uploaded 1 day ago
Lightweight PCM audio processing for voice-assistant style pipelines: format conversion, level metering, resampling, and voice activity detection (RMS-based and WebRTC-based), both as one-shot and streaming APIs.
Changelog
# Changelog
<!--
Placeholder for the next version (at the beginning of the line):
## **WORK IN PROGRESS**
-->
## 0.2.1
- Add public mirror. No functional changes
## 0.2.0
- Add WebRTC VAD backend (`hannah_webrtc_vad_init` / `hannah_webrtc_vad_feed` / `hannah_webrtc_vad_free`), vendored from [libfvad](https://github.com/dpirch/libfvad) under `third_party/libfvad/` (BSD license). Frequency-based, distinguishes speech from music/background noise where the RMS-based VAD can't; runs independently alongside it on a separate state instance.
## 0.1.1
- `hannah_resample`: apply a second-order Butterworth low-pass before downsampling to suppress aliasing (frequencies above the destination Nyquist previously folded back into the output)
- `hannah_resample`: reject negative `in_samples`/`out_samples` (previously only `== 0` was caught)
## 0.1.0
- Initial release of AudioLib
- Batch API: stereo-to-mono downmix, RMS energy, linear-interpolation resampling, voice activity detection (VAD)
- Streaming VAD state machine with debounced onset/offset detection for chunk-by-chunk processing (e.g. I2S DMA buffers)
- ESP-IDF component support via `CMakeLists.txt`
- Unit test suite (`make test`) and GitLab CI pipeline for merge requests
Links
Supports all targets
To add this component to your project, run:
download archive
idf.py add-dependency "nurpech/audiolib^0.2.1"
Stats
-
Archive size ~ 52.52 KB
-
Downloaded in total 0 times
-
This version: 0 times
Badge
