search

Hiroshiba / realtime-yukarin

An application for real-time voice conversion
MIT License
331 stars 51 forks source link

Can u introduce your thought of real-time? #4

Closed CarolinGao closed 4 years ago

CarolinGao commented 4 years ago

Hi, I have a voice conversion model, I want to convert it to a realtime model.

  1. I am confused what do "start_time" and "extra_time" in your code mean.
  2. I want to record audio from microphone , process the audio data and play the processed audio at the meanwhile. How can I design the code? Thank u very much!
Hiroshiba commented 4 years ago

Sorry for the delay in replying!

1) In this voice conversion system, the voice is separated and converted sequentially. start_time is the start time of the separated voice. extra_time is the padding time of the voice. Increasing this value will increase the quality of the conversion, but will also increase the delay.

2) With PyAudio, you can convert your voice in real time. See run.py for more information.