Adds full frame implementation to this node operations.
No functional changes.
No performance changes.
Details
Details
- Reviewers
Jeroen Bakker (jbakker) - Commits
- rBe33814ef6bae: Compositor: Full frame Levels node
Diff Detail
Diff Detail
- Repository
- rB Blender
- Branch
- cmp-node-levels (branched from master)
- Build Status
Buildable 15911 Build 15911: arc lint + arc unit
Event Timeline
| source/blender/compositor/intern/COM_MultiThreadedOperation.cc | ||
|---|---|---|
| 18 ↗ | (On Diff #38929) | Better have a setter for clarity and remove the parameter from functions. |
| source/blender/compositor/operations/COM_CalculateMeanOperation.cc | ||
| 198 | We could add a join code path to remove the mutex. (every thread works on their own instance of PixelsSum and at the end it would join them together. | |
Comment Actions
- Merge branch 'master' into cmp-node-levels
- After merge fix
- Add execute_work overload with join function and use it to sum results
| source/blender/compositor/operations/COM_CalculateStandardDeviationOperation.cc | ||
|---|---|---|
| 111 | Normally a reduce function has 2 parameters of the same type. Here (two PixelsSum) one is const, the other one is writable. | |
Comment Actions
- Implement reduce function as review said
- Merge branch 'master' into cmp-node-levels
- Use iterators
- Avoid zero division