SXVR-25MC - DeBayer Matrix

[Aidi]

Hello there,

Could someone confirm what the correct bayer matrix selection is for the SXVR 25MC camera?

I am reading conflicting information as to what the correct setting should be when going through the debayering process.

Rgds
Aidan

[aworonow]

Usually, if you try them all, one will be obviously correct and the others not so much

Alex

[Aidi]

I did try that.  Two gave me red or blue screen.  The other two were pretty much the same

[Niall Saunders]

Try printing three large squares of colour onto a sheet of paper, one in pure Red, one in oure Green and one in pure Blue.  Now, hold each of these - individually - over the camera aperture whilst pointing at a relatively 'white' light source, and take a RAW exposure

Load the threeimages into PixInsight and zoom into the top-left corner of each image. There you will see the individual pixels that make up the RAW (greyscale) Bayered image, as acquired through the Colour Filter Array (the CFA). The naming convention is something like "RGBG" which, in this example, would have the top-left pixel under a Red filter. Next pixel (to the right, but still on the top row) would be a Green pixel. Then, moving on to the second row, this CFA example would have a Blue pixel in the first column, and another Green pixel in the second column. This pattern then repeats over the entire image - and you can see that there are more Green pixels than Red or Blue pixels (not unusual for many CFA patterns).

So, if you are looking at the image through the Red sheet, the example above would have a much brighter pixel  at the first column of the first row. Cobversely, the pixel immediately below this would be, relatively, much darker. On the Blue-exposed image, this situation should be completely reversed - darker first pixel of the first row with a brighter pixel at the start of the second row. Finally, the Green-exposed image should have pixels of a similar brightness as the second pixels on both the first and second rows.

Note: Just be sure that you have definitely zoomed to the maximum level possible, and that you have either panned or scrolled the whole way to the very top left of every image.


It is just possible that your imager is using a CFA that is not one of the four 'common' matrices that PixInsight can already deBayer - a future release of PixInsight (due shortly) will offer the 'missing' four. However, as I have no experience of your imager, all I can say is try the suggestions offered and see how you get on.

Don't hesitate to get back to us if you are still struggling.

[Jredmond]

I have the same camera.  I wondered if all SXVR M25C's had the same Bayer Matrix.  I came across a post by Terry Platt of StarlightXpress explaining that each individual camera could be different because the individual sensor arrays would not necessarily have the same color in the same spot, for example, the upper left corner.  I had the same problem figuring it out that you reported, but finally tried the method Terry recommended in the manual: take a daylight picture of a an outdoor scene with green trees, blue sky, a red object.  I found in my case the pattern is not RGGB but GRBG.  That pattern may not apply to your camera

John.

[Niall Saunders]

I came across a post by Terry Platt of StarlightXpress explaining that each individual camera could be different

Well John, if that really is the case with Starlight Xpress imagers, it would by a very good reason not to invest in one!!

At the end of the day, the CFA is associated directly with the imaging sensor itself - not the electronics supporting and interfacing that sensor within any given imager. Companies like StarlightXpress usually 'buy in' a sensor from a CCD/CMOS sensor manufacturer (such as Sony and Kodak, to name but two). It is the manufacturer who specifies the CFA location with respect to the pixel locations.

However, that might not stop the imager manufacturer from changing their internal design - placing the actual i.c. in a different orientation without making internal software changes to compensate for the physical changes.

But - and this is more common than you might think - it can often be the image-capture software that is 'at fault', where images are stored (in FITS and TIFF files) such that any hope of relying on a given CFA layout becomes meaningless.

It is then that users have to carry out experiments to determine which CGA pattern has been used (as you described in your post). Of course, what then happens is that the user jumps through all the necessary hoops to establish the 'correct' CFA - and then changes to a different suite of image capture software.

Now they have to repeat the process all over again!

[Aidi]

After some trial and error and assistance from a fellow user of the same camera, mine debayers correctly with the GRBG matrix.