Canon - dynamic range / banding. Future?

Messages
198
Name
Chris Richards
Edit My Images
No
With the release of the 7D2, it will be interesting to see if any work has been done to increase the dynamic range.

The 5D3 suffers from banding - and the 5D2 even more so. So, is this something that Canon are actively trying to improve?
 
Know what you mean but think that the banding only occurs if you underexpose by quite a lot
dynamic range is important though I agree I think especially for wildlife as you sometimes have to shoot in difficult lighting
in landscape you can work round it by bracketing
 
Know what you mean but think that the banding only occurs if you underexpose by quite a lot
dynamic range is important though I agree I think especially for wildlife as you sometimes have to shoot in difficult lighting
in landscape you can work round it by bracketing

Yeh Pete - it mainly occurs when an image is underexposed.

Dynamic range is so blimin important :(
 
With the release of the 7D2, it will be interesting to see if any work has been done to increase the dynamic range.

The 5D3 suffers from banding - and the 5D2 even more so. So, is this something that Canon are actively trying to improve?

Banding on the raw, straight from camera? Shouldn't be doing.. it's still a 14bit raw, and the dynamic range has nothing to do with banding anyway.. banding is a colour depth issue, usually caused by poor processing.
 
Banding on the raw, straight from camera? Shouldn't be doing.. it's still a 14bit raw, and the dynamic range has nothing to do with banding anyway.. banding is a colour depth issue, usually caused by poor processing.

Whatever the reason or phrase, it is a well known fact that the 5d2s / 5d3s suffer from banding when you attempt to drag detail out of the shadows. If you've every processed Nikon images (from equiv camera bodies) you'll find the images much more flexible. Have a Google.
 
User error.. nothing to do with the camera. A) Stop under exposing, and B) stop over processing your images.

There's nothing wrong with the 5D MkII or III.

Sure... I can Google it.. I'll just find posts by people who don;t know what they're doing, and blaming the gear for their own incompetence.
 
This is not about exposing correctly. Any image can consist of bright areas, mid tones and dark areas. So unless you can expose for all in one exposure, the comments re exposure are pointless.

It is a known fact that 5d2 and 5d3 struggle to drag details out of shadows - and exhibit banding. Nikon don't seem to have the same problems. You can't argue this. Fact.
 
nah - just fix it in post ;)
ok I'll get me coat ... :coat:

Armour, maybe :D My point was that it's just not as easy as that.

And... abuse a Canon and another camera in the same way and you're arguably much more likely to see banding in the Canon shots but of course you're doing something wrong and there's obviously nothing wrong with the Canon. Perish the thought and just ridicule the incompetent complainant.
 
This is not about exposing correctly. Any image can consist of bright areas, mid tones and dark areas. So unless you can expose for all in one exposure, the comments re exposure are pointless.

It is a known fact that 5d2 and 5d3 struggle to drag details out of shadows - and exhibit banding. Nikon don't seem to have the same problems. You can't argue this. Fact.


Ok then. If getting lights, darks and shadows all exposed correctly I guess you'll have problems regardless of make or model
 
Ok then. If getting lights, darks and shadows all exposed correctly I guess you'll have problems regardless of make or model

That's the problem. You don't. Canon's tend to be behind the best of the rest at the moment.

Granted it's an issue you may be more likely to see if you expose for the highlights and then boost the shadows post capture but with other brands this is not so much of an issue so just blaming the user isn't really a fix all, is it? It's one way of dealing with it. Another is to buy a Nikon. However, people with a bag full of Canon kit and those who just prefer some aspect of Canon kit and want to stick with it would like to see Canon nearer the competition in this respect.
 
That's the problem. You don't. Canon's tend to be behind the best of the rest at the moment.

Granted it's an issue you may be more likely to see if you expose for the highlights and then boost the shadows post capture but with other brands this is not so much of an issue so just blaming the user isn't really a fix all, is it? It's one way of dealing with it. Another is to buy a Nikon. However, people with a bag full of Canon kit and those who just prefer some aspect of Canon kit and want to stick with it would like to see Canon nearer the competition in this respect.

