art-from-the-machine / Mantella

Mantella is a Skyrim and Fallout 4 mod which allows you to naturally speak to NPCs using Whisper (speech-to-text), LLMs (text generation), and xVASynth / XTTS (text-to-speech).
https://art-from-the-machine.github.io/Mantella/
GNU General Public License v3.0
164 stars 41 forks source link

Add a warning when the mic recording filters out >10 seconds of audio #238

Open art-from-the-machine opened 4 months ago

art-from-the-machine commented 4 months ago

Sometimes players experience mic issues like the below when audio_threshold in config.ini is set to the default "auto" value:

02:09:58.517 Level 27: Listening... 02:11:42.227 INFO: Processing audio with duration 01:42.888 02:11:43.015 INFO: VAD filter removed 01:36.360 of audio

A warning needs to be displayed when VAD filter removes >10 seconds of audio. This is a problem where the player's mic picks up on too much background noise.