Alternative implementation of ALF memory reduction

This alternative implementation uses only 83% of peak memory compared to master branch implementation. The change is lossless bd-rate wise

Merge request reports

Loading