Canon may be behind at the moment , but if you're recovering in any brand you're losing quality fast, with no reason to do so. So yes, ultimately it's a user issue
 
This is not about exposing correctly. Any image can consist of bright areas, mid tones and dark areas. So unless you can expose for all in one exposure, the comments re exposure are pointless.

It is a known fact that 5d2 and 5d3 struggle to drag details out of shadows - and exhibit banding. Nikon don't seem to have the same problems. You can't argue this. Fact.

Not fact.

I've done extensive dynamic range tests on both 5D2 and 5D3, no banding on either. Neither have I heard this 'fact' before.
 
Canon may be behind at the moment , but if you're recovering in any brand you're losing quality fast, with no reason to do so. So yes, ultimately it's a user issue

You keep saying it's a user issue so how do you recommend the user solve the problem? And "Buy a Nikon" isn't a valid answer :D

I suppose the numpty user should take multiple exposures and then combine them post capture. Ah! Silly of me to overlook this simple answer. Doh! So what if a Nikon can get the shot in one take, Canon's are like fine wine and need to be appreciated over multiple shots. It's just not an issue is it? :D
 
You keep saying it's a user issue so how do you recommend the user solve the problem? And "Buy a Nikon" isn't a valid answer :D

I suppose the numpty user should take multiple exposures and then combine them post capture. Ah! Silly of me to overlook this simple answer. Doh! So what if a Nikon can get the shot in one take, Canon's are like fine wine and need to be appreciated over multiple shots. It's just not an issue is it? :D

:banana::clap:


I think Hugh is either being awkward OR has just completely missed the point.
 
You keep saying it's a user issue so how do you recommend the user solve the problem? And "Buy a Nikon" isn't a valid answer :D

I suppose the numpty user should take multiple exposures and then combine them post capture. Ah! has silly of me to overlook this simple answer. Doh! So what if a Nikon can get the shot in one take, Canon's are like fine wine and need to be appreciated over multiple shots. It's just not an issue is it? :D


I don't know if that's a serious question or not? You're seriously saying you don't know how to get a correct exposure regardless of brand?

Do Nikon users not need flashes, grads etc in some situations anymore? Are they magically better photographers? I think not. Seriously if someone is unable to get the exposure they want then they have bigger problems then changing brand would solve.

At a guess, I've processed maybe 12,000 files from various flavours of 5d in the last three years. All from various 2nd shooters who are all at various standards. I can safely say no banding has ever been an issue or even noticed when you expose as you intend
 
I don't know if that's a serious question or not? You're seriously saying you don't know how to get a correct exposure regardless of brand?

Do Nikon users not need flashes, grads etc in some situations anymore? Are they magically better photographers? I think not. Seriously if someone is unable to get the exposure they want then they have bigger problems then changing brand would solve.

At a guess, I've processed maybe 12,000 files from various flavours of 5d in the last three years. All from various 2nd shooters who are all at various standards. I can safely say no banding has ever been an issue or even noticed when you expose as you intend

You're being deliberately vague and even obtuse aren't you?

Look, yes Canon's are nice cameras and yes it's possible to get lovely pictures out of them and yes there are workarounds with this alleged banding issue but the fact is that some of the competition just doesn't display this issue to the effect that some Canon's do.

And I'm out...
 
You're being deliberately vague and even obtuse aren't you?

Look, yes Canon's are nice cameras and yes it's possible to get lovely pictures out of them and yes there are workarounds with this alleged banding issue but the fact is that some of the competition just doesn't display this issue to the effect that some Canon's do.

And I'm out...


Nope, just lost as to why people think exposing correctly is a big deal? :).
 
If you need 14 stops of DR you can use dual iso with magic lantern. In 6D canon improved sensor and I have to push deep shadows really hard to nonotice any banding.
 
Think that the point is that for most of the time it's not an issue and you just have to get the exposure correct or bracket the exposure and blend if the scene has a wide dynamic range
But just occasionally certainly for wildlife you have to shoot in bad light with high contrast and you can't get the whole scene exposed correctly
To be honest tho if the light isn't good the shot probably wasn't worth taking anyway and having a "better" camera would have not made much difference
I've never had banding, though I don't have a 5D, I use a 550D and 7D
 
There's no such thing as "expose correctly" only "expose as the photographer intended."
 
it is a known and common problem regarding the canon SNR.
http://www.fredmiranda.com/5DIII-D800/index_controlled-tests.html
just one example of a 100 posts if you search for it

That review is a couple of years old now, and it's puzzled me ever since.

