Mystery RGB matrix

[Josh Lake]

I'm helping an online friend process his data in PI (always trying to spread the good word!  8)) and I got a full data set from him, uploaded to the Endor share a jlake_mysteryRGBmatrix.

The data was shot with a QSI 583c, so some debayering will be involved. The shooter wasn't quite sure what the matrix was, but I think it might be GRBG.

The 'mystery' comes from the fact that there is very little color in the image (shots were only 180 s), so trying all 4 debayer options doesn't lead to a clear answer. I've gone ahead and processed a full image the normal way using both the RGGB and GRBG options... and both ways led to the attached -- a green bar in the Cigar instead of a red one.

So obviously my Debayer matrix is incorrect, but I don't know of an easy way to figure out what it is. This is complicated by the fact that I used Kai and Juan's wonderful Batch Calibration script, and I'm not sure if the "Up-bottom FITS" option is correct for this one... I think it is normally on for most imagers.

Any help figuring out the proper matrix and trying to get some color out of this image would be wonderful!

[chris.bailey]

Normally the wrong debayer pattern will leave a bayer pattern behind when you zoom right in. The right one will just show pixels with no grid.

OSC images are often quite washed out requiring a fairly have curves stretch on Saturation.

[Josh Lake]

Good tip, I'll try zooming in on a single frame.

Edit, after checking:

That's odd, 2 of the options seem to just show the pixels and 2 show the grid. In my experience, it's been 1 good, 3 bad, so I'm a bit confused on this one... 

[chris.bailey]

Of the two options that work, one will be a bit Joseph and his amazing Technicolouresque and the other will look a bit washed up but with colour calibration will show expected colours. Takes me ages every time I swap cameras to remember which matrix it should be even though there are only four options!
 

[Josh Lake]

Hmm, so for me, that would be BGGR instead of RGGB. It seems odd, then, that green would end up where red was... okay, I'll give it a shot.

[chris.bailey]

You also need to know whihc way up the FITS file was created against how it is read into PI. Odd results in debayering can also be caused by the file being upside-down so effectively you have 8 options. Have to say I resorted to a daylight image to fully sort out my OSC.