Hello everyone,
this is my first post here, I hope it is appropriate.
I am still trying to find a solid and repeatable LRGB workflow in Siril, and I would like to ask whether my current approach makes sense, or whether I am overcomplicating it and introducing unnecessary sources of error.
At the moment, my workflow is roughly this:
I start from the raw stacked masters for L, R, G, and B.
I use RGB Composition on the raw L, R, G, B masters, with alignment enabled, so that all channels are registered consistently.
I then take the aligned luminance master and process it separately. This usually includes background extraction, denoising, aberration removal, sharpening, and then stretching.
Separately, I compose an RGB image from the (aligned before) R, G, and B masters only, without luminance.
On this RGB image, I perform plate solving and then (S)PCC.
After that, I process the RGB image, again with operations such as background extraction, denoising, aberration removal, mild sharpening, and stretching, usually less aggressively than for luminance.
I then split the stretched RGB image back into its R, G, and B channels.
Finally, I recombine the processed/stretched luminance and the processed/stretched R, G, B channels into a final LRGB image.
So, in short, I process luminance and RGB separately, and only at the end I recombine them.
My doubts are these:
Is this workflow conceptually correct?
Is splitting the already processed RGB image back into channels and recomposing with L a reasonable approach, or is it a sign that I am doing something in a clumsy way?
Is there a better / simpler / more robust way to build an LRGB image in Siril?
Could this workflow itself be one of the reasons why my final LRGB results often look somewhat washed out compared to the standalone RGB?
What worries me is that this workflow feels quite heavy and error-prone. There are many steps, many intermediate files, and many opportunities to make inconsistent choices between the luminance branch and the RGB branch.
Also, my final L+RGB recombinations are often less vivid than I expect: the result is usually smoother and more detailed, but also somewhat desaturated or “flat”. I am not sure whether this comes from my processing choices, from stretching L and RGB independently, or from the recombination method itself.
I would be very interested to know whether this is a sound workflow, or whether more experienced users would recommend a different strategy.
Thanks a lot.