Whatever is going on there, I wouldn't call it a banding problem - just normal shadow noise. Other comparisons of the 5D3 vs D800, including my own and also with the 5D2, don't show that to anything like the same extent. All I can think of is that something has happened in the 'normalising' of the Canon file when it was interpolated up to match the Nikon's resolution. The other thing is what we're looking at there are huge enlargements, about 12-14x for the small sections. On my screen now, that's a section from a print 10-12 feet wide and that reveals details not visible, or not significant, at normal output sizes.
 
This is not about exposing correctly. Any image can consist of bright areas, mid tones and dark areas. So unless you can expose for all in one exposure, the comments re exposure are pointless.

It is a known fact that 5d2 and 5d3 struggle to drag details out of shadows - and exhibit banding. Nikon don't seem to have the same problems. You can't argue this. Fact.

That's not the cause of banding though.

BUT you get banding in the Canon system

If you mistreat the images, yes... as you will from any system.
 
Just seen that linked test... You're referring to the noise pattern in the shadows... that's not really what most people think of as banding. Banding is when you get marked "bands" in smooth gradients in skies etc... which is a bit depth issue.

canon-mk3-2.jpg



Seriously.... if you're recovering THIS much shadow detail and expect it to look nice, you have very high expectations. That's a VERY contrasty scene. No camera can pull this off without a penalty, even the D800 will look rough here (and does).. just not as bad as the Canon. However... yes, this IS a dynamic range problem, but calling it banding is very misleading. It's just noise caused by too much recovery of shadow detail that's right down at the bottom end of the sensor's range.
 
But, it is the degree of mistreatment. The Canon files are less mailable.


After reading the test, I've acknowledged the problem is indeed dynamic range... see above. It was the OP's choice of name that threw me, and probably others. This is not banding... this is just a noise issue.
 
After reading the test, I've acknowledged the problem is indeed dynamic range... see above. It was the OP's choice of name that threw me, and probably others. This is not banding... this is just a noise issue.

Here is why it is called banding. This is a dark frame - no lens, just the lens cap - raw file opened in Lightroom with the original on the left and a +2 stop exposure boost on the right. Clearly the noise is not random, but arranged into vertical stripes or "bands"....

Capture-2.JPG


The visible presence of bands makes the noise more intrusive and less easily removed than with cameras with far less noise or truly random noise.

The 5D3 is better, but not entirely free from the problem....

Capture-3.JPG


I don't have a sample handy, but from memory the 5D2 is closer to the 7D than the 5D3.
 
Banding on the raw, straight from camera? Shouldn't be doing.. it's still a 14bit raw, and the dynamic range has nothing to do with banding anyway.. banding is a colour depth issue, usually caused by poor processing.

It depends on the dynamic range of the display medium, noise levels and the bit depth of the channels.
 
Here is why it is called banding. This is a dark frame - no lens, just the lens cap - raw file opened in Lightroom with the original on the left and a +2 stop exposure boost on the right. Clearly the noise is not random, but arranged into vertical stripes or "bands"....

Capture-2.JPG


The visible presence of bands makes the noise more intrusive and less easily removed than with cameras with far less noise or truly random noise.

The 5D3 is better, but not entirely free from the problem....

Capture-3.JPG


I don't have a sample handy, but from memory the 5D2 is closer to the 7D than the 5D3.

I think something else may be going on here because if the "banding" (not the right term to apply to this) was this bad the camera would be almost unusable.

And I certainly wouldn't expect to see this at 100 ISO - I regularly use between 800-3200 on my 1Ds MkII and have never seen this even at 100% crop.
.
.
 
Last edited:
This is a 100% crop with shadow clipping warnings enabled. Since the sensor received no light the image on the left should be a solid blue tone. All the non blue parts have some sort of signal, all of which was fabricated by the sensor and associated electronics..

The other sort of banding which some people are referring to is posterisation, which is to do with insufficient bit depth and stretching the tonal range of the image to the point where step changes between brightness levels are visible. That is not the topic of this discussion.
 
Last edited:
Back
Top