pstruebi c681e4ce39 feat: refactor audio input to use dedicated reader thread instead of per-frame executor ...

- Replaced per-frame `run_in_executor` calls with single background reader thread in `ThreadedAudioInput`
- Reader thread continuously calls `_read()` and enqueues data via `call_soon_threadsafe` to asyncio.Queue
- Reduces per-frame scheduling overhead and context-switch jitter while preserving async API
- Added thread lifecycle management: lazy start on first `frames()` call, graceful stop in `aclose()`
- Update

2025-11-19 18:52:37 +01:00

..

__init__.py

first commit

2025-11-17 13:33:17 +01:00

io.py

feat: refactor audio input to use dedicated reader thread instead of per-frame executor

2025-11-19 18:52:37 +01:00