New Canon 5d mk3 horizontal banding at iso200 - faulty body?

Started Jul 18, 2012 | Discussions thread
tmr Senior Member • Posts: 1,520
Re: underexposure does not exist.

Thanks for the comparison. I did some test shots trying to get camera settings similar to the op. On a shot that was underexposed and then pushed in post by 1 stop I saw a bit of banding but nowhere near the amount in the op shots. With proper exposure and HTP disabled I was hard pressed to find any banding.

Here is a link to the test shots: http://www.pbase.com/tmr4/banding_test

You need to look at them in original size for this exercise. Here is a small size sample for context only.

Hopefully DPR updates their gallery to take unmodified 5d3 files. I didn't want to do any other modifications to these.

MayaTlab0 wrote:

tmr wrote:

MayaTlab0 wrote:

tmr wrote:

I don't understand the complaints against Canon. The issue here is clearly caused by the use of LR. Canon isn't responsible for that. Your complaint should be to Adobe. Many have pointed out how to avoid the problem by using Canon software. That my not work for your workflow but that doesn't make it a 5d3 problem.

It's there as well in DPP, just less noticeable, mostly because i believe the tone curve is less open in the shadows and you can't push them as far as LR.

I recommend staying with LR, but using Dfine or another specialised de-banding tool in the workflow.

Perhaps, but the op isn't pushing anything here and he is getting significant banding with no pp. As I don't have LR I can't do my own study but several folks have posted both jpgs from LR and the original raws and the raws in DPP really don't show any banding while it is clearly visible in the LR jpgs. And I do not see any banding in my own 5d3 shots of a similar nature to the op, though I don't have any with HTP enabled and at 200 iso so perhaps there is something in that.

I understand that many find LR to be quick and more functional, but if it can't render the raw to ones satisfaction then another alternative is order. Your suggestion of using Dfine makes sense for folks who need it. Using DPP on a one off basis for those shots where using LR is particularly problematic could be another.

One of the two reasons DPP doesn't exhibit as much banding is because its default shadow curve isn't as open as the one on LR. So in essence the default LR image is pushed a little harder than the one from DPP.

As an example to show that DPP is just as prone to banding as LR, here is one of the OP's raw processed in LR (first) with everything default except 0 sharpening, 50 chroma noise reduction, pushed two stops, and in DPP, with everything default except 0 sharpening, 0 luminance noise reduction (to put it in line with LR), 4 chroma noise reduction, pushed 2 stops as well. banding is very well visible in both :

When not pushed two stops, DPP effectively shows a little less banding, but I think that's mostly because its shadows seem to cut earlier than LR to me. The other reason why DPP doesn't show as much banding is that its noise reduction algorithms are quite a bit more aggressive than LR's ones, at the cost of details. The only way to effectively address banding is to use a de-banding software. Here's just an example with the same LR processed shot, not pushed 2 stops, passed through Dfine :

 tmr's gear list:tmr's gear list
Canon EOS 400D Canon EOS 5D Mark III Canon EF 50mm f/1.2L USM Canon EF 85mm f/1.8 USM Canon EF 135mm F2L USM +4 more
Post (hide subjects) Posted by
tmr
tmr
tmr
tmr
tmr
tmr
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow