Skip to content

Experimental Spatial Audio refinement: center preservation and unified peak safety - #5

Merged
Batmancpu merged 6 commits into
mainfrom
experiment/spatial-audio-refinement-20261115
Nov 15, 2026
Merged

Batmancpu merged 6 commits into
mainfrom
experiment/spatial-audio-refinement-20261115

Conversation

@Batmancpu

Copy link
Copy Markdown
Owner

Scope

  • Reversible second layer over the existing stable + experimental Spatial Audio engine
  • Preserve center and concentrate extra widening above 140 Hz; avoid common-mode crossfeed coloration
  • Delegate normal Float32 DSP-chain peak limiting to its final loudness stage, keeping direct/PCM16 safety
  • Add 7 audio regression tests, ensure both CI pipelines apply and restore the layer, preserve branding/signature and on-demand-only builds
  • New release flag audioRefinement forces version bump when enabled/disabled

Hold conditions

Do not merge if production unit tests fail. CI tests do not replace blind listening or battery benchmarks; improvements are candidates, not proven perceptual or energy benefits.

Rollback

Set fc/spatial-audio/refinement/config.properties to enabled=false, then build a higher-version signed APK. Prior experiment and original engine remain available.

@Batmancpu
Batmancpu merged commit 1f6ec55 into main Nov 15, 2026
1 check passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant