Need Advice: Noise Issues on a Canon 7D

[sm4him]

First of all: NO, I have not switched sides, nor do I have any plans to do so! I'll stick with my Nikons, thank you very much.

I have a friend who has just gotten into photography in the last year. She is a birder and wants to get good pictures of the birds--which is how we met.
I've given her advice in general on bird photography, but am hoping for some advice on an issue she is having with her camera.

She has a 7D and a 400mm lens. She is getting a LOT of noise issues, sometimes even at very low ISO settings.
I've tried to help her troubleshoot this but so far the only thing I've been able to determine is:

--It's not due to cropping, adjusting exposure or other processing issues.

So now, I'm trying to figure out whether it's technique or the camera. After some internet searching, I'm seeing a lot of posts about noise issues with the 7D, but the ones I've read have all been very mixed messages as far as whether the problem was really the camera or not.

I have seen some of these photos myself and she's right, some of them are VERY noisy, unacceptably so for low ISOs (400 and lower, even). But it's not producing *consistently* noisy pictures--a lot of them are, but then some of them seem just fine, even at ISO settings of 800 or 1000.
So, I'm really not sure whether the problem is something in her technique (settings) or a problem with the camera. The camera is still under warranty, so she could probably get a replacement if needed, but I'd like to be a little more confident that the camera is even the problem here.
She has already been on the phone with Canon twice, the last time for nearly 3 hours trying various things and sending them sample photos--they are telling her it's NOT the camera, yet nothing they've told her to adjust has seemed to resolve the issue.

So, Canon Users:

1) Is there a known issue with this noise problem at low ISOs on the 7D? If there is, I'll probably just advise her to insist on a replacement.

2) Any suggestions on what might be causing the issue that I could test? I plan to meet up with her at some point soon and switch cameras--let her use mine and I'll use hers, to see what kind of results we get. But since I don't use Canons, I'm really not sure what menu settings to try adjusting as a "test."

Thanks for any help!

 

[Gavjenks]

Is your "friend" doing long exposures? Even at 100 ISO, a 5 minute long exposure will be pretty noisy if you don't correct for it at all, in any brand or sensor. Just an issue of electricity running through the sensor for too long and heat it up, then thermal noise.

That's the only thing I can think of that would explain it off the top of my head. If the sensor were simply busted, then it doesn't seem like you would be getting clean shots ever at 800, etc.



Maybe shoot a whole bunch of images in a row with the same scene and settings: does noise change between them?

 

[sm4him]

Is your "friend" doing long exposures? Even at 100 ISO, a 5 minute long exposure will be pretty noisy if you don't correct for it at all, in any brand or sensor. Just an issue of electricity running through the sensor for too long and heat it up, then thermal noise.

That's the only thing I can think of that would explain it off the top of my head. If the sensor were simply busted, then it doesn't seem like you would be getting clean shots ever at 800, etc.



Maybe shoot a whole bunch of images in a row with the same scene and settings: does noise change between them?

:lmao: I *knew* I was gonna get that "friend" comment!!

Nope, no long exposures. If anything, just the opposite. Most of these images are somewhere between 1/800 sec. and 1/2000 sec. exposure times. She shoots almost exclusively birds. That's what is so WEIRD. She'll get two or three pretty decent shots (at least in terms of noise) mixed in with 8-10 really noisy ones.
I'm NOT sure what, or how much, she is adjusting in-camera in between shots, though.

Good plan to shoot a bunch of the same scene and settings and see if the noise varies significantly; I'll do that as one of my tests when we're able to meet up.

 

[Dao]

Are the photos usually underexposed and she pull them back in POST?

 

[sm4him]

Are the photos usually underexposed and she pull them back in POST?

No. That was my initial thinking was that she was underexposing, and that between correcting the exposure and cropping in post, a lot of noise was being introduced. But I've had her show me some that were not processed, other than converting them to jogs--the noise is there in some of them, not in others.
To be sure, I DO think sometimes it's being made WORSE by cropping or adjusting exposure, but the problem exists independent of that.

 

[pixmedic]

Just tell your friend to switch to Nikon. problem solved! :thumbup:

 

[sm4him]

Just tell your friend to switch to Nikon. problem solved! :thumbup:

I've tried that! You just can't talk ANY sense into SOME people! :lmao:

 

[Gavjenks]

Example images would help if at all possible.

The only real sources of noise are:
1) Thermal, addressed above (long exposures. Extremely rapid and prolonged can cause more noise too, like holding the shutter for many seconds on end)
2) Digital conversion errors (unlikely to change in degree from shot to shot as you are describing. Although perhaps POSSIBLE with some bizarre software bug, you should avoid concluding that too soon.)
3) Amplification side effects, i.e. ISO causing other types of noise to be exaggerated. But you're already of course taking ISO into account. This also includes cropping, which is essentially amplification by allowing you to see smaller flaws, but you also said you saw it in unedited (presumably uncropped?) frames, so not that.
4) Electronic interference. I dunno, is her specialty MRI machine photography? Does she live directly under high tension power lines? Lol. This seems unlikely to change noticeably shot to shot except in ridiculous circumstances.
5) Quantization noise - I.e. shooting dark and then brightening versus shooting right and then darkening, "expose to the right" etc. Quite possibly involved, but already addressed above and you said no.
6) Simply noise inherent in the image. Due to photons not being emitted perfectly uniformly from a typical light source. There would be no reason for this to change shot to shot, though, unless again your friend is in extraordinarily weird circumstances. Like, for example, taking some shots in sunlight and other shots purely with collimated laser light... =P

7) Seemingly the most likely at this point: one of the things above, but changing from shot to shot due to her having some sort of auto setting that the camera is deciding to change from shot to shot. For example, if the camera were to do something like auto histogram centering when/if it can't get a proper exposure, in full auto mode? I don't THINK the 7D does that, but it might. I don't own one. Or something similar. In that example, you might think her exposure was correct looking at the "unedited" (but actually already edited by the camera) image, while in fact it was too low and pulled up, causing quantization noise.


I would:
* Have her go through all her settings and turn off everything automatic in every sub menu, then switch to manual mode on the dial as well.
* Find an example of a mysteriously high noise image and a good one. Seek out similar environments / subject content to those two shots for testing.
* In the environment from the noisy shot, shoot a series of controlled images of subject at the same time, with varying ISOs and so forth. Shoot several frames at each setting.
* In the environment from the non-noisy shot, do the same thing.
* Look for any inconsistencies. If a single setting shows inconsistency, then it may indeed be a software bug or something wrong with the camera. If not, then probably not. If you see a strange but consistent pattern of noise with certain settings, then try to consider why that might be, and/or post here about what they are.

 

[Overread]

Odd it shouldn't generate noise randomly - the only thing that sparked to my mind was if she was using Auto ISO and the camera is picking in-between ISOs which are oft just under/over exposed from a normal full stop ISO and then bumped up/down as needed (and thus can be more noisy than a normal ISO).

I do recall some high ISO discussion with the 7D but it gets horribly muddled between beginners under exposing and people wanting fullframe quality or comparing it to Nikon.

She could try comparing it to another 7D in case she does have some software or internal problem - the other options are getting Canon to look at it diret or getting a 5DMIII and enjoying great low noise and cropping a bit.


As an aside is this noise a 100% views? I know a fair few people get scared by software/noisier images at 100% view when they changed up from 12mp to 18 or higher mp cameras and its not the camera that is at fault, but rather that the larger MP size means that a 100% view is even more enlarged. I tend to view at around 60% with my 7D as I find that most similar to what I was used to with the 400D in terms of visual quality.

Note the 7D did require some shift in how I edited to get the most out of it. All cameras are like this and have their own little quirks.

 

[jaomul]

The 7d is sometimes said to be not great on its blue channel. If you Google blue channel noise and 7d I think you will get sone hits. I had a 7d and generally found it Good at hi iso but some shots I have with birds and blue skies were noisy at lower iso than I would have expected

 

[JacaRanda]

Just tell your friend to switch to Nikon. problem solved! :thumbup:

I've tried that! You just can't talk ANY sense into SOME people! :lmao:

:banghead: layball::blackeye:

 

[Gavjenks]

Oh right color channel too. That's not just the 7d, that's anything, but yeah. If the noisy images are all really one color and the non noisy ones are another, then that could be it: different noise in different channels due to sensitivity to wavelengths being different blah blah

 

[sm4him]

Example images would help if at all possible.

