Image calibration of "Pure Raw" files

Luca Amerio

Member
Aug 26, 2017
18
2
Hi everybody

I'm calibrating a bunch of dark frames using Image Calibration. Under [Format Explorer] > [Raw], I set it to [Pure Raw].

Since my Superbias is not debayered and I want my master dark to be as well, should I enable the [Enable CFA] option in the Image Calibration process?

The problem is probably that I did not understand what this option does. Is it just a setting to optimize noise evaluation etc, or does it means that the output will be debayered?
 

pfile

PTeam Member
Nov 23, 2009
6,170
180
you can but it is not recommended. the noise in the dark subs can cause some pixels to be less than the corresponding master bias pixel and then you end up with a bunch of pixels clamped to 0 when you calibrate the dark subs. that's bad. when you integrate the uncalibrated darks first, the SNR is boosted and the residual noise is not enough to make any pixels in the dark less than the corresponding pixels in the master bias.
 

Luca Amerio

Member
Aug 26, 2017
18
2
It actually makes very sense. I mean, the problem lies in the fact that the value of the pixels is unsigned. I actually never thought about it that way.

If I understood correctly, you then calibrate your master dark using master bias afterward, do you? Or do you just use an uncalibrated master dark to calibrate your lights? In case, if you do calibrate it, what process do you use? Always ImageCalibration?

I will now real Bulrichl guide to see the whole process. Maybe I will find myself the answers there.
Thank you for the insight.

PS: any idea regarding the original question? What does that “Enable CFA” tick does?
 

Luca Amerio

Member
Aug 26, 2017
18
2
There are tutorials that suggest to pre-calibrate the individual dark frames or the integrated dark frames AND don't even give a reference to truncation of negative values or how to avoid clipping in the pre-calibrated MD (namely by the application of a pedestal). Even the otherwise excellent tutorial [10] does. This is bad advice.
I always followed EXACTLY that tutorial 😅

Anyway, I’m reading your guide and it’s the best piece of explanation I ever found on the topic. Its exhaustiveness is astonishing.
Thank you so much for this effort and sorry for opening the n-th thread on the calibration topic.

BUUUUTTT my original question was about what the “Enable CFA” tick in the ImageCalibration process does :) I’m still keen on understanding that.
 

ngc1535

PixInsight Ambassador
Feb 1, 2014
655
106
AdamBlockStudios.com
BUUUUTTT my original question was about what the “Enable CFA” tick in the ImageCalibration process does :) I’m still keen on understanding that.
I am going to take a stab at the answer. The Flat can be scaled on a per channel basis (look at the Flats section). This requires the CFA pattern to be known. There might be another reason? Not certain. If there isn't another reason..maybe this check should be the Flats section. But there probably *is* a reason.

-adam
 

pfile

PTeam Member
Nov 23, 2009
6,170
180
Screen Shot 2021-01-22 at 1.28.26 PM.png

according to the source code this tickbox seems to trigger a bunch of checks to make sure the bayer pattern "makes sense" (meaning, is not something nonsensical like CMYK, but is either X-trans or one of the legal permutations of R/G/B) and also enables checking consistency of the CFA pattern between the lights/flats/darks during calibration. it used to be that the bayer pattern from the input files would get propagated to the output files before this checkbox existed, but now maybe that doesn't happen unless the box is checked. i'd have to experiment since i don't clearly see the code path for keyword propagation.

also with respect to what adam is saying, if you untick "enable CFA" then the "Separate CFA flat scaling factors" checkbox is disabled. so they are definitely related.

rob
 
  • Like
Reactions: Luca Amerio