Strange colors in my live stack image ?

[Opmc73]

Hi Robin,

I am figthing since a long time with strange circles of colors in my pictrues. I got that since I am doing master flat with SC.
Initially I was looking the responsible in my focal reducer, in my flat screen, etc... but now I wonder if it does not come from SC.

I have finally redo the stack of last nigth with pixinsigth using the same flat and dark I used in SC and look the result side by side in the capture side by side : left is the stack done by Pix and rigth is the stack done by SC. We can clearly see a kind of additionnal yellow circle color . I was wondering if SC is really using the Mono flat generated because this is very strange and I have dozen of captures with that strange colored circles. I got that when I used the RC8 and now the RC10 as I am using in both case the same camera , focal reducer and flat screen I was looking the issue with those element, but the other common poinr is sharpcap, so I submit the question to you (When doing flat I select mono flat, use bias , let SC adjust the histogram for me, I use the same offset that in my dark and pictures, and remove the dark + let SC put the masterbias in the process).

Capture d’écran 2024-10-29 140352.jpg (323.04 KiB) Viewed 854 times

[admin]

Hi,

does your flat frame show a distinct change in brightness at the point where the edge of the circle is showing up in SharpCap? If so, that would tend to tie in with your idea that it is related to the flat frame correction. It's probably best to check the flat frame in some sort of FITS viewer to see what the structure is - also look at the pixel level - if you see a grid pattern then the flat is RAW, but no grid means mono (SharpCap will not use a 3 channel colour flat to correct a RAW image, so that's not going to be an issue).

It's also possible that the effect could come from some minor problem with the way that the image is being dark subtracted (which happens before the flat frame is applied). If the dark is slightly out due to different camera settings (gain, exposure, brightness/black level, colour balance) then the subtraction will not result in the part of the signal coming from photons hitting the sensor, so following on from that the flat correction will not do the right thing.

However, given that PI is not showing the problem then it's possible that one of the calculations happening in SharpCap is slightly out, giving a biased result even with correct data. Because deep sky imaging involves stretching the image strongly, even very small errors can become noticeable in the stretched image.

If you want to share enough of the raw frames and calibration frames (say via google drive) to reproduce the problem, I'd be happy to take a look and try to work out what is going on in more detail, since it might lead to an improvement in SharpCap.

cheers,

Robin

[Opmc73]

Hi,

does your flat frame show a distinct change in brightness at the point where the edge of the circle is showing up in SharpCap? If so, that would tend to tie in with your idea that it is related to the flat frame correction. It's probably best to check the flat frame in some sort of FITS viewer to see what the structure is - also look at the pixel level - if you see a grid pattern then the flat is RAW, but no grid means mono (SharpCap will not use a 3 channel colour flat to correct a RAW image, so that's not going to be an issue).

It's also possible that the effect could come from some minor problem with the way that the image is being dark subtracted (which happens before the flat frame is applied). If the dark is slightly out due to different camera settings (gain, exposure, brightness/black level, colour balance) then the subtraction will not result in the part of the signal coming from photons hitting the sensor, so following on from that the flat correction will not do the right thing.

However, given that PI is not showing the problem then it's possible that one of the calculations happening in SharpCap is slightly out, giving a biased result even with correct data. Because deep sky imaging involves stretching the image strongly, even very small errors can become noticeable in the stretched image.

If you want to share enough of the raw frames and calibration frames (say via google drive) to reproduce the problem, I'd be happy to take a look and try to work out what is going on in more detail, since it might lead to an improvement in SharpCap.

cheers,

Robin

Robin,

You will find the zip with my data from this nigth it is not too big because I took only 11 frames of 60s. The scope is a RC10 with focal reducer. You will see a kind of ring in the flat, this is because I removed the baffle extension to see if the stange color had something to do , but no it just introduced some reflexion in certains case with brigth stars out of the field (M101 pictures by example).

Hope this will be usefull.

