-
Notifications
You must be signed in to change notification settings - Fork 1.2k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Highlights raster mask #16575
Highlights raster mask #16575
Conversation
dadfd4a
to
5daf6a1
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Tested and working on my side, a minor improvement proposed. TIA.
Make highlight "reconstruction" data available as a raster mask to be used at other parts of the pipeline. The mask used in highlight module - uniform or drawn - is modified by the process ratio of output and input data of that module. Works with all hightlights algorithms. Mask is gaussian-blurred for less noise.
5daf6a1
to
f94f976
Compare
fixes as requested, squashed and force-pushed. Comments: masking in pre-demosaic modules is broken for non-raws atm. Was also in master, will fix in another PR. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks!
Ah, didn't notice that. Is that since a long time? |
I simply don't know. I think we oversaw the colorspace default in one of the earlier modules in the pipe. Likely there was simply noone using masks in those modules on sraws. I just stumbled across this by chance. |
Reproducer: use an sraw file, define any mask in highlights, use -d pipe. Module csp is RAW for all modules early in the pipe. |
Adding support for the problem "What can we do with remaining color casts?".
Instead of
This PR offers a simple raster mask generation of the highlights module. It works for all internal algorithms that increase clipped photosite channels like opposed, segmentation based or laplacian. Respects the used mask in HLR module and modifies it.
To check: