Colour Banding Issue

[ST1]

I would welcome anyone's thoughts on the colour banding visible in this attached image.

The attached jpg is from a raw file with EXIF: ISO800, f16, 6.00sec, I have only lifted the exposure by around 2.5 stops to allow the banding to become visible in the attached jpg.

Having spoken to the Canon UK technical people they have suggested that it may have been caused by me not having long exposure noise reduction enable in the camera settings. Sadly I can't return to the site and re take the image, to see what effect the use of LE noise reduction may have.

I'm wondering if the banding was caused by some external influence, to the left and behind the camera there is a large Radome that may have military or commercial applications running. To the right of the shooting position there is a number of telescopes within dome style enclosures (they open to allow the scope to see the sky [not sure of the correct term for them ])

The following link is to my website showing a processed image from the same day and location without any banding.

https://www.peterstockton-photograph...hh/i-Jf2b6xr/A

The link below is to the attached image on my website as I couldn't see how to attached anything other than a thumbnail size jpg to this post.

https://www.peterstockton-photograph...hh/i-4mSLjwX/A

[pnodrog]

Is the banding apparent on adjacent and/or similar images?

[ST1]

Is the banding apparent on adjacent and/or similar images?

Hi Paul

They are visible on adjacent images the jpg I posted was one a set of bracketed images all three of which have the banding.

[DanK]

What camera?


Sent from my iPhone using Tapatalk

[ST1]

What camera?

Canon 5D Mark 1V

Sent from my iPhone using Tapatalk

[DanK]

I short with a 5D Mark IV, and it doesn’t need LE noise reduction with exposures as short as 6 seconds. I shoot several times that long without it. Will think more.


Sent from my iPad using Tapatalk

[DanK]

I have trouble seeing the banding. When I expand the image on my ipad, I see a blue vertical band in the sky toward the left. Is that what you are referring to?

AFAIK, LE noise is of two types. The more common is hot pixels. These are bright dots that appear random but in fact occur at the same place in every image. LE noise reduction works by taking a black exposure, which will have the hot pixels but nothing else, and subtracting it from the prior image. Banding is entirely different.

The second type of LE noise is thermal noise that results from the sensor heating up. This can take the form of odd color distortions, and I suppose it could create a band. I have literally never experienced it with my 5D IV, even though I do a fair amount of long-exposure work. I did experience it with a 50D on a hot summer night. I think the exposure was more than 10 minutes, maybe 20 minutes.

Here's what I would do:

1. dim the light in a room to a level where you can take a properly exposed photo of a blank wall at 6 seconds.
2. either by manipulating the light or by changing the aperture, take a second shot at 2 seconds and a third at 1/2 second, keeping the exposure about the same.

Examine all three raw files for banding.

If there is no banding, I'd repeat this with a lower exposure, say, 2 or 3 stops below the metered level, and boost the brightness in post before examining the images.

[ST1]

I have trouble seeing the banding. When I expand the image on my ipad, I see a blue vertical band in the sky toward the left. Is that what you are referring to?

AFAIK, LE noise is of two types. The more common is hot pixels. These are bright dots that appear random but in fact occur at the same place in every image. LE noise reduction works by taking a black exposure, which will have the hot pixels but nothing else, and subtracting it from the prior image. Banding is entirely different.

The second type of LE noise is thermal noise that results from the sensor heating up. This can take the form of odd color distortions, and I suppose it could create a band. I have literally never experienced it with my 5D IV, even though I do a fair amount of long-exposure work. I did experience it with a 50D on a hot summer night. I think the exposure was more than 10 minutes, maybe 20 minutes.

Here's what I would do:

1. dim the light in a room to a level where you can take a properly exposed photo of a blank wall at 6 seconds.
2. either by manipulating the light or by changing the aperture, take a second shot at 2 seconds and a third at 1/2 second, keeping the exposure about the same.

Examine all three raw files for banding.

If there is no banding, I'd repeat this with a lower exposure, say, 2 or 3 stops below the metered level, and boost the brightness in post before examining the images.

Thanks for the suggestions Dan, I shall give them a try. The blue line you can see is the one I was referring to, it's also present in the lower part of the frame within the cloud area too.

[BobGilbody]

Hello Peter,

There is a red spot 'glow' to the left of the vertical blue line. The vertical blue line is difficult for me to see in the bright area of the frame, but obvious in the darker areas.
It might be that the cloud cover has affected the red spot 'glow' for a short period of time; assuming the clouds are moving a bit. At a slow shutter speed all the frame will be exposed at the same time for most of the exposure.

Then again maybe it's the effect of random yobba rays emanating from the the domes.

I learnt of yobba rays from 'Space Patrol'.

Hope this is of some help. I don't think it makes much sense. Please ignore this post.

Bob

[pnodrog]

Hi Peter,

what shutter mode were you using - Mechanical, electronic front curtain or fully electronic?

Other than being very skeptical about it having anything to do with noise reduction, at this stage it has me puzzled.

Do the bands that are on the other images all occur in the same place?

P.S. The band looks more cyan than blue which at a wild guess could indicate the problem has something to do with the red channel.

[DanK]

PS: I searched in vain for the old photo with thermal noise. However, I have a hunch that LE noise reduction won't remove it because it probably has a random component, and LE noise reduction relies on noise being fixed. That's just my hunch, however.

[Stagecoach]

P.S. The band looks more cyan than blue which at a wild guess could indicate the problem has something to do with the red channel.

Does the fact that the cyan band appears to disappear in 'brighter' areas along its route give a clue?

[pnodrog]

Does the fact that the cyan band appears to disappear in 'brighter' areas along its route give a clue?

Probably but unfortunately I'm unable to interpret it....

[ST1]

Hello Peter,

There is a red spot 'glow' to the left of the vertical blue line. The vertical blue line is difficult for me to see in the bright area of the frame, but obvious in the darker areas.
It might be that the cloud cover has affected the red spot 'glow' for a short period of time; assuming the clouds are moving a bit. At a slow shutter speed all the frame will be exposed at the same time for most of the exposure.

Then again maybe it's the effect of random yobba rays emanating from the the domes.

I learnt of yobba rays from 'Space Patrol'.

Hope this is of some help. I don't think it makes much sense. Please ignore this post.

Bob

Bob that red glow that you mentioned is the lights of something in the valley below the clouds

[ST1]

Hi Peter,

what shutter mode were you using - Mechanical, electronic front curtain or fully electronic? Mechanical shutter Paul

Other than being very skeptical about it having anything to do with noise reduction, at this stage it has me puzzled. Me too Paul

Do the bands that are on the other images all occur in the same place? They do

P.S. The band looks more cyan than blue which at a wild guess could indicate the problem has something to do with the red channel. I agree the band is more Cyan than blue

Paul see my responses to your questions within your quoted text above. The original image was the sixth I made that morning. The combined shutter speed of the six images 104 seconds. I had the camera set to Aperture Priority, with the metering mode as centre weighted average. Shooting height was 1801.6 metres above sea level.
I'm beginning to wonder if the GPS receiver in the camera was picking up signals from the nearby telescopes/Radar site. But I maybe clutching at straws with those thoughts.

[ST1]

Does the fact that the cyan band appears to disappear in 'brighter' areas along its route give a clue?

It doesn't give me a clue Grahame, does it to you?

[Stagecoach]

It doesn't give me a clue Grahame, does it to you?

No it does not Peter.

I undertook a search on pixel row failures and everything I found showed a very clear narrow sharp line.

Placing the image in curves and adjusting the mid slider towards black emphasizes things and appears as if a strong ray of cyan/blue light travelling through the clouds, certainly a mystery.

[BobGilbody]

I enjoyed looking at the photos of old Russia. https://en.wikipedia.org/wiki/Sergey_Prokudin-Gorsky
I know next to nothing about pixel things but it does seem that digital cameras take black and white photos which are then manipulated to give colours. Whether or not this is additive or subtractive I would not know.
Some of the old Russian photos are shown with the edges of the photo not cropped. Although the image is still there it is yellow or green or red. It seems to me that the disappearance of the band in the 'white' portion of the image that Peter presented is important.
I guess white is the result of all the blue green and yellow pixels being full. Where the image is not white some 'malfunction' of the sensor is providing a blue end result. This suggests to me that the band effect is due to something going wonky with the sensor.
It would be interesting to see images showing banding with longer longer exposure times.

Got it wrung: blue green and red filters used. Can't escape senility eh

[pnodrog]

Paul see my responses to your questions within your quoted text above. The original image was the sixth I made that morning. The combined shutter speed of the six images 104 seconds. I had the camera set to Aperture Priority, with the metering mode as centre weighted average. Shooting height was 1801.6 metres above sea level.
I'm beginning to wonder if the GPS receiver in the camera was picking up signals from the nearby telescopes/Radar site. But I maybe clutching at straws with those thoughts.

Thanks for answering my questions, unfortunately I am no closer to answering yours. The fact that you were using mechanical shutter eliminates some possibilities but there are still plenty of others on which to speculate. Hopefully it will remain a unrepeated mystery.

[William W]

I have been reading this conversation from the Opening Post . . .

Paul see my responses to your questions within your quoted text above. The original image was the sixth I made that morning. The combined shutter speed of the six images 104 seconds. I had the camera set to Aperture Priority, with the metering mode as centre weighted average. Shooting height was 1801.6 metres above sea level.
I'm beginning to wonder if the GPS receiver in the camera was picking up signals from the nearby telescopes/Radar site. But I maybe clutching at straws with those thoughts.

Don't dismiss that train of thought. I remember reading a conversation on another Photo Forum around 2009 ??? (ref: EOS 5DMkII) concerning colour banding issues - the gist was the (contaminated) images were taken near a military site and the conversation went into depth on that matter.

I have been looking but can't find the thread - maybe because the indexing of old threads is a bit haywire at the moment. The site is Photo.net.

Anyway - perhaps researching (documented?) effects of RF and other high band interference on EOS DSLRs might spring fruit - we need a resident expert in this area to offer ideas - any takers? Where's Ted?

I am still searching . . .

WW