Debayering?

[magnusl]

Hi!

I'm a bit curious about debayering in PI. I'm rather new to PI, and have learned to debayer my DSLR subs after calibration, and before registration and integration. It works nicely for me. However, I also hear and read in some places about people who do NOT debayer. They calibrate, register and integrate without any such step.

Maybe there is something with the setting for raw files involved in this. I use the Pure Raw, setting, and with that my cr2-files are displayed as grey. However, others seems to use the default "e-Bayer RGI" setting, and seems to get nice results without debayering.

Anyone here who can clarify this, and the place for debayering in current PI?

Magnus

[pfile]

well, if those people are letting DCRAW debayer the file as it is read from disk, then they Are Doing It Wrong. the only way to properly calibrate an image from an OSC is to do it while it is "raw", that is, not debayered.

the results might look OK or actually be OK as far as those folks are concerned, but according to best practices, debayering before calibration is incorrect.

or, maybe you are talking about people doing Bayer Drizzle, but from the description, probably not?

rob

[magnusl]

Hi!

Actually, I am talking about people NOT debayering at all. For instance, have a look at this youtube tutorial by Richard Bloch (very good, as I see it) - where there is no debayering (and have a look at the comments below too): https://www.youtube.com/watch?v=zU5jJgjKuQQ.

This question comes off a discussion on a swedish list, where the same issue turned up - and we are not able to sort the debayering issue out. If cal - debayering - registration and stacking is correct, how come some people produce at least reasonalby good results with cal - registration - stacking and no debayering? What is PI doing in those situations, that we seem to have a hard time figuring out?

Best,

Magnus

[Cosmick]

Actually, I am talking about people NOT debayering at all. For instance, have a look at this youtube tutorial by Richard Bloch (very good, as I see it) - where there is no debayering (and have a look at the comments below too): https://www.youtube.com/watch?v=zU5jJgjKuQQ.

Digital cameras only take greyscale images. You only get the colour once it has been debayered. He is starting with already debayered files (RGB in the image window). He may not have debayered the images but his camera or software have.      ;-)

[magnusl]

Hi!

Right. Are you saying that PI has definitely not done it?

I don't really understand what the format setting for RAW means. I have understood it as only making a difference for display, not for the actual images - but if so, why is it important for me to set it to Pure Raw?

So my thinking goes: if set to "e-Bayer RGI", could that make "manual" debayering unnecessary (as these examples might imply), or is it a question, as you say, of another process doing the debayering, for instance the camera?

I'm sorry if I sound a bit unclear here, but this is more a question of trying to understand something, more than having a working solution... And I guess when I understand it better, I can also formulate the questions better......

Magnus

[Cosmick]

Right. Are you saying that PI has definitely not done it?

No. I am saying that it has been done either by the camera (eg. Saving as 16 bit tiff) or by Pixinsight or some other software.
If I remember correctly the defaults in Pixinsight can be set to open RAW files from digital cameras as RGB colour images. This just means that they are being debayered in the background as they are being opened.

I don't really understand what the format setting for RAW means. I have understood it as only making a difference for display, not for the actual images - but if so, why is it important for me to set it to Pure Raw?

The RAW format settings are telling Pixinsight in which format you want to open and work with the image - either as a greyscale image with the bayer matrix intact or as a debayered RGB image. For image calibration the correct option is as a greyscale image with the bayer matrix intact (as mentioned above by Rob).