Green band (fixed-pattern noise) in photos with Canon 5D Mark IV

yadavtarun03 wrote:

TLDR;

• Photos from Canon 5D Mark IV show a band (fixed-pattern noise).

Problem

I recently noticed that long exposures taken at higher ISOs with my Canon 5D Mark IV (purchased in Oct 2016, still in the warranty period) demonstrate a green band close to one edge.

This is quite visible in under-exposed images that are edited in Lightroom to increase the exposure (starts showing up with +1-stop, often when coupled with other changes).

.

5DIV, 200s, ISO 800, f/16 Exposure +0.66, Dehaze +100 in Lightroom to show green band at bottom

5DIV, 420s, ISO 800, f/16 Dehaze +100 in Lightroom to show horizontal green band along bottom edge

5DIV, 120s, ISO 640, f/11 Exposure +3, Dehaze +100 in Lightroom to show vertical green band along right edge

When I first reached out to Canon, I didn't know what this band was. I've since found out about Fixed-Pattern Noise

What you are seeing there is not Banding or Fixed Pattern noise. It's always in the same place and only related to one row of pixels. Likely an issue with the electronics associated with that row. Will require a sensor change to fix.

Mako2011 wrote:

What you are seeing there is not Banding or Fixed Pattern noise. It's always in the same place and only related to one row of pixels. Likely an issue with the electronics associated with that row. Will require a sensor change to fix.

Agreed - there is something wrong and it's not what we would normally refer to as "banding". Faulty sensor (or electronics associated with it) is a good guess.

Steve Balcombe wrote:

Mako2011 wrote:

What you are seeing there is not Banding or Fixed Pattern noise. It's always in the same place and only related to one row of pixels. Likely an issue with the electronics associated with that row. Will require a sensor change to fix.

Agreed - there is something wrong and it's not what we would normally refer to as "banding". Faulty sensor (or electronics associated with it) is a good guess.

On re-reading then Wikipedia article, it sounds like the distinction between what I have and FPN is that FPN presents as a more regular grid pattern?

Additionally, as I noted in the post, Canon has already looked at this and said it performs to spec. I’d welcome any suggestions to help prove more definitively that the unit is defective.

Thanks!

yadavtarun03 wrote:

Additionally, as I noted in the post, Canon has already looked at this and said it performs to spec. I’d welcome any suggestions to help prove more definitively that the unit is defective.

Thanks!

I can't help you with that specific case, but Canon's position isn't definitive obviously and I'm not even sure that they actually believe what they answered you. It's just one of their BS standard starting point for "negotiations". If I had never pushed passed the usual "it's within specifications" BS that nearly all customer services always serve me, I'd have had nothing done, ever.

When I know that something of this kind is hard to prove (for example if I don't have two copies of the same product), I usually try to elevate the issue. One of the most efficient way I've found to get things done quickly is to find the company's email format, re-construct the mails of all the top management (regardless of whether they're qualified to answer or not), and then bombard all of them with a carefully explained mail where I clearly state the problem, why it affects me, and what I want them to do.

Since 99% of the time I was right and my product wasn't actually "within spec", I then always make a point to prove to them afterwards that their first answer was complete BS by showing them the effect of their own repair.

That being said, can you have access to another 5DIV from a friend, a shop, or from another source ? If you have found a way to reproduce your 5DIV's problem consistently, then comparing two 5DIVs could help in starting to build up evidence.

Thanks for your thoughts and observations Joe. I'm also considering reaching out to their social media accounts on Twitter and Facebook - hopefully that'll help.

Thanks for your thoughts and observations. I'm also considering reaching out to their social media accounts on Twitter and Facebook - hopefully that'll help.

As for a second 5DIV — I'd thought about that and that was one of my reasons for posting online. I don't have easy access to one and I was hoping someone might be able to try and reproduce. I will try and escalate and get this solved  again but if that fails, I might rent a body for a day or two to test with. I'd rather not spend more but if that's what it takes to get a fix, so be it

yadavtarun03 wrote:

Steve Balcombe wrote:

Mako2011 wrote:

What you are seeing there is not Banding or Fixed Pattern noise. It's always in the same place and only related to one row of pixels. Likely an issue with the electronics associated with that row. Will require a sensor change to fix.

