Which OpenAL distance model behavior applies to existing sources after a global change?

Answer

When alDistanceModel() is called, the new distance model is used immediately for attenuation calculations of **all** active sources, including those that are already playing. No additional source‑side call or buffer re‑queue is required; the update takes effect on the next audio mix buffer.

Confirmed facts

• The OpenAL specification treats AL_DISTANCE_MODEL as global state.
• Attenuation is recomputed each time the mixer needs a source’s output level; there is no per‑source cached distance model.
• Therefore a change from, e.g., AL_INVERSE_DISTANCE to AL_LINEAR_DISTANCE changes the volume of existing sources on the very next frame.

Likely explanation

OpenAL does not store a distance model per source; it reads the global setting during each mix step. When the global setting changes, the next mix step automatically uses the new model, so the attenuation curve updates synchronously with the audio pipeline.

Steps to verify

1. Create a source playing a steady tone (e.g., a looping buffer).
2. Call alDistanceModel(AL_LINEAR_DISTANCE) while the source is playing.
3. Listen for an immediate volume change; it should be audible as soon as the next mix buffer is processed (typically within a few milliseconds).
4. Optionally, measure the output RMS before and after the call; the difference will match the expected attenuation change.

If you need to know the exact latency (e.g., the size of the current mix buffer) to predict the delay, provide that buffer size; otherwise the behavior described above holds for all standard OpenAL implementations.