Investigate: speed-of-sound propagation delay for in-world sounds and speakers #896

New issue

Open

opened 2026-10-20 21:08:07 +00:00 by sky_line · 0 comments

sky_line commented

2026-10-20 21:08:07 +00:00

Owner

Investigate whether it's worth simulating sound arrival delay (distance / 343 m/s) for in-world audio.

Today

Already delayed by distance: thunder (#783). Room-acoustics reverb pre-delay and late-reverb delay, and the reverb distance decay, are also derived from 343 m/s.
Played instantly, with only level and filtering changing: in-world sounds and parcel speakers (#892). Paths around corners and through doorways (#858) change level and tone but add no delay.
Doppler: OpenAL doppler is configured, but check that sources actually set velocity.

Questions

Speakers. The feeder already schedules sample-exact starts, so a per-speaker delay looks cheap. At a large stage, far speakers and delay towers would line up as they do in real life. Does it improve the experience, or just add latency?
Script-timed sounds. Delaying them shifts script-synchronised audio, such as music players and animation sync. Should the delay be limited to distances above a threshold, or to particular sound kinds?
Diffraction paths. Could #858 paths carry their extra path-length delay, a small precedence effect toward the opening?
Doppler. Make moving sources (vehicles, projectiles) actually shift pitch.
Cost and complexity. Decide which layer would own the delay: per-source start offsets, or a delay line per speaker role.

Deliverable

A recommendation: worth it or not, and for which sounds.

Investigate whether it's worth simulating sound arrival delay (distance / 343 m/s) for in-world audio. ## Today - **Already delayed by distance:** thunder (#783). Room-acoustics reverb pre-delay and late-reverb delay, and the reverb distance decay, are also derived from 343 m/s. - **Played instantly, with only level and filtering changing:** in-world sounds and parcel speakers (#892). Paths around corners and through doorways (#858) change level and tone but add no delay. - **Doppler:** OpenAL doppler is configured, but check that sources actually set velocity. ## Questions 1. **Speakers.** The feeder already schedules sample-exact starts, so a per-speaker delay looks cheap. At a large stage, far speakers and delay towers would line up as they do in real life. Does it improve the experience, or just add latency? 2. **Script-timed sounds.** Delaying them shifts script-synchronised audio, such as music players and animation sync. Should the delay be limited to distances above a threshold, or to particular sound kinds? 3. **Diffraction paths.** Could #858 paths carry their extra path-length delay, a small precedence effect toward the opening? 4. **Doppler.** Make moving sources (vehicles, projectiles) actually shift pitch. 5. **Cost and complexity.** Decide which layer would own the delay: per-source start offsets, or a delay line per speaker role. ## Deliverable A recommendation: worth it or not, and for which sounds.

sky_line added the

feature

label

2026-10-20 21:08:07 +00:00

sky_line added this to the Slipstream project

2026-10-20 21:08:08 +00:00

sky_line referenced this issue

2026-10-20 22:11:28 +00:00

Parcel speakers: smooth sync during movement (no drift-then-snap) #898