Well, this is a Python script which also uses FFMPEG. We translate using Google Translator. We also use Google's speech recognition. Currently works only on Windows. For now it sucks, since it doesn't do any threading.
Tested on Python 3.6.8. I don't know if it works on other versions of Python. Package versions below are the confirmed ones on which I tested the software.
- pafy 0.5.5
- requests 2.25.0
- json 2.0.9
- SpeechRecognition 3.8.1
- googletrans 4.0.0-rc.1
- cutlet 0.1.17
You will also need ffmpeg, which you can download from here, or their website.
- Download the script
- Install the modules
- Open it in the cmd/powershell/git (any terminal, remember that it has to be on Windows)
- Copy and paste the Youtube Channel ID of the person who is streaming.
- Enter the input language (use a language code, for example "ja", which stands for Japanese)
- Enter the destination language (also use a language code, for example "en", which stands for English) After these steps, the script should start trying to detect the speech and to translate it. If you get an python exception saying invalid source language, then the language code you've provided was incorrect.
The script is still in pre-alpha version. In the future, I plan on doing things in threads, stop the duplicate sentences,words problem, and many more. I also may make a gui for this script in the future.