Adjustable VAD threshold: Speech to Text now documents a vad_threshold parameter (streaming query param and batch multipart field) to tune the voice-activity gate that skips non-speech audio. Lower values help with quieter or noisier spe…
Changelog
Speech to Text / Voice
Adjustable VAD threshold: Speech to Text now documents a vad_threshold parameter (streaming query param and batch multipart field) to tune the voice-activity gate that skips non-speech audio. Lower values help with quieter or noisier speech (e.g. narrowband telephony); 0 disables the gate.
Defaults and behavior are covered in the Speech to Text model docs and REST API reference (batch and streaming), including Voice streaming.
Docs
No pages added or removed.
Release notes updated with the VAD threshold entry; related API and capabilities pages refreshed to match.
