OSC cam: CFA pattern visible in bias and dark frames...? => solved.

[limburg11]

Hello to the forum,

my name is Bernd and I am new to PixInsight and new to serious astrophotography.

I just got my new ASI294MC Pro OSC camera and started generating bias and dark frames (for getting master bias, super bias, master dark). I came across the following strange behaviour when I looked at a single bias frame (see attachment, a PI screenshot, don't know yet how to integrate the attached image into the post). When I stretch and zoom in, I can clearly see the CFA pattern.

How is that possible? There is no light interfering with the RGB filters, so all pixels should be equal (with noise, of course). The dark frames show the same pattern.

Thanks for any clarification and best regards,
Bernd

[bulrichl]

Hi Bernd,

I responded to your post in Astronomie.de.

Bernd

[limburg11]

Hi Bernd,
just read it. Thanks!
If I find the reason for this obviously wrong bias frame, I will post it here.
Regards,
Bernd

[limburg11]

Found the reason for the implausible bias frames.

The ASI driver in the Linux/Raspi version of Indilib has default values for settings called WB_R and WB_B. Seem to be some white balance factors, running from 0 to 100, unity = 50. Default was WB_R = 52 and WB_B = 95. So I think the driver multiplied a RGGB gain vector (52, 50, 50, 95)/50 = (1.04, 1.00, 1.00, 1.90) to the pixels prior to saving it to the FITS file. Therefore, the blue pixels were overemphasised by a factor of 1.9, boosting them in intensity and made them clearly stand out of the noise floor of the bias frame. So it was not really the Bayer pattern that could be seen in my image but the positions of the blue pixels in the CFA because they were almost twice as bright as the rest.

I corrected the values of WB_R and WB_B to 50, respectively, and took another bias frame. And voilá, it looks as it should.

CS, Bernd

[bulrichl]

As I wrote in your Astronomie.de thread, I cannot imagine that a camera driver modifies the raw intensity values by applying different scaling factors to the color channels of an OSC camera. I suppose that the capturing software writes modified values to the FITS file. In any case, this is a serious bug in either the driver or the capturing software.

Bernd

[limburg11]

I placed this question in both the Indilib and ZWO manufacturer forums. No answer as of now, but I‘ll keep trying and possibly get it fixed.

(The other) Bernd

[wvanreeven]

The ZWO website states that they are out of office until January 31 due to the Chines new year. I suppose you will receive an answer from them after that.


Wouter