Stages applied, in source order: copy for mono at the target rate, downmix for stereo at the target rate, resample for mono at another rate, and downmix-resample for stereo at another rate.
Downmix coefficients with one row per output channel and one column per source channel: [[1]] for mono and [[0.5, 0.5]] for stereo.
Mixer-facing channel count; the decoder always folds to mono.
Positive safe-integer output samples per second, validated by the ingest decoder from the caller's request.
The deterministic conversion
decodeProductionAudioAssetapplied to turn a WAVE source into mono mixer input, recorded so a render plan can check it against the source format it claims.