Green band in long exposure on 5DMK4

Aug 23, 2019
3
0
Hi all, I took this long exposure a few nights ago just before sunset and the light was poor. Exposure time was 266 secs, ISO100. I noticed when I pulled the shadows, I had this green/cyan vertical line through the image. It was also on a few others but less pronounced. Nothing showing on normal exposures. I've read on other forums this has been seen before but I can't see that Canon have acknowledged it as a sensor issue? Anyone know if there is a fix or that if this is def a sensor issue? Thanks
Note: Exposure and shadows pulled up on this image to make line clearer
 

Attachments

YuengLinger

EOS 5D MK IV
Dec 20, 2012
2,505
610
Southeastern USA
How much did you pull the shadows?
How old is your 5DIV? Did you buy it new from an authorized dealer?
And have you sent this image to Canon service to ask about it yet?
Also, have you looked at the surface of the sensor, maybe with a magnifying glass? This looks almost like something on the sensor, though, just brainstorming here, it could also be a column of pixels producing a slight difference of exposure. (If that's possible.)
 
Aug 23, 2019
3
0
How much did you pull the shadows?
How old is your 5DIV? Did you buy it new from an authorized dealer?
And have you sent this image to Canon service to ask about it yet?
Also, have you looked at the surface of the sensor, maybe with a magnifying glass? This looks almost like something on the sensor, though, just brainstorming here, it could also be a column of pixels producing a slight difference of exposure. (If that's possible.)
Hi thanks for reply
Shadows were pulled up about 4 stops
Camera less than 2 years old, bought direct from Canon store
No, haven't sent to Canon yet but plan to
The fact I can't see it on any normal exposure images at all, I took before & after this leads me to think its not on the surface of the sensor, although I haven't looked tbh
 

SecureGSM

2 x 5D IV
Feb 26, 2017
1,239
263
Happy to share, but can't seem to attach a CR2 file? I can save as a DNG & attach - any good?
RAW please, I would like to understand what parameters were used. also, did you use any filters at all?

are you able to upload to mega.nz and provide a link? I will download and have a go...

p.s. seems to be the red part of the spectrum that is affected. there are multiple faint vertical lines to the left and to the right.
4 stops is heaps. plus a long exposure...
Cyan Line 2_DxO.jpg
 
Last edited:

privatebydesign

Would you take advice from a cartoons stuffed toy?
Jan 29, 2011
7,788
876
119
To me it just looks like one of the bands of read out electronics off the sensor got hot and distorted the signal. I suspect it was after a lot of live view and or other long exposures. Give the camera time to cool down and it goes away which is why you don't normally see it.

Turn LENR off and do 5 ten minute exposures in a row, I suspect you will see a similar thing in the later images but not the earlier images.
 

LSXPhotog

EOS RP
Apr 2, 2015
377
204
www.diossiphotography.com
And everyone thought I was crazy...

This problem also influenced the sensor on the EOS R until a firmware update quietly fixed the issue. I am surprised that the 5D Mark IV hasn't been treated to the same fix if it was something purely with firmware.
 
  • Like
Reactions: Del Paso

Joules

EOS RP
Jul 16, 2017
277
183
Hamburg, Germany
I doubt this is an actual problem. It is just what you can expect from a recent Canon sensor.

Here is the upper portion of a panorama I took yesterday. It is exposed for the amazing sunset and pushed too the limit to show the issue (I have a properliy exposed ground section too, this is just to make the point). Settings: 1/15s, ISO 100, f8. In LR: +4 stops exposure, -100 lights, +100 shadows. Shot on an 80D with a Samyang 16mm 2.0. This is a DNG Panorama created with LR and converted to JPEG, but the issue also shows up if I push the RAW files themselves with these settings.
 

Attachments

privatebydesign

Would you take advice from a cartoons stuffed toy?
Jan 29, 2011
7,788
876
119
I doubt this is an actual problem. It is just what you can expect from a recent Canon sensor.

