/openhack

Automate dubbing of Khan Academy videos

Primary LanguagePythonMIT LicenseMIT

Found in Translation - OpenHack 2017 in Malmö

Description

This project aims to automate the dubbing of Khan Academy videos into Swedish. The process to do this involves:

  1. Download a playlist from YouTube at once.
  2. Transcribe speech into English with timestamps (Voice-to-Text).
  3. Translate the text into Swedish.
  4. Use Speech Synthesize to generate Swedish voice (Text-to-Voice).
  5. Mix together the generated audio clips and synchronize the audio to the video.
  6. Upload all the updated videos in the playlist at once.

Future improvements

  • Better translations as well as Speech synthesize using i.e. Microsoft Cognitive Services.
  • More resilient automation.

Members

  • Albin Heimerson
  • Cajsa-Lisa Ivarsson
  • David Wessman
  • Hampus Månefjord
  • Josefin Lindström
  • Louise Linné
  • Miriam Ahlberg
  • Ross Linscott

Dependencies

  • Python 3 - for connecting to Microsof Cognitive Services and YouTube.
  • PHP - for scraping YouTube for transcripts and translations.
  • FFMPEG - for editing videos using MoviePy.

Run instructions

  • Run master.py with the playlist link as command line argument