My 5D2 Has Low Iso Banding

toxic wrote in post #10586390
Actually, the solution isn't using ISO 400. That's a workaround to make up for Canon's mistake.

Why should anyone spend $2500, and then have to buy and mess around with ND filters instead of being able to use ISO 100? Especially for a vaunted portrait camera, where low ISO performance can be more important than high ISO. I guess that's why it's cheaper than the 5D was.

Of course, if you don't need to mess around with sync speeds or slow shutter speeds, then go ahead and use ISO 400 and don't worry about it. But that doesn't make it not a problem.

Right. The best solution is for Canon to manufacture the right electronics and ICs, issue a recall, and repair/replace folks 5D2s (and 7Ds) at their cost. It would be better customer service than saying, "screw you, the 5D2 is almost into year 3 of it's product life and we're going to hope we didn't screw up yet again with the 5D3 which you can buy to replace your flawed 5D2"

you really expect canon to "repair" something that is the operators fault lol

Poe wrote in post #10586154
I don't have a 5D2, so I can't supply.

just about sums it up perfectly

poah wrote in post #10588263
you really expect canon to "repair" something that is the operators fault lol

I would think it to be quite difficult to get banding in iso 100 shots. With that being said this isnt a problem found in all the 5dii's. But obviously in some. Else we would see a slew of comments about it when the camera first came out. I would suggest upon buying the camera to do some pixel peeping and see if you have one of the effected cameras. If so then exchange it until you get one that doesnt have it. I was just looking at my 50d pictures and couldnt find anything like this problem at iso 100. But I also dont understand why the 5dii didnt upgrade its af system. Its a good camera but it could have been so much better. People were expecting it to be a d700 killer and that did not happen.

poah wrote in post #10588263
you really expect canon to "repair" something that is the operators fault lol

Yeah, it's your fault if your operating system crashes when you leave your computer turned on for too long...

It's the photographer's fault for wanting to be able to dodge and burn, because as we all know they have never been able to do that before...

kcbrown wrote in post #10543026
After doing some significant searching, I'll be darned if I can find any two shots that show a clear difference between the bodies I was testing.

As it happens, even the closest shots I was able to find weren't close enough -- framing was different, or the lighting subtly different, or something.

Somehow, the difference was plain to me but I think I developed that impression as a result of taking a number of shots in different settings and using that to get a feel for the difference.

The next time I get the opportunity, I'll do a truly controlled test.

I did a controlled test of the second 7D I bought versus the first one. The first one had been to the Canon service center in Irvine twice and had the sensor assembly replaced during the second trip there. The second one has not seen any service by Canon at all. I am keeping the second one based on testing such as this.

Everything is identical: the target, the distance to the target, the framing of the target, the light on the target, the lens, the aperture, the shutter speed, the ISO (100), the color balance. Absolutely everything. Despite that, there appear to be subtle response differences between the two sensors. Even the color response was slightly different. The shots below were converted to grayscale in order to eliminate the color response difference between the two sensors.

The following test shots were converted from RAW files using ACR 6.1 and processed in Photoshop Elements 8. During conversion of the first two samples, I used all the defaults supplied by ACR save for blacks and contrast. ACR defaults to blacks of 5 and contrast of 25. I set both blacks and contrast to zero, presumably implying no adjustment of them within the converter. NO EXPOSURE ADJUSTMENTS HAVE BEEN PERFORMED TO THE FIRST TWO IMAGES. The banding is visible (if subtle) despite that. THE FIRST TWO IMAGES ARE WHAT YOU GET WITHOUT THE CONVERTER SUPPRESSING THE SHADOWS.

Here's a crop of the shot from the first camera (the one that twice went to Canon):

timbop wrote in post #10585076
I've posted several times the suggestion that owners could deal with the issue by "not going there" and shooting at mid ISO. You know, suggesting a technique that they could implement that would avoid their problem and let them persue a life of religious fulfillment. Kind of like learning how to use flash, or noise ninja, or whatrever.

that doesn't work for a LOT of situations though. Landscaping? your shutter speed just got significantly shorter. Yes you could use an ND to solve that problem in theory, but what if you were already using it? How 'bout the fast prime guys that shoot outside at times? 1.2 and 1.4 can be quite limiting at 100 with 1/8000th as the upper limit, take 2 stops away and you wouldn't be able to shoot wide open w/o that ND filter again...

And I shot Nikon for 9 months to get away from this. But their base ISO is 200. Yes you deal with it, but even the bump from 100 to 200 with a 1.4 lens is problematic. So going to 400 would just make it twice as bad. This is NOT a solution, it's a band aid.

but, I've also switched back for the 7D. It's the best all around camera available. Sure I wish it was FF, but that's the ONLY spec I don't like. Yes it's noisy at low ISOs, but you don't see it in prints. So I'm dealing with it as I dealt with noise on my 5Dii at base ISOs before. I'll live. But that doesn't make any of this any less canon's fault. These cameras have faults, it's crucial people know them so they can work with them.

