Source: pymicro-vad
Maintainer: Home Assistant Team <team+homeassistant@tracker.debian.org>
Uploaders:
 Edward Betts <edward@4angle.com>,
Section: python
Priority: optional
Build-Depends:
 debhelper-compat (= 13),
 dh-sequence-python3,
 pybuild-plugin-pyproject,
 python3-all,
 python3-setuptools,
Build-Depends-Arch:
 python3-pytest <!nocheck>,
Standards-Version: 4.7.2
Homepage: https://github.com/rhasspy/pymicro-vad
Vcs-Browser: https://salsa.debian.org/homeassistant-team/deps/pymicro-vad
Vcs-Git: https://salsa.debian.org/homeassistant-team/deps/pymicro-vad.git
Testsuite: autopkgtest-pkg-pybuild

Package: python3-pymicro-vad
Architecture: any
Depends:
 ${shlibs:Depends},
 ${misc:Depends},
 ${python3:Depends},
Description: Self-contained voice activity detector for audio streams
 This library provides real-time detection of speech within raw audio data
 streams, distinguishing between speech and background noise using compact
 machine learning models derived from microWakeWord. It is designed to process
 small audio frames, typically ten milliseconds in length, represented as
 16-bit mono PCM sampled at 16 kHz. Operating independently without external
 service requirements, the detection process enables systems to identify when
 speech is present in continuous audio. This makes it suitable for use in
 embedded devices, voice assistants, and other scenarios where efficient, local
 voice activity detection is necessary. The detection algorithm operates
 entirely on device and does not require network connectivity or interaction
 with cloud-based services, ensuring privacy and low latency. Input is expected
 to be short chunks of audio data, and the output is a probability indicating
 the presence or absence of speech for each processed frame.
