86 replies to this topic

[alcyone00]

Got R8 few weeks ago and wanted to test some astro with it, previously I used Canon 77D and all other equipment is the same, Star Adventurer 2i and Samyang 135mm or Canon 70-300mm lens. Never had this problem before.

It's visible only after ABE or DBE in Pixinsight, there is some walking noise so I don't know if that's the cause for magenta color patches or something else with the camera sensor or some setting?

Like I said i did astro from same place for past 2 years but with different camera and never had that much magenta in the images.

 

It looks like lower ISO make it worse.

 

ISO 250, f2.8 with Samyang 135mm

40x45" stacked in DSS and only used DBE in Pixinsight so it's not fully processed because I just wanted quick preview.

 

ISO 1600, f8 Canon 70-300mm @ 300mm

56x30" stacked in DSS and only used DBE in Pixinsight.

 

 

[Dan Watt]

I also have an R8 but I don't have this issue. I do use WBPP to stack with Pixinsight. Perhaps it's an issue with DSS? 

[alcyone00]

I also have an R8 but I don't have this issue. I do use WBPP to stack with Pixinsight. Perhaps it's an issue with DSS? 

I tried with WBPP and Siril, same result...

[BPoletti]

Could it possibly be that those red areas are actually Ha emission areas that weren't captured earlier?  That is a fairly high Ha emission area.  

[alcyone00]

Could it possibly be that those red areas are actually Ha emission areas that weren't captured earlier?  That is a fairly high Ha emission area.  

Hmm I don't think so, I just tried different part of the sky and magenta pattern looks the same on different stack, also it's bortle 6 and I don't use any filter.

[sharkmelley]

What is your processing technique? Do you put all your raw lights, flats, darks and biases into DSS? Or do you convert the lights to TIF using LightRoom or RawTherapee, then stack the TIFs?

[alcyone00]

What is your processing technique? Do you put all your raw lights, flats, darks and biases into DSS? Or do you convert the lights to TIF using LightRoom or RawTherapee, then stack the TIFs?

I put all together into DSS, no converting.

[Tapio]

Actually I'm not seeing it (in this phone).
Are you dithering?
And maybe ABE/DBE needs tweaking (parameters) - someone could try your data?

[sharkmelley]

I don't know about the colour mottle but boosting the colour saturation I think I can see at least one big red ring in the image:

 

 

Possibly it's a lens artifact or possibly it's in-camera processing like we found in the EOS R and EOS Ra, which causes rings but only at low ISO:

• https://www.markshel...ConcentricRings
• https://www.cloudyni...non-ra-testing/

Unfortunately, very little is known about the behaviour of the Canon EOS R8.  If you upload a raw light and a raw flat it may be possible to determine what is causing the ring.

 

Mark

 

[alcyone00]

Actually I'm not seeing it (in this phone).
Are you dithering?
And maybe ABE/DBE needs tweaking (parameters) - someone could try your data?

I did dither on 135mm photo but only in RA because it's not possible in DEC with star adventurer.

 

I don't know about the colour mottle but boosting the colour saturation I think I can see at least one big red ring in the image:

 

CanonR8_Rings.jpg

 

Possibly it's a lens artifact or possibly it's in-camera processing like we found in the EOS R and EOS Ra, which causes rings but only at low ISO:

• https://www.markshel...ConcentricRings
• https://www.cloudyni...non-ra-testing/

Unfortunately, very little is known about the behaviour of the Canon EOS R8.  If you upload a raw light and a raw flat it may be possible to determine what is causing the ring.

 

Mark

Yes it's a new camera, libraw is still not supporting R8 if that helps. I will upload some raws tonight, also I didn't do flats so there is some vignetting but you are right about ISO, I think it's less visible on higher ISO and when there is no vignetting. I tried ISO 3200 and f5.6 with Samyang last night and there was almost no magenta but "blobs" pattern was slightly visible.

[alcyone00]

Did some tests tonight and maybe "found solution" 

 

This is stacked image with DBE only, 45 minutes, ISO 3200 and f7.1 on Samyang 135mm f2.8

I also did some manual dithering and moved frame every few minutes.

 

 

 

These are random part of the south sky with different settings, 15 minute stacks but difference is in f stop and ISO. 

 

 

ISO 400 f2.8, 30 sec per frame

 

ISO 3200 f8, 30 sec per frame.

 

Higher ISO and smaller aperture helps a lot, no flats are taken.

I also turned off noise removal in camera settings but don't know if that helped because I am not sure if it affect RAWs. Will do more tests tomorrow.

Edited by alcyone00, 09 September 2023 - 11:23 PM.

[sharkmelley]

Yes it's a new camera, libraw is still not supporting R8 if that helps.

But DSS uses libraw.  If the EOS R8 is unsupported by libraw then how are you opening the raws in DSS?

[ErwinL]

LibRaw does read unsupported CR3 frames but does not apply the correct bias. Also, image and frame sizes do not match. The color correction data might also be wrong, I guess.

[alcyone00]

But DSS uses libraw.  If the EOS R8 is unsupported by libraw then how are you opening the raws in DSS?

I get warning message that its not supported but it stacks them.

 

LibRaw does read unsupported CR3 frames but does not apply the correct bias. Also, image and frame sizes do not match. The color correction data might also be wrong, I guess.

Yes, frames don't match and I get border around but when compared to stack from Pixinsight colors looks the same.

[bulrichl]

Can you please upload some light and dark frames in raw format to a filehoster and post the link here?

 

Also check whether the lens distortion option is disabled.

