Changes in pixels during cooling?

[fwm891]

Started out last night doing a guiding assist check in PHD2 before trying a few subs on M27. I captured in SGP and decided to set my usual cooling on an ASi533 OSC at -12° C and just start exposing while the sensor cooled. I noticed that some of the pixels were bordered by a black ring - very worried at tis point as I'd not seen anything like this before. Decided to let the captures continue and noticed that the rings faded. They didn't completely disappear until well after the sensor was showing it was at -12° C.

First 3 frames are sequential final frames some 10 mins after the others. All subs are 5 mins.

Anyone seen this effect before.

Edited July 22, 2021 by fwm891

[vlaiv]

That is most likely artifact from display interpolation.

If image was scaled, capture application probably used some sort of interpolation algorithm to do the scaling. Some interpolation algorithms don't deal particularly good with lone pixels that have big variation with respect to surroundings. Say you have hot pixel (that goes away with cooling) - then it will have value much larger than surrounding pixels.

Look what happens with a single hot pixel when image size is reduced by 50% using bicubic interpolation:

[markse68]

I get very strange patterns across the frame after cooling and have to leave it a while to settle or scrap the first bunch of subs. But mines an old ccd so probably a different thing- maybe frosting

Mark