README.md

metadata

tags:
  - speech
  - text
  - alignment
language: en
license: cc-by-4.0

English Accent Audio-transcript Dataset.

Audio is extracted from movies, shows, speeches, talks to capture diverse accents - mainly scottish and indian

This dataset contains 30-second audio clips with aligned transcript text.

Structure

Each entry includes:

• audio: 30-second speech segment
• text: Corresponding transcript
• start_time / end_time: Segment timestamps

License

Licensed under CC-BY-4.0.