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:

Download a playlist from YouTube at once.
Transcribe speech into English with timestamps (Voice-to-Text).
Translate the text into Swedish.
Use Speech Synthesize to generate Swedish voice (Text-to-Voice).
Mix together the generated audio clips and synchronize the audio to the video.
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