experimental ‘Dark Scaling’ option

[Jean-Francois]

Hello Robin,

I install the last version with the new experimental ‘Dark Scaling’ option.

Simple question ... how to use it ?

I have the sensor analysis for my camera ... and I did not remark a difference.

Jean-Francois

[Jean-Francois]

Hello Robin,

Thanks for the detailed description viewtopic.php?p=45807#p45807

I repeat with a second camera ... it works. Then with the first camera and it does not work with it.

OK with QHY600.
Not OK with QHY174GPS.

Regards,
Jean-Francois

Here with the QHY174GPS camera:

DarkFrameScaling_1.png

[admin]

Hi,

let me give that a test with my QHY174 to see if I can repro the issue of the control not appearing. I did have a bug where it worked if you selected the dark after changing to 16 bit mode, but not if you selected the dark first, but I thought I had fixed that fully before yesterday's version.

cheers,

Robin

[admin]

Hi,

OK, it turned out to be a mis-handling of the ratio between the true ADU of the camera (ie 12 bit sensor in range 0..4095) and the image ADU (16 bit - 0..65535). That meant that for 16 bit sensors everything worked fine, but for 12/14 bit sensors the error crept in. Guess who ended up testing on a 16 bit sensor

I have a fix for this now - please try this version : https://d.sharpcap.co.uk/download.html? ... 0&arch=x64

Note that version also contains some unrelated internal changes to focuser scans that *should* be OK, but if you encounter new focuser related issues with that version then a) let me know and b) go back to the previous version!

cheers,

Robin

[Jean-Francois]

Hello Robin,

QHY174GPS works now in 16 bit. But not in 8 bit modus.

In 16 bit ... I have the following:
- dark with 10 ms (OK, that is not really a "dark" ... but fine for testing).
- light frame is 100 or 1000 ms ... the scaling is most of the time below 0.1x but sometimes above 1x
- do you set the threshold at 10% of the pixels ? ... the scaling changed from 0.062 to 1.086

Jean-Francois

DarkFrameScaling_2.png

DarkFrameScaling_3.png

[admin]

Hi,

8 bit mode is deliberately excluded - the loss of sensitivity in pixel values would mean there is relatively little data to calculate the scaling from. Anyway, there is no real serious use case for long exposures (needing darks) in 8 bit mode.

I'd suggest trying longer exposures - 1s or more. The scaling is calculated by looking at 'warm' pixel brightnesses, which tend to scale linearly with exposure (and relative gain, and non-linearly but predictably with temperature). In a 10ms image, there's not really much time for that signal to build, so the scaling will be pretty unreliable and often give rubbish values or low correlations.

cheers,

Robin