Voice-Dataset-Collection
Voice-Dataset-Collection copied to clipboard
fix(frontend recording): fixing audio capture and lossy encoding
Description
This change addresses issue #4. Resolves the current bug of naming blobs with wav extensions but leaving the MediaRecorder for the browser to manage the actual audio encoding of the incoming stream.
Changes
- Removes the misnamed wav encoding statements.
- Adds raw stream capture in PCM using audio contexts and buffers
Recording.tsx - Adds Wav encoding utility to convert PCM data to wav
frontend/src/utils.wavEncoder.ts