No Bayer options seem to work for my D5300

[Andrew Klaassen]

Hi,

I believe I have tried every Bayer option for NEF files coming out of my Nikon D5300, but all of them give me wonky colours. I'm wondering if there's a setting I've missed somewhere.

Here's what the colours should look like:

https://drive.google.com/file/d/1InjjHc5JD_8RTXcsCv2kd6gFgU9Swx29/view

Here's what they look like with the various Bayer pattern options:

https://drive.google.com/file/d/1cseYtSgl3DyV-ZPC_PbHOgAzs_oPj1Jy/view

I also tried with all of the "De-mosaic method" options (AstroC, AstroM, AstroSimple, Bilinear, SuperPixel) and "Raw conversion" options (LibRaw full area, Libraw cropped area, Dcraw), and got the same results for each of these Bayer patterns.

Is there something else I can try?

Thanks.

[Axel Thomas]

Hi,
similar to what I encounter with a D5100. The images have a strong green background. I have not been able to try any Bayer patterns as the Test pattern button is always inactive.
Any help is appreciated!

Axel

[Andrew Klaassen]

I have found a couple of potentially useful things:

• In the main viewer window, I had set Minimum and Maximum to something other than 0 and max, and that was screwing up what I was seeing. Dragging those slider all the way out at least got my colours to some approximation-ish of the original.
• To do another test, you have to Ctrl-Z after you do a test. (You can also reload the image.)

What I'm trying now is seeing how well it does if I use the free NX Studio application from Nikon to export as TIFF. Would still be nice to get it fully working in ASTAP, though.

[han.k]

Hi Andrew,

In auto or RGGB is looks okay. I assume the colours will becomes more normal if you remove the stretching in the viewer and place the minimum at zero. See attached. The yellow building door should become more yellow then greenish. Zero for minimum is required because the program tries to find the average (sky) background to subtract. This doesn't work well for an image of a building.

Once you take an astronomical image, the option "Auto leveal" will adjust the colours to average white after debayering.

If this option auto level is not used all colours are represented with equal gain and images like you used for testing could be a little out of colour balance.

You could share the raw and I will have a look.

Han

[Andrew Klaassen]

Thanks, not stretching does seem to help a lot. Here's what it looks like now:

https://drive.google.com/file/d/1U2ow9jMbsz6GITG5Ey4nyUY_u13raGng/view

Is that about what I should expect? I've attached the original NEF raw file - thanks for taking a look.

[han.k]

Astronomical images have a high background value mainly caused by light pollution. You want to subtract that. Furthermore the signal of deep-sky object is very weak. The stretching is optimized for that.

[han.k]

Yes. you can make it a little better by moving the minimum slider a little up to the first peak of the histogram and stretch a little, maybe 4. See attached.

There is not individual adjustment for R, G, B stretch but the auto level will do that by making the stars average white. Any fine tuning you could do in Photoshop/Gimp

Han

[Andrew Klaassen]

Thanks for the extra info. One more question coming out of that: Does my manual stretching affect how the images are processed during stacking? Or does the stacking process do its own automatic stretching?

[han.k]

For info, here a description what a camera would do for terrestial images. This is not used for astronomical images:

https://www.nxp.com/docs/en/application-note/AN1904.pdf

[han.k]

The stretching influences only the canvas/view. The image itself is not modified. If you save the FITS there is no stretching. If you export as 16 bit TIF/PNG could can selectan unmodified or stretched version in the save options. So you can further process the image in an other program

If you save as JPG, PNG and BMP (ctrl+J ) then you get a 8 bit image which is a copy of the view /canvas. Use this if the image is fine and final.

Han