Restoring timecodes to speaker diarized accurate transcription.
## Steps to prep
### Option 1
- Download videos
- Run videos through STT
- using https://github.com/bbc/digital-paper-edit-electron) for UI
- STT credentials from https://app.assemblyai.com using 5hours monthly free tier
- correct speaker segmentstion (https://github.com/bbc/react-transcript-editor)
- export in DPE json format to get corrected speaker segmentation
### Option 2
- install pocketsphinx locally
OpenNewsLabs/pocketsphinx-stt - run alignement locally with @bbc/stt-align-node
- import in @bbc/react-transcript-editor demo
- add speaker labels and review speakers segmentation
- export dpe json
### option 3