Adaptive Color Transform in Video Encoding

The patent applies ACT selectively to specific coding units based on local content characteristics rather than uniformly across all coding units. The encoder evaluates each coding unit to determine if ACT is beneficial, applying the transform only where color component decorrelation provides meaningful quality improvement. This local application approach maintains high encoding quality where needed while avoiding unnecessary complexity in regions where ACT would not provide benefit.

The patent dynamically adjusts the application of ACT based on content analysis. The encoder performs rate distortion optimization checks to determine the optimal configuration for each coding unit, enabling or disabling ACT adaptively rather than using a static configuration. This dynamic approach allows the encoding system to respond to varying content characteristics, applying ACT only when it provides measurable quality improvement.

The patent applies ACT partially rather than universally. By performing rate distortion optimization checks and selectively applying ACT only to coding units where it provides benefit, the encoder avoids the excessive action of applying ACT to all coding units regardless of need. This partial application reduces processing time by skipping unnecessary evaluations and transforms for coding units where ACT would not provide meaningful improvement.

The patent enables the encoder to skip unnecessary ACT evaluations and applications for coding units where color components are already de-correlated or where ACT would not provide benefit. By identifying such units through content analysis and rate distortion optimization checks, the encoder rushes through these regions without applying ACT, thereby reducing overall processing time while maintaining encoding efficiency in regions where ACT is beneficial.