Here is the upper portion of a panorama I took yesterday. It is exposed for the amazing sunset and pushed too the limit to show the issue (I have a properliy exposed ground section too, this is just to make the point). Settings: 1/15s, ISO 100, f8. In LR: +4 stops exposure, -100 lights, +100 shadows. Shot on an 80D with a Samyang 16mm 2.0. This is a DNG Panorama created with LR and converted to JPEG, but the issue also shows up if I push the RAW files themselves with these settings.
To get a better understand for those that just end up here because of a Google search for 'Canon sensor issue' could you post the base shots (as jpegs) to illustrate the changes you have made in post, many people won't realize just how extreme your adjustments are.

A small jpeg of the finished pano would be nice too.
 

Joules

EOS RP
Jul 16, 2017
277
183
Hamburg, Germany
To get a better understand for those that just end up here because of a Google search for 'Canon sensor issue' could you post the base shots (as jpegs) to illustrate the changes you have made in post, many people won't realize just how extreme your adjustments are.
That's a good point. Extreme adjustments is putting it lightly. In the RAW image the foreground is pitch black with absolutely 0 detail. You could argue that I even underexposed the sky by about a stop or so. Attached is the unedited version of a frame in the right third from the previous panorama. Just a straight conversion from RAW to JPEG with default lightroom settings.

I can't post the finished image yet, because I don't have enough free disk space currently for photoshop to process these high res panoramas :( So here's a temporary edit of the properly exposed lower half. Just image the blown out portion was filled with the sky from the above image. That's the vision at least, I hope I can pull it off once I have cleaned up some space.

This was taken about 20 minutes after sunset by the way. The only light on the heather hills is indirect light coming of those wispy clouds. Foreground exposure settings are 1s f7.1 ISO 100. It's still pushed pretty hefty to get the look you see here with -100 lights, +85 shadows and -0.6 exposure. Maybe I'll change that later on. It's just a quick edit so far. Just had to think of this thread when I saw the artifact in the severly underexposed ground of the sky shot.
 

Attachments

Last edited:
  • Like
Reactions: privatebydesign

LSXPhotog

EOS RP
Apr 2, 2015
377
204
www.diossiphotography.com
I would suggest not shooting any camera with the intent of pushing shadows 100% and exposure by more than 4-stops. This specific issue is not something I have on any other Canon camera I own, and it has only impacted 3-4 realworld shots, but it's still annoying and something that should be known by anyone that has the camera to avoid the conditions that may bring this..."unique characteristic" out.

As I mentioned, this also was present on the EOS R but after a firmware update it went away.

To avoid this from happening, never push shadows beyond 70% when pushing your exposure by more than 3 stops. There are other ways of getting the same shot that don't involve crushing the RAW file.
 
  • Like
Reactions: YuengLinger

SecureGSM

2 x 5D IV
Feb 26, 2017
1,239
263
Seems like a non-issue to me.
However, I keen to understand what specific camera settings caused this? I pushed shadows on a canon 5D IV file by crazy 2.8 stop once for a family friend. No Banding was detected. Not a long exposure file though. Underexposed due to misfired flashgun. So around 1/60 sec max.
 

LSXPhotog

EOS RP
Apr 2, 2015
377
204
www.diossiphotography.com
Seems like a non-issue to me.
However, I keen to understand what specific camera settings caused this? I pushed shadows on a canon 5D IV file by crazy 2.8 stop once for a family friend. No Banding was detected. Not a long exposure file though. Underexposed due to misfired flashgun. So around 1/60 sec max.
It is most pronounced when there is a heavily bright exposure on one side of the frame versus the adjacent side....and then you push the exposure an excessive amount. Not really an issue at all, honestly. But still annoying that it even happens at all.
 

SecureGSM

2 x 5D IV
Feb 26, 2017
1,239
263
Ah, ok. Sounds like this may be a result of an interference Of some sort. Dual pixel sensor? Subframe one, subframe two... we do not know as to how these two subframe form a final RAW image internally. I am not a physicist or an optician.
Interesting. Does this issue present in dual pixel RAW files as well? Just wondering if there are any problematic DPRAW files to play with.