WebVTT generation functionality
Opened this issue · 0 comments
agrathwohl commented
Description
A Hypercore output from StreamAnalyserCore
will contain time-aligned metadata about an audio recording session, including any transcription text generated in real time (see: AUD-39).
This information is sufficient to produce a WebVTT-formatted string. We should produce a new function that can convert a Hypercore output by StreamAnalyserCore
into a WebVTT file. This will enable time-synchronized audio playback with text annotations.
Acceptance Criteria
- The ideal functionality would be provided within our new audio corestore, requiring both the main
AudioBuffer
feed and an associatedStreamAnalyserCore
feed to be in the store before the function executes. - Function should write clipping, silence, and voice activity to the WebVTT output as WebVTT comments.
- Function should write each transcription text object contained in the
StreamAnalyserCore
as a WebVTT Cue with a cue text payload.
Additional Notes
This function should be exported independently of the class(es) making use of it.