We transfer link : (it is 3 or 7 days valid don't know the rule now before it was 7 but I saw 3 somewhere?)

https://we.tl/t-ZLgT8HaeLd

Kind regards
Olivier

[Opmc73]

Hi Robin,

it seems I got better results doing 2 things:

1) I did again my flat but setting the exposure manually because SC automatic calibration for flat was setting exposure time to 1 sec or was changing the gain to minimize the exposure time. I did my new flat with unity gain (100 for touptek 2600) and set the time to about 300ms (and did the ligth accordingly to have the histogram ok). Acolleague in a forum told me to set exposure to 17ms, but with my flat screen it is too low for SC..

2) I took my picture withe a gain that was 2x the unity gain (Touptek max is 300000 , I set 200) . Then I don't see anymore those strange colors because it seems to me it enhance the background ligth and this is hidden the circle that are still a little there if I push the histogram.

[admin]

Hi,

thanks for sharing the files and also for the later report on how you have made improvements. I will have a look at the files over the next few days to see if there is anything I can add to the understanding of the problem, or to improve SharpCap's processing.

cheers,

Robin

[Opmc73]

Hello Robin,

I finally figured out what was causing this issue : the dark I was using was not good. I have redo it and now it is much better. But I have still that issue on some part of the sky , it seems it is when there is ligth polution (I have a village nearby) , it creates a gradient that for any reason is circular , certainly where the shadow of the secondary mirror is present.

Also the histogram start very short on the left, is this normal ? It comes like this from the 1st frame... sometime not, and often yes. See below .

Capture d’écran 2024-10-31 212815.jpg (428.51 KiB) Viewed 827 times

That does not explain why pix did better but now it is very satisfying in SC.

Kind regards
Olivier

[admin]

Hi,

I wonder if your dark image is still not quite matching your light image - I can see a faint blue ring in that image, which may be the result of the dark being slightly off causing the flat correction to behave weirdly, then the colour balance adjustments making it look blue.

The histogram being very close to the edge is also a bit odd - the dark subtraction will move it further left, but you would usually expect some gap remaining because the light frames will have light pollution in them, raising the background level above that of the dark frames.

It's probably worth checking the background levels of the dark and light frames in a FITS viewer that gives pixel readouts to see how close they are.

cheers,

Robin

[Opmc73]

Hi,

I wonder if your dark image is still not quite matching your light image - I can see a faint blue ring in that image, which may be the result of the dark being slightly off causing the flat correction to behave weirdly, then the colour balance adjustments making it look blue.

The histogram being very close to the edge is also a bit odd - the dark subtraction will move it further left, but you would usually expect some gap remaining because the light frames will have light pollution in them, raising the background level above that of the dark frames.

It's probably worth checking the background levels of the dark and light frames in a FITS viewer that gives pixel readouts to see how close they are.

cheers,

Robin

Hi Robin,

it i stange, sometimes I have that colored circle, sometimes not with SC. I have redone all my darks yesterday and image were better, but still have that circle.
The histogram very close to the edge is done by SC...I have no control on it and I don't why it does that. Sometime it is normal not on the edge and most often it is like that... it is a mysterious thing for me...
I observed that if I let sharpcap calculate the exposure time , offset etc for me the everything is ok.

Moste of the time I do shoot with 60s exp, gain 100 , ofset 396 and temp -10°c and I am using the dark accrodingly.

Kind regards
Olivier

[Opmc73]

Robin,

This is my shhot at the moment. I have perfromed before the analysis of the bottom of sky with the histogram tool and followed the recommandation of SC : gain 300 , offset 396 etc.. no strange colors for the moment. But as you can see, the histogram start on the very left edge as I have each time. I can not do anything to have the left side of the histogram it seems it is managed by sharpcap automatically

[admin]

Hi,

can I just check that your dark frames are being captured with exactly the same settings as the light frames? The obvious settings are exposure and gain, but offset and colour balance settings should also be the same. In particular, if the dark frames used a different offset then that will lead to issues with the dark subtracted image having a histogram peak that is too far left (or right).

cheers,

Robin