Purple banding with long exposure (R5 + 24-70 f/2.8L)

[fr34k]

Hi guys,

I wanted to compose an image using two long exposures, one for the sky (Milky Way) and one longer one for the ground. But for the longest one I attached here, you can clearly see a purple fringe at the left edge. What could that be caused by and how can I avoid that happening? There are some hints of blue fringing on the right, but the purple one is really visible.
If you boost the image for the sky, you can also see the same type of fringe. Is my sensor broken? It always happens on one side of the sensor, no matter the orientation but all "normal" type images look spectacular and do not display anything weird.

R5, 24-70 f/2.8L @ 24mm, f/2.8, ISO 6400, 300s.

 

[candyman]

Not sure what caused the purple fringe on the left side. But 300 seconds seems a lot for creating a star trail and freeze the water in the foreground. Even when stacking. Did you try with less long exposuretime? What do you see?

 

[scyrene]

There used to be a thing called "amp glow", especially on older bodies like the 5D3; my understanding is it was caused by heat from the electronics of the camera bleeding into the sensor, especially visible in long exposures of very dark scenes/ultra high ISO. I thought it was much less of an issue nowadays but perhaps your settings were enough to make it visible.

 

[neuroanatomist]

There used to be a thing called "amp glow", especially on older bodies like the 5D3; my understanding is it was caused by heat from the electronics of the camera bleeding into the sensor, especially visible in long exposures of very dark scenes/ultra high ISO. I thought it was much less of an issue nowadays but perhaps your settings were enough to make it visible.

Amp glow was my first thought when the image loaded.

 

[fr34k]

Not sure what caused the purple fringe on the left side. But 300 seconds seems a lot for creating a star trail and freeze the water in the foreground. Even when stacking. Did you try with less long exposuretime? What do you see?

On that day the fringe was about proportional to the exposure time.
This exposure was not set to be correct for the stars, but to brighten the environment (still dark, but to have some silhouette), so I can overlay it with another one I took which was correctly exposed for the Milky Way (24mm, f/2.8, 8~10s, ISO 6400).

Where we were, the environment was pitch black and the only thing illuminating was the night sky (without moon of course), so I needed an enormous exposure time. I have noticed that fringe on two occasions, both times the weather was around 25°C (77°F) and I took quite a few long exposures, but I would not say it got worse the more I took.

 

[fr34k]

There used to be a thing called "amp glow", especially on older bodies like the 5D3; my understanding is it was caused by heat from the electronics of the camera bleeding into the sensor, especially visible in long exposures of very dark scenes/ultra high ISO. I thought it was much less of an issue nowadays but perhaps your settings were enough to make it visible.

Amp glow was my first thought when the image loaded.

thanks for the explanation. I've never heard of that before. It was decently warm and I did take lots of images, but I cannot say that it got worse during the evening.

It was ultra dark there. If we wouldn't have been dimly illuminated by the night sky (sans moon) it would've been absolutely black like in the forest we treaded through to get there. You couldn't see your hand in front of your eyes even if you wanted.

 

[neuroanatomist]

Reports on the interwebs vary, but the R5 doesn't appear to be a serious offender in terms of amp glow (the R6 does, however). You mostly find this info on astrophotography sites (long exposures of dark scenes).

Sort of a side note since it shouldn't apply to your camera, but RF lenses (except the 28-70/2) have an IR LED inside them that makes them poor choices for long exposure IR photography. According to the source (article on Kolari Vision, one of the companies that convert cameras for IR), this should not affect a regular camera. But 300 s is in the time frame where you are able to use a front filter for daytime IR photography with a regular sensor. Seems unlikely (the pattern is different), but it's easy to test if you have an EF lens and the adapter (shoot 300 s exposures with both lenses with the lens cap on and see if the glow is there/different)

Theoretically, it could also be a light leak from another source. You could also try that 300 s dark exposure with no lens and the body cap in place (you may need to enable the setting to Release shutter w/o lens).

 

[scyrene]

Sort of a side note since it shouldn't apply to your camera, but RF lenses (except the 28-70/2) have an IR LED inside them that makes them poor choices for long exposure IR photography. According to the source (article on Kolari Vision, one of the companies that convert cameras for IR), this should not affect a regular camera. But 300 s is in the time frame where you are able to use a front filter for daytime IR photography with a regular sensor.

Wow! What is its purpose?

Theoretically, it could also be a light leak from another source. You could also try that 300 s dark exposure with no lens and the body cap in place (you may need to enable the setting to Release shutter w/o lens).

Yeah my suggestion would be take a dark frame (lens/body cap on), which can be used to cancel out effects like this in postprocessing, at least to some extent. Best done under the same conditions as the original shots (ie just before or after).

 

[scyrene]

Wow! What is its purpose?

No worries, I clicked through to the article which explains it. Thanks @neuroanatomist !

 

[fr34k]

Reports on the interwebs vary, but the R5 doesn't appear to be a serious offender in terms of amp glow (the R6 does, however). You mostly find this info on astrophotography sites (long exposures of dark scenes).

Sort of a side note since it shouldn't apply to your camera, but RF lenses (except the 28-70/2) have an IR LED inside them that makes them poor choices for long exposure IR photography. According to the source (article on Kolari Vision, one of the companies that convert cameras for IR), this should not affect a regular camera. But 300 s is in the time frame where you are able to use a front filter for daytime IR photography with a regular sensor. Seems unlikely (the pattern is different), but it's easy to test if you have an EF lens and the adapter (shoot 300 s exposures with both lenses with the lens cap on and see if the glow is there/different)

Theoretically, it could also be a light leak from another source. You could also try that 300 s dark exposure with no lens and the body cap in place (you may need to enable the setting to Release shutter w/o lens).

I did two exposures, one with lens and one without - almost identical. (ISO 6400, 300s) But neither display the fringe on the lower side of the sensor. It's not as warm in my room right now as it was on those days, but regarding explanation of the origin, I would've at least expected a litte, instead the left side, especially the corners seem to have some brightness to them, but nothing of that sort.

w/l



w/o

 

[fr34k]

Yeah my suggestion would be take a dark frame (lens/body cap on), which can be used to cancel out effects like this in postprocessing, at least to some extent. Best done under the same conditions as the original shots (ie just before or after).

If I would've known, I'd done so, but that's not possible anymore (that place is so f'n remote, yikes, and in a far away country also, so 2x impossible). Next time I'll know what to do, so I'll take it as a learning experience and somehow work with what I've got.
I really appreciate the help. Thanks!

 

[neuroanatomist]

t's not as warm in my room right now as it was on those days, but regarding explanation of the origin, I would've at least expected a litte, instead the left side, especially the corners seem to have some brightness to them, but nothing of that sort.

The heat that causes amp glow isn’t due to the ambient temperature, but rather due to the camera electronics (CPUs under high load can approach the temperature of boiling water, which is why most computers have some form of active cooling, i.e. a fan). Ambient temperature affects it only to the extent that it increases or reduces the rate at which the heat inside the camera is dissipated.

For your test images above, if you just turned the camera on and took those two exposures, you may not see the problem. When you were on location, probably you had been shooting. Lots of pictures, or even more a fair amount of time spent composing images with the screen or EVF active, would warm up the camera’s electronics. Perhaps, after you next go out and spend some time shooting, you could try the test again.

 

[fr34k]

Some official info:
I send my camera off to be serviced for another issue and asked them to investigate the issue somewhat.
Their response was essentially the same, this being thermal noise and a limitation of the camera.