The only real sources of noise are:
1) Thermal, addressed above (long exposures. Extremely rapid and prolonged can cause more noise too, like holding the shutter for many seconds on end)
2) Digital conversion errors (unlikely to change in degree from shot to shot as you are describing. Although perhaps POSSIBLE with some bizarre software bug, you should avoid concluding that too soon.)
3) Amplification side effects, i.e. ISO causing other types of noise to be exaggerated. But you're already of course taking ISO into account. This also includes cropping, which is essentially amplification by allowing you to see smaller flaws, but you also said you saw it in unedited (presumably uncropped?) frames, so not that.
4) Electronic interference. I dunno, is her specialty MRI machine photography? Does she live directly under high tension power lines? Lol. This seems unlikely to change noticeably shot to shot except in ridiculous circumstances.
5) Quantization noise - I.e. shooting dark and then brightening versus shooting right and then darkening, "expose to the right" etc. Quite possibly involved, but already addressed above and you said no.
6) Simply noise inherent in the image. Due to photons not being emitted perfectly uniformly from a typical light source. There would be no reason for this to change shot to shot, though, unless again your friend is in extraordinarily weird circumstances. Like, for example, taking some shots in sunlight and other shots purely with collimated laser light... =P

7) Seemingly the most likely at this point: one of the things above, but changing from shot to shot due to her having some sort of auto setting that the camera is deciding to change from shot to shot. For example, if the camera were to do something like auto histogram centering when/if it can't get a proper exposure, in full auto mode? I don't THINK the 7D does that, but it might. I don't own one. Or something similar. In that example, you might think her exposure was correct looking at the "unedited" (but actually already edited by the camera) image, while in fact it was too low and pulled up, causing quantization noise.


I would:
* Have her go through all her settings and turn off everything automatic in every sub menu, then switch to manual mode on the dial as well.
* Find an example of a mysteriously high noise image and a good one. Seek out similar environments / subject content to those two shots for testing.
* In the environment from the noisy shot, shoot a series of controlled images of subject at the same time, with varying ISOs and so forth. Shoot several frames at each setting.
* In the environment from the non-noisy shot, do the same thing.
* Look for any inconsistencies. If a single setting shows inconsistency, then it may indeed be a software bug or something wrong with the camera. If not, then probably not. If you see a strange but consistent pattern of noise with certain settings, then try to consider why that might be, and/or post here about what they are.

Some good suggestions there. Rather than trying to get HER to do these though, my plan is to switch cameras and do them myself, or at least WITH her. That way, *I* can control what is getting done and what's not. For instance, she SAYS she isn't adjusting exposure in post…but is it at all possible that she's doing some sort of "auto correct" and not even realizing that affects exposure?

 

[Overread]

The 7D does have some auto exposure correction features but that should only be affecting JPEGs not the RAW files.

 

[sm4him]

Odd it shouldn't generate noise randomly - the only thing that sparked to my mind was if she was using Auto ISO and the camera is picking in-between ISOs which are oft just under/over exposed from a normal full stop ISO and then bumped up/down as needed (and thus can be more noisy than a normal ISO).

I do recall some high ISO discussion with the 7D but it gets horribly muddled between beginners under exposing and people wanting fullframe quality or comparing it to Nikon.

She could try comparing it to another 7D in case she does have some software or internal problem - the other options are getting Canon to look at it diret or getting a 5DMIII and enjoying great low noise and cropping a bit.


As an aside is this noise a 100% views? I know a fair few people get scared by software/noisier images at 100% view when they changed up from 12mp to 18 or higher mp cameras and its not the camera that is at fault, but rather that the larger MP size means that a 100% view is even more enlarged. I tend to view at around 60% with my 7D as I find that most similar to what I was used to with the 400D in terms of visual quality.

Note the 7D did require some shift in how I edited to get the most out of it. All cameras are like this and have their own little quirks.

She's not using Auto ISO, to my knowledge, at least not for the most part. However, I can't guarantee that isn't where the problem is. Perhaps she DOES sometimes use it and THAT's when the noise issue is happening. That'll be a good thing for me to test as well.
I *do* know that it's not a matter of 100% views. It's clearly visible, even on a zoomed out Flickr image. I'd put a link to her flickr page, but it wouldn't really help, as many of those ARE cropped and otherwise adjusted so even where there are noise problems, there's no way for any of you to know whether those particular photos had noise issues to start with or not.

I think perhaps, if I can't help her figure it out by using the camera myself for some tests, I'll suggest that she rent another one just for comparison.

The 7d is sometimes said to be not great on its blue channel. If you Google blue channel noise and 7d I think you will get sone hits. I had a 7d and generally found it Good at hi iso but some shots I have with birds and blue skies were noisy at lower iso than I would have expected

Oh right color channel too. That's not just the 7d, that's anything, but yeah. If the noisy images are all really one color and the non noisy ones are another, then that could be it: different noise in different channels due to sensitivity to wavelengths being different blah blah

Huh. THAT is certainly something I hadn't even considered. VERY good "food for thought." So, I'll add that to the list of "tests" to run. Thanks!