Stacked image has grid like pattern

[Cey42]

I was running 4.0.8949 and decided to upgrade to 4.0.9132.

Camera is ZWO 294MM. The images below were reproduced using Folder Monitor feature. Near a full moon causing gradient.

When I got to my first target and started stacking I noticed a grid like pattern (more obvious when you run this in SharpCap). See image below.

9132.jpg (521.04 KiB) Viewed 632 times

I thought I had made an error in the flat capture, so I redid the flat but I continued to see the issue.

I then downgraded to 4.08.8949, used the same flat and got the below. No change in histogram. Which looks much smoother.

8949.jpg (374.5 KiB) Viewed 632 times

You can get fits, flat and dark files here, https://1drv.ms/u/s!AmzOfpCoABCRgdVK0-S ... g?e=PGTlIY

Thanks,
Cey

[admin]

Hi,

thanks for the report - I have had a dig into it and the good news is that it is not an issue with the images themselves or with the flat/dark correction process, it's just a display issue to do with the way the image is scaled when the scale factor is less than 100%. In fact you can see the grid without dark/flat/live stacking, it's just less distinct because you can't stretch the image as much without correcting the vignetting.

The newer version of SharpCap is using a different technology to display the image and the scaling options work a bit different.

Right now I am using a 'nearest neighbour' option to give the blocky zoomed in pixel look when the zoom is >200% and a 'linear' blending option for zooms below 200%. If I change from 'linear' to 'high quality' then all the grid effects vanish, but I am worried about the impact of that for high speed imaging work (high fps). I will have to try to arrange the code so that 'high quality' can be used for longer exposures and the simpler approach can be used for high speed imaging.

Anyway, if you look at the saved images you should not see the grid (or if you do, it indicates that your viewing application is suffering from the same thing).

cheers,

Robin

[Cey42]

Thanks for the explanation. I have played with .9132 and did see he blockiness go away at 100%. Seems worse between 25-33%. At least now I know how to work around it.

Hopefully you can figure out how to use the old algo again.

Thanks,
Cey

[admin]

Hi,

I've made the code switch over between 'fast' and 'good' automatically at about 4fps (faster frame rates get 'fast', slower gets 'good'). I think that the effect is actually coming from a very faint grid shaped brightness pattern in the pixel data from the sensor which is normally invisible, but a combination of strong image stretching and certain zoom levels make it noticeable.

cheers,

Robin

[Cey42]

Thanks for 'fixing' this. Look forward to testing it out.

Cey

[Cey42]

Fix looks great.

Thanks,
Cey