avoc 0.0.5

Voice changer for desktop

AVoc: Local Realtime Voice Changer for Desktop

A speech-to-speech converter that uses AI models locally to convert microphone audio to a different voice in near-realtime.

Suitable for gaming and streaming.

Quick Start

Drag your voice model files into the window.

Features

• Import of the voice models provided by the user
• Switching between voices
• Pitch adjustments
• Hