The “preprocess -cfa” works great and it was more of a theoretical question.
Unfortunately you didn’t explain the difference between “seqfind_cosme_cfa” and “preprocess -cfa” to me.
I want to give you some more feedback:
I use an ASI183MC CFA non cooled camera and there are also some hot pixel clusters.
While Siril removes them automatically, not even Pixinsight does it automatically (i.e. without DefectMap)
see this sample of 10 Lights and 10 Darks (Gain 200, 60 seconds, 25°C sensor)
top line are single images (RGB corrected with DarkFrame)
bottom line are stacks of 10 images
Bold is the regular Siril workflow
(1) single frame of ASI183MC, not debayered raw cfa image
(2) Image (1) processed with Siril commands (DeBayer RCD)
preprocess light -dark=dark_stacked
seqfind_cosme_cfa pp_light_ 3 3
preprocess cc_pp_light -debayer
(3) Image (1) processed with Siril command (DeBayer RCD)
preprocess light -dark=dark_stacked -cfa -debayer
(4) Image (1) processed with PixInsight Weighted Batch Preprocessing (WBPP, DeBayer VNG) and CosmeticCorrection (Auto Detect Hot Sigma=3, Cold Sigma =3)
(5) Siril Stack of 10 Images from (2)
stack r_bkg_pp_cc_pp_light rej 3 3 -norm=addscale -out=…/result
(6) Siril Stack of 10 Images from (3)
stack r_bkg_pp_light rej 3 3 -norm=addscale -out=…/result
(7) PixInsight stack of 10 Images from (4) with WBPP
my comment:
(5) You can clearly see that seqfind_cosme_cfa generates darker correction pixels. If you stack more images, the effect becomes even more pronounced. The pixel cluster was not removed well.
(6) Everything is great.
(7) The pixel cluster was not removed well. These green spots are even more noticeable with larger stacks.