[ SUMMARY ]
・Please expose the local audio output latency (time from audio scheduled by the engine to sound at the user’s speakers/headphones) to Udon.
・With this value, world creators can apply a per-client offset to video playback so audio and video stay in sync on each device.
[ PROBLEM ]
・Audio output latency varies widely by user (hardware, OS/driver, platform, buffer sizes).
・Udon currently cannot read that latency, so creators rely on manual sliders or fragile periodic seek/pause nudges.
・Result: noticeable desync during watch parties, karaoke, concerts, rhythm content, and story scenes.
[ PROPOSAL (Udon API surface; read-only, local) ]
・float GetLocalAudioOutputLatencySeconds()
Returns the estimated end-to-end local output latency in seconds (local only).
・(optional) double GetLocalDSPTime()
Exposes local audio DSP time for precise scheduling.
・(optional) bool TryGetLocalAudioOutputLatency(out float seconds)
Safe getter with failure fallback.
・(optional) Low-frequency refresh or change notification (approx. 1–2 Hz) if latency can vary at runtime.
[ ADDITIONALLY HELPFUL (not required) ]
・int GetOutputSampleRate()
・void GetDSPBuffer(out int bufferSizeSamples, out int numBuffers)
[ HOW CREATORS WOULD USE IT ]
  1. Read GetLocalAudioOutputLatencySeconds() on each client.
  2. When starting or resyncing video, apply that value as a local offset (small seek or short pause) so visual frames align to what the user actually hears.
  3. Keep existing network sync as is; this request enables only a local correction.
[ KEY USE CASES ]
・Karaoke / lip-sync shows: align mouth movements and lyrics to the user’s real heard audio.
・Watch parties / cinemas: reduce clap-test drift between audio and video.
・Rhythm / DJ / club worlds: visuals land on beat for each attendee.
・Cutscenes / story events: tighter A/V cohesion improves immersion.
[ EXPECTED BEHAVIOR AND CONSTRAINTS ]
・Read-only numeric data; no privacy issues.
・Local-only value; creators implement their own offset logic.
・Reasonable update cadence (on change or around 1–2 Hz) is sufficient.
・If unsupported on a platform, return 0 or a defined “unsupported” state.
[ WHY THIS MATTERS ]
・A single float (local audio output latency) lets creators deliver professional-grade A/V sync.
・It removes device-dependent guesswork and yields a consistent viewing/listening experience with a minimal, safe API.
[ ACCEPTANCE CRITERIA ]
・Udon can query a local audio output latency value (seconds).
・On typical hardware, creators can keep A/V sync within approximately ±20–40 ms using this value.
・Clear, documented fallback behavior when the value is unavailable.