 

Bernd

[alcyone00]

Can you please upload some light and dark frames in raw format to a filehoster and post the link here?

 

Also check whether the lens distortion option is disabled.

 

Bernd

https://drive.google...?usp=drive_link

 

Light, dark and bias frames, test I did last night. Magenta patches are visible slightly after DBE, I didn't dither this time. You will also notice some drifting between frames and walking noise because PA wasnt perfect.

 

Lens correction was enabled but does it matter if lens is manual without electronic contacts? Camera can't recognize which lens are attached.

[sharkmelley]

https://drive.google...?usp=drive_link

 

Light, dark and bias frames, test I did last night. 

Those uploaded files are all ISO 3200 and I can't see any rings in them.  It might be useful to upload an ISO 250 light and flat from your image in the original post.

[alcyone00]

Those uploaded files are all ISO 3200 and I can't see any rings in them.  It might be useful to upload an ISO 250 light and flat from your image in the original post.

I have only lights, no flats, if that's fine I can upload

[sharkmelley]

I have only lights, no flats, if that's fine I can upload

Sure, a couple of lights will be fine.

[alcyone00]

Sure, a couple of lights will be fine.

https://drive.google...pr5?usp=sharing

 

ISO 250, f2.8 with Samyang 135mm

[sharkmelley]

https://drive.google...pr5?usp=sharing

 

ISO 250, f2.8 with Samyang 135mm

I averaged your 5 ISO 250 lights then binned the data (to reduce noise) and divided the (bias-subtracted) red channel by the (bias-subtracted) green channel. Here is the result after stretching:

 

 

There are two obvious features:

• A set of fairly sharply defined concentric rings
• A "mottled" background

Neither the rings nor the mottling appear in the green or blue channels.  It's too early to speculate on the cause of the rings and I also have no idea of what might be causing the red mottle.  However, your experience so far seems to suggest that a high ISO gives much better results, so that might be the best way to proceed.

 

Mark

[alcyone00]

I averaged your 5 ISO 250 lights then binned the data (to reduce noise) and divided the (bias-subtracted) red channel by the (bias-subtracted) green channel. Here is the result after stretching:

 

CanonR8_ISO250_RdivG.jpg

 

There are two obvious features:

• A set of fairly sharply defined concentric rings
• A "mottled" background

Neither the rings nor the mottling appear in the green or blue channels.  It's too early to speculate on the cause of the rings and I also have no idea of what might be causing the red mottle.  However, your experience so far seems to suggest that a high ISO gives much better results, so that might be the best way to proceed.

 

Mark

For mottled background I also don't know, maybe sensor noise pattern? 

Could light pollution + lens vignetting cause those rings? I noticed on old camera I would get those rings only when light pollution is bad, like on humid nights for example, when I do astro outside of town from dark areas I never noticed those even on wide apertures, its like light pollution combined with lens vignetting increase those rings, but then again why it appears only on red channel, its weird...

[sharkmelley]

For mottled background I also don't know, maybe sensor noise pattern? 

Could light pollution + lens vignetting cause those rings? I noticed on old camera I would get those rings only when light pollution is bad, like on humid nights for example, when I do astro outside of town from dark areas I never noticed those even on wide apertures, its like light pollution combined with lens vignetting increase those rings, but then again why it appears only on red channel, its weird...

The mottling could be sensor related because other sensors certainly do have weird effects like this, especially in the red channel.  Flat frames are essential in order to combat this.

The only way to know for sure about the rings is to take a series of ISO 100 raw flat frames at different exposures using the test protocol I explained in this post:

• https://www.cloudyni...ngs/?p=12570173

If the number of rings changes from frame to frame then we know for certain that the cause is not optical but is caused by some kind of undocumented in-camera raw data processing.  If you shoot the sequence, I'm more than happy to analyse them.

 

Mark

[alcyone00]

The mottling could be sensor related because other sensors certainly do have weird effects like this, especially in the red channel.  Flat frames are essential in order to combat this.

The only way to know for sure about the rings is to take a series of ISO 100 raw flat frames at different exposures using the test protocol I explained in this post:

• https://www.cloudyni...ngs/?p=12570173

If the number of rings changes from frame to frame then we know for certain that the cause is not optical but is caused by some kind of undocumented in-camera raw data processing.  If you shoot the sequence, I'm more than happy to analyse them.

 

Mark

https://drive.google...?usp=drive_link

 

Aperture is f5.6, Samyang 135mm, focused to infinity. Thank you.

[sharkmelley]

https://drive.google...?usp=drive_link

 

Aperture is f5.6, Samyang 135mm, focused to infinity. Thank you.

That's a really useful set of test files - thanks for making the effort!

 

Unfortunately it confirms beyond doubt that in-camera processing is producing rings in the raw data.  As before, I've debayered and binned the data (to reduce noise), then divided the (bias-subtracted) red channel by the (bias-subtracted) green channel. Here is the result after stretching:

 

 

You can see a very clear progression of sharply defined concentric rings. Doing the same with the blue channel instead of the red channel gives the following:

 

 

Something might be going on in the blue channel but I'm not sure what.

 

You've already said that you don't notice the rings and mottle at higher ISOs, so that's probably the best workaround.  Certainly with my EOS R there are no rings at ISO 1600 and above in a stacked image.

 

We now know for certain that both the EOS R/Ra and EOS R8 produce rings at low ISO.  There is therefore a possibility that all Canon mirrorless cameras are similarly affected.

 

Mark

Edited by sharkmelley, 12 September 2023 - 02:12 PM.