Problems with flats

[MikeB838]

Hi Robin,

before I start a session, I take darks and flats. By using them during capturing (Live Stack), all was okay (until version 229 incl.).
Starting with 251 my dust donuts are no longer eliminated and are visible. I thought I did sth. wrong (eg. having an old dark
frame active during flat frame capturing), but I checked it last evening with version 285 and I think I did it correctly, but the
dust donuts are still visible (with using the flat frame).

Are there some changes or have I to do sth. different with the new versions?

In case: what information or files do you need?

Thanks in advance!

cheers,
Mike

[admin]

Hi Mike,

I can't see anything changed in the code between those two versions that could affect flat correction (or at least nothing that looks like it should affect it). It might be worth checking the camera settings files of your flats, darks and raw frames to see if you have changed something that will throw the corection off. Such things might include

* Using the 'flip' controls on a camera
* Adjusting the offset/black level/brightness control to a new value
* Changing camera colour balance settings.

If none of that helps, then the next step would be to test the flat correction in the two different versions using the folder monitor camera, some saved raw images and your usual darks/flats. Either you could do it, or send me a set of files to test. Using the folder monitor camera and previously saved files should give the same results in both versions if nothing has changed in SharpCap - if it doesn't then I will need to dig deeper into what might have changed.

One other thing to note is that flat correction is rarely perfect - usually some remnant of the dust donut/vignetting is left behind. If you have started using a stronger live stacking stretch than before then that can make this sort of thing visible when it was there before, but too faint to see.

cheers,

Robin

[MikeB838]

Hi Robin,

thanks a lot for your explanation!

I compared the setting files and see that darks and flats have DisplayStretchEnable=1, my frames have DisplayStretchEnable=0
But this was the same, when I could not see the donuts in the results.

Surely I change the settings in the Histogram tab (mostly black and mid level changes), to get a good view and see if my target (mostly Asteroids)
are visible in the frames. Sometimes I adjust the colour, but seldom. But that usage has not changed lately.

I will take a closer look the next days. If I get no results, I may ask you to look at my frames, thank you very much!

Cheers,
Mike

[admin]

Hi Mike,

the display stretch values aren't important in this respect, as those controls only affect the processing of the image when it is about to be shown on screen in SharpCap, so no effect on the underlying image that is captured, processed and saved.

cheers,

Robin

[MikeB838]

Hi Robin,

I do not change exposure, gain, offset and colour balance (besides auto colour balance in the Histogram tab in live stack) during my session.
And flip is: None
But I see the same behaviour in SC 285 and SC 229. So it seems to be by chance, that it looks okay in 229 for me.

In the saved *.png (at the end of the session), no donut is visible. When I reprocess the session in the folder monitor, I see a donut. But this is when I
adjust the mid level slider a lot to the left, to bring out fine details.

I experimented with the offset a long time ago and had set it to 60. The next time at the telescope I will check, if I can adjust it for the darks.

I add M81 the inverted PNG and a Screenshot (Mid level slider too far to the left to show the donut)

M81_inverted_Stack_21frames_630s.jpg (155.31 KiB) Viewed 4854 times

m81_folder_monitor.jpg (281.65 KiB) Viewed 4854 times

But I think we can close this for now!

Thank you very much for your support!

Cheers,
Mike

[MikeB838]

Hi Robin,

I have to open my thread again. Last night at my telescope (with SC 319) all looked very well (regarding flats).
Today I used SC 331 on my desktop PC to help me process my stacks to check Asteroid positions in Astrometrica. And the donuts are back again.

Both stacks (at scope and in Folder Monitor) are the addition of 5 frames 30 seconds each.

I fiddled the settings (Folder monitor) to have nearly the same settings as last night, but the donuts stayed.

I copied dark and flat to my desktop machine and select the same as at the scope.

For me it seems that in the Folder Monitor code there is still some problem or it seems to be handled differently there.
In the code for the Live Stacking all seems well.

May be you can take a look? Thanks a lot in advance!

Here I add both screenshots and Camera setting files of the stack:

Stack created at scope with live stacking

Stack.jpg (570.79 KiB) Viewed 3993 times

Stack created in folder monitor

foldermonitor.jpg (570.73 KiB) Viewed 3993 times

I add flat and dark here too. The donut at lower left in the flat looks really dark, but is not saturated. If it is overcorrecting then I would expect it at the scope session too.

Master Flat

masterflat.jpg (345.85 KiB) Viewed 3993 times

Master Dark

masterdark.jpg (578.56 KiB) Viewed 3993 times

If I can check something else, please let me know.

Cheers,
Michael

[kaymann]

+1 Same here...

[admin]

Hi,

OK, I can look into this. Can you share with me the following files...

1) Master flat file and the capture settings that goes with it
2) Master dark file and the matching capture settings
3) About 3-5 image files (raw images, unprocessed) and their capture settings files.

Also, let me know whether you used bias, dark flats or nothing when capturing the master flat.

cheers,

Robin

[MikeB838]

Hi Robin,

thank you very much for looking into this.

The needed files are at:
http://www.twus.de/data.zip

In folder 4387 (Asteroid Tanaka) are the raw files of the stack.

I selected "None" as Black Level offset correction when capturing the flats.

I hope this helps

cheers,
Michael

[Menno555]

No answer here but I tried stacking the RAW captures and master flat and master dark in Siril and the stack is fine.
SC Folder Monitor gives indeed the results Michael is describing.

Menno