Images showing up Green!

[ANKulin]

I am not very experienced with PixInsight, having played around with it infrequently now and then. I built a new PC last summer and just re-installed it on the compouter this past week trying to process a wide-field capture of Orion and area. What am I doing wrong here?

I am attaching three screen captures:

1. RAW Light Image of Orion (CR2, from a Canon 60D with IR filter removed - so it has an overall tint to it)
2. Regular photograph of a Sunset (taken with a Canon 5D3 - no mods so a regular camera)
3. A screen capture taken from PixInsight with three photos.
   1. The Raw Orion Capture at the left with no processing at all (so basically photo from bullet 1 in this list)
   2. The sunset photo in the middle (it's rotated in PixInsight)
   3. A processed (calibrated) version of the same Orion CR2 file at the right
4. The three photos are presented with grey scale versions at the top (non-debayered) and the debayered images below

With debayering, everything is coming out with green tints. Even non-processed raw files.

Hopefully it is just a setting somewhere that I have set incorrectly.

Any ideas??

 

[nico1038]

Hello @ANKulin

These are not unexpected results. Your raw images are not color calibrated.
If you want to get more reasonable colors before color calibration try to unlink the R,G,B channels in STF.

 

[fredvanner]

Actually, you have the correct settings and the images should look green before colour calibration. Most non-astro applications will load raw files with the camera built-in white balance as usually used for terrestrial photos. Astro applications like PixInsight are usualy configured to load images with the raw sensor signal values. Since virually all colour camera sensors are most sensitive at green wavelengths, the raw images will look greenish when debayered. This is corrected after preprocessing (calibration; debayering; registration) and integration, using a process such as SpectroPhotometricColorCalibration (SPCC).

 

[pfile]

it's normal... due to the 2 sets of green pixels per bayer quad, the green SNR is higher and thus under autostf images will sometimes appear green. for terrestrial photographs you'll have to figure out how to apply the canon white balance to the images by hand. for astronomical images, you can use SPCC to color calibrate your images and the green cast will be gone.

if you unlink the channels in STF you might see something that looks a little more normal but this is no substitute for proper color calibration.

rob

 

[pfile]

wow i suspect we just tested the database locking capabilities of the forum software, lol.

 

[fredvanner]

due to the 2 sets of green pixels per bayer quad,

The green cast has nothing to do with the two green cells per bayer group (this is an urban myth that I have tried to exorcise before); it is entirely to do with the QE of the sensor at different wavelengths.

 

[ANKulin]

Wow. First off I am impressed that I received 5 responses in something like 15 minutes. It must be cloudy everywhere.

So it sounds like I have a ways to go yet before I get to colour calibration and that I should have no problem carrying on with the processing in spite of the green-ness of these images after debayering (we'll see about that of course, as I am sure that I can screw something up still!). And that once I do the colour calibration I should get something natural looking after all that effort.

 

[fredvanner]

in spite of the green-ness of these images after debayering

Note @pfile's comment above - selecting an unlinked STF (Ctrl-Clk on

) will display a reasonably "normal" colour rendering of the image until you colour calibrate it (after which a linked STF will display the corrected colour).

 

[fredvanner]

The green cast has nothing to do with the two green cells per bayer group (this is an urban myth that I have tried to exorcise before); it is entirely to do with the QE of the sensor at different wavelengths.

If you don't believe me, try this:

• use IMAGE > New image to create a uniform 0.5 neutral grey greyscale view (e.g. 1024 x 1024);
• debayer it with "RGGB" selected.

I'll bet the result is neutral grey (in spite of the 2:1 green cell ratio).

 

[cloudbait]

The green cast has nothing to do with the two green cells per bayer group (this is an urban myth that I have tried to exorcise before); it is entirely to do with the QE of the sensor at different wavelengths.

I image through separate RGB filters, and most unprocessed (channel combined) masters are green (although STF makes odd decisions sometimes, so I can get blue or red occasionally).

 

[pfile]

The green cast has nothing to do with the two green cells per bayer group (this is an urban myth that I have tried to exorcise before); it is entirely to do with the QE of the sensor at different wavelengths.

fair enough, had forgotten about this

 

[fredvanner]

I image through separate RGB filters, and most unprocessed (channel combined) masters are green

Of course, you have the option (not available with OSC) to compensate by using different exposures in the different filters, so you situation is more flexible.

 

[cloudbait]

Of course, you have the option (not available with OSC) to compensate by using different exposures in the different filters, so you situation is more flexible.

Sure, but I never do. No compelling reason. (This wasn't a complaint... I was just pointing out that the behavior is normal, and not only found with OSC cameras having doubled green pixel counts. As you have explained.)