Agreed - there is something wrong and it's not what we would normally refer to as "banding". Faulty sensor (or electronics associated with it) is a good guess.

On re-reading then Wikipedia article, it sounds like the distinction between what I have and FPN is that FPN presents as a more regular grid pattern?

Additionally, as I noted in the post, Canon has already looked at this and said it performs to spec. I’d welcome any suggestions to help prove more definitively that the unit is defective.

Call again. Ask to speak to a supervisor. Call another service center if needed.

I had to send my 135L three times to Canon. The first two times I could not convince them that there was a problem.

J A C S wrote:

Call again. Ask to speak to a supervisor. Call another service center if needed.

I had to send my 135L three times to Canon. The first two times I could not convince them that there was a problem.

Thanks, I'll keep at it!

MarshallG wrote:

yadavtarun03 wrote:

Additionally, as I noted in the post, Canon has already looked at this and said it performs to spec. I’d welcome any suggestions to help prove more definitively that the unit is defective.

Try the exact same thing on another 5D Mark iv.

You're taking some very, very long exposures. It may be that ALL the Mark iv's do this; your 150 second "corner case" could have escaped Canon QA. Or maybe it's a glitch in the software; who knows? The first thing you should do is use another 5D iv's side-by-side with yours to see if the problem is unique to your body or not. There's no point in exchanging your camera if the problem will follow it.

Thanks! I'd definitely love to do that but I don't have easy access to another unit. I could rent from something like BorrowLenses.com but it's USD 123 (3-day minimum) before tax + delivery and I'd very much like to avoid that if possible. It's certainly a last resort option though!

By posting, I was hoping that 5DIV owners might be able to try and report if they see this problem with their cameras.

@Keeef has replied suggesting posting to the Astrophotography forum and I think I will try that to see if there are any users who can confirm or refute that this problem is endemic to the model.

Keeef wrote:

Ask in the Astrophotography talk forum if anyone has seen anything similar.

Thanks for the suggestion - that sounds like a good idea!

Blake Cook wrote:

See what happens if you open the photos with DPP.

Thanks for the suggestion - I gave this a quick try:

• I haven't used this editor before and I couldn't tell how to edit the photo to see the same final result as my edits in Lightroom
• As with other viewers, I do see the green band without any edits (albeit faint). I can make it a little more obvious by fiddling around with the contrast and color temperature.
• Why do you think opening with this program might make a difference? I expect the problem to be inherent to the image itself.

http://www.canonwatch.com/sensor-banding-issue-canon-eos-5d-mark-iv/.

check the long expose examples in the comment section .

AmbleYonder wrote:

http://www.canonwatch.com/sensor-banding-issue-canon-eos-5d-mark-iv/.

check the long expose examples in the comment section .

That is not the same issue the OP is having. Note the difference in color/size/edges

Blake Cook wrote:

yadavtarun03 wrote:

Blake Cook wrote:

See what happens if you open the photos with DPP.

Thanks for the suggestion - I gave this a quick try:

• I haven't used this editor before and I couldn't tell how to edit the photo to see the same final result as my edits in Lightroom
• As with other viewers, I do see the green band without any edits (albeit faint). I can make it a little more obvious by fiddling around with the contrast and color temperature.
• Why do you think opening with this program might make a difference? I expect the problem to be inherent to the image itself.

DPP is a Canon product that is designed for its sensors. Adobe Raw seem to take at least year to start to get colours right for Canon bodies in recent years. 1Dx processing isn't right yet.

DPP can't fix the OP's issue though. It's not a processing artifact. The line can be seen on the back of the camera when one looks

-- hide signature --

Blake in Vancouver
Canon and Zeiss Stuff. Mac Stuff & annoying PC & Windows stuff.