mcut SDK

mcut SDK / transcription-local/src / WHISPER_SAMPLE_RATE

Variable: WHISPER_SAMPLE_RATE

const WHISPER_SAMPLE_RATE: 16000 = 16_000;

Defined in: packages/transcription-local/src/wav.ts:8

Minimal WAV reader → 16kHz mono Float32Array (Whisper's input format). The editor extracts clip audio as 16kHz WAV already (extractAudioToWav), so this covers the hot path without any AudioContext — it runs in workers and tests alike. Non-WAV input falls back to decodeAudioData upstream.