I'm waiting for the day canon makes a studio FF body. I know most would say it's the 1DsIII, and it's close, but basically, concentrate on the low ISOs. Start at 50. Give us whole stops only. Stop at 800 and DO NOT offer "extended" range. Cut the bull**** and give us some amazing range w/o any noise. Keep the MP reasonable. the 5D (classic) and 1Ds and 1DsII are the closest we have to this IMHO. They were made in an "era" where 3200 was considered overly fast so they have amazing low ISOs and good high ISOs. But now, with the 5Dii and 7D, it's obvious in this level camera canon's making sacrifices at low ISOs to get the high ISOs looking better. Just like we don't really need 18mp, we don't really need native 6400 on an APS-C sensor, but the marketing department loves these numbers and pushes for them.

jacobsen1 wrote in post #10588592
Just like we don't really need 18mp, we don't really need native 6400 on an APS-C sensor, but the marketing department loves these numbers and pushes for them.

Heh. Speak for yourself.

I've made practical use of ISO 12800 before, shooting night baseball (high school) with my 70-300 IS. It was amazingly useful, though because it was the 50D, the ISO 12800 shots didn't come out all that well. The ISO 6400 shots came out okay after postprocessing, but they did take some work.

I clearly need a faster lens for that kind of work, but you use what you've got. The 7D in that situation would have done noticeably better than my 50D did.

I've also used ISO 6400 at a concert with my 85 f/1.8. I was at f/1.8 and shooting at 1/60th of a second, and still needed ISO 6400 to get the shots.


Yeah, I very much want native ISO 6400 on my sensor if I can get it, whether it be a crop sensor or full-frame. Fortunately, the ISO 100 quality of the 7D is still quite good despite the banding in the deep shadows.

kcbrown wrote in post #10588362
Yeah, it's your fault if your operating system crashes when you leave your computer turned on for too long...

It's the photographer's fault for wanting to be able to dodge and burn, because as we all know they have never been able to do that before...

nothing wrong with wanting to do a bit of PP - odd how I and countless others don't see any banding in any of our shots

thats probably why I never see any banding since I don't shoot black frames

kcbrown wrote in post #10588435
The 5D2's maze pattern noise is much, much worse to deal with than this.

jacobsen1 wrote in post #10588592
that doesn't work for a LOT of situations though. Landscaping? your shutter speed just got significantly shorter. Yes you could use an ND to solve that problem in theory, but what if you were already using it? How 'bout the fast prime guys that shoot outside at times? 1.2 and 1.4 can be quite limiting at 100 with 1/8000th as the upper limit, take 2 stops away and you wouldn't be able to shoot wide open w/o that ND filter again...
...

Come on, that's a specious argument. Portraits in broad daylight at 1.4 I can believe, but landscapes?

At least be honest about it: at this point continuing this thread has the singular goal of **** about canon. We get it. The horse is dead, so please put the whips away.

timbop wrote in post #10589490
Come on, that's a specious argument. Portraits in broad daylight at 1.4 I can believe, but landscapes?

you're kidding me right? I'd argue it makes a bigger difference with landscapes than it does ports. You can't fake long exposure times in PS (not that you can DOF easily either). I shoot a lot of water and rocks. Especially on a crop body where I shouldn't stop down as much (diffraction) you're saying I should also give up two stops for banding as well? I guess you don't like any blurred water in your landscapes then?

poah wrote in post #10589406
thats probably why I never see any banding since I don't shoot black frames

Those examples were primarily for jacobsen1's benefit, since he wanted to see evidence of copy variation between 7Ds.


You don't ever shoot twilight shots? Night shots? Shots of subjects against a dark background? Outdoor portraits involving shade and sunlight? Nothing of that sort? Hey, if you don't ever take shots that have tonal values identical to the ones I posted and don't ever have to do anything to those tonal values except make them even darker, then good for you.

If you do take shots like that, then you may be affected. Or you might not be. There's copy variation, after all.

Remember, the first two images I posted HAVE NOT HAD ANY EXPOSURE ADJUSTMENT. The banding is still visible anyway. All I did was drop the blacks to zero and eliminate the default contrast bump.

Now, if you always bump your black levels and/or increase your contrast, and thus depress the shadows even more than they are by default, and don't ever do anything to push the shadows at all, then sure, you won't ever see anything. But if you're shooting a 7D, you will if you do anything to push up shadow regions that contain those tonal values. That can include reducing the contrast (ACR allows you to go negative on the contrast slider), directly pushing the shadows, dodging, adding fill light, etc. All of those things have to be added together in order to get the total effect on the shadows, and it takes less than a stop of that (remember, in total) to make the banding plain. At least, on the 7D.


And why do you ignore the twilight shots that have been posted that show pattern noise without any exposure adjustment whatsoever?


I'm glad you're happy with the way your 5Dmk2 performs, and that you never see any issues such as these with it. I hope that remains the case, that you don't ever actually hit this problem.

But please don't insult our intelligence by claiming that because it's not a problem for you, it can't possibly be a problem for anyone, and that if it is then they must be doing something wrong.

poah wrote in post #10588263
you really expect canon to "repair" something that is the operators fault lol

It's not. It's canon's shotty QC.

poah wrote in post #10588266
just about sums it up perfectly

Just because I don't have a 5D2, doesn't mean that the banding doesn't exist. Others have already proved it exists. In this thread and this one for example: https://photography-on-the.net/forum/showthre​ad.php?t=750731

kcbrown wrote in post #10591386
I'm glad you're happy with the way your 5Dmk2 performs, and that you never see any issues such as these with it. I hope that remains the case, that you don't ever actually hit this problem.

But please don't insult our intelligence by claiming that because it's not a problem for you, it can't possibly be a problem for anyone, and that if it is then they must be doing something wrong.

Ignorance is bliss, some say.

and here, I thought banding was what they put around fat people's stomachs.