Y-DAT: Youtube Video Download and Transcription

Allows the user to download Youtube video using it's URL (using youtube_dl library) and then generate it's transcript using SpeechRecognizer library

Installation

Make sure you have Python 3.5 or higher installed along with pip.

Cloning the repository

git clone https://github.com/thedhruvrawat/y-dat.git

Installing the requirements

To install the requirements, run

pip install -r requirements.txt

Downloading the video

To download the video from a link, run

python downloader.py -url <YOUTUBE_VIDEO_URL>

Generating the transcript

To recognize the speech in video, run

python3 recognizer.py -video <VIDEO_NAME>

By default, the video will be saved by the name video.mkv

The generated text will be saved in the file transcript.txt, along with this an audio.wav file containing solely the audio of the downloaded video will also be generated in the same folder. Works only for English language.

Best results obtained when the audio is free from any kind of background sounds

Further Reading

• Youtube-dl
• SpeechRecognizer
• Librosa
• Russian-Subtitles-Generator

License

MIT