and 500D I can comfortably push 2 stops in all situations I have encountered.
RDKirk wrote:
You show that crop a good three stops pushed over the full image. In the full image, the shadow in her hair behind her neck is about Zone 2, while you've got lightened close to Zone 5 in the crop.
In the full image, you've done precisely what I said should be done--you've printed the shadow down to near black--in the crop, you've raised it enough to reveal the noise.
Yes - all Canon cam have some shadow noise at low iso but not as bad as 5d2 sensor that what i try to say here. Most my 5D2 headshots have hair (sometime skintone) colours issues while 5D and 1Ds2 doesn't not. Here is L/R angles of the same image. Note that, this has +1/3 stops from AV mode in day light
I went back to look at the speech file to see what extreme processing I had done to it. It was shot with highlight control which amounts to about a stop of noise, so call it iso 1600. It is also pushed a stop and a half or so: http://cyberphotographer.com/5d2/speechcrisscrosspalettes.jpg
It's not even on minimum contrast. The intrusive criss-cross pattern is going to spoil a print at any size. My 1Ds3 never failed in this way. Don't get me wrong, the 5D2 is a good camera, but its files can't take manipulation in the way that some other Canon bodies' files can.
brainiac wrote:
Don't get me wrong, the 5D2 is a good camera, but its files can't take manipulation in the way that some other Canon bodies' files can.
+1, This is all about - and it does kill portrait photographer during his/her post work
(sold the first). Note that, i've always shot with NR disabled