Newbie DSLR - What am I seeing on the screen, CFA, RGB, SOS?

[gsedun]

It's all a bit confusing - I'm sure you've heard that before.  PixInsight truly is a great tool and I'm looking forward to getting to know it better over the years.  However, I'm a bit baffled by when to do things with DSLR Images.  I've searched around a lot on the forum and Resources but I am still a bit uncertain of what I am seeing on the screen with DSLR data (I am totally fine with LRGB data is fine - I've been doing that for years).

When I calibrate the original RAW file I get a  Calibrated FIT   file.  This file, apparently is not a real RGB file but a CFA file? (whatever that is - I can't find an explanation for it anywhere.)

Why am I supposed to debayer the image file AFTER calibrating?  Shouldn't that be done before calibration so the separate Cal files can be applied or does PI separate the different channels in the RAW file and calibrate them separately?  Some folks are writing modules to separate the channels out from the RAW files first, which seems to make sense but that isn't what others do.

Does PI actually create three separate RGB  Biases, Darks and Flats when using RAW images to create them and then apply them separately each channel?

Debayering creates separate RGB Channels ( correct?)  but are these combined on the screen to so I can VIEW a color image while running the Subframe Selector script or any other process, for example?

I have to separate out the RGB channels after I've made my final stack and I then have to combine them into an RGB image before I can do any processing? 

IT would be really great to have a simple explanation about all of this that is HIGHLIGHTED as a DLSR USERS READ THIS FIRST file somewhere.

Thanks

Garry

[TobiasLindemann]

Hi Garry,
what you see is the color filter array (cfa) of your sensor http://images.google.de/imgres?imgurl=https://upload.wikimedia.org/wikipedia/commons/thumb/3/37/Bayer_pattern_on_sensor.svg/2000px-Bayer_pattern_on_sensor.svg.png&imgrefurl=https://en.wikipedia.org/wiki/Color_filter_array&h=1300&w=2000&tbnid=F0qG93c0gx5klM:&tbnh=90&tbnw=138&docid=TIKiUliKl8EgmM&usg=__7lYw5pFvxxwUCGgdKFFA8p3wck8=&sa=X&ved=0ahUKEwjx6O2B6OHNAhVDiRoKHf-yAAwQ9QEIQDAD
The Raw-file is saved as a greyscale image. With the default settings, PI lays the cfa over your image. This looks a little bit weird, but this is the way, the raw-file looks like. For better viewing, you can change some settings in the file-explorer to prevent PI from lay the color filters over your image. So you have a greyscale image, which is still raw, but looks not so weird.
In the FormatExplorer double-click on the DSLR-RAW and then select "pure raw"

For calibration you should use the Raw-files. Debayering calculates a rgb-value for each pixel and interpolates the neighbor pixels for that. So the debayered pixels are not the real pixels, but you need the real pixel values for calibration. https://pixinsight.com/doc/tools/Debayer/Debayer.html

You don't have to split the channels. Just calibrate with the undebayered raw files, then debayer it and than do the StarAlignment and so on. If you use the PreprocessingScript, it does everything automatically in the right way.

Greetings
Tobias

[pfile]

the splitting CFA channels is something someone might do if they have used a DSLR with an Ha filter. in that case there's only meaningful data in the red pixels of the bayer matrix. in order to save processing time and space, you can extract the red pixels out of your lights and calibration frames early on, so that you're not wasting resources working on the G and B channels which have very little data.

rob

[gsedun]

Thanks guys - it's starting to make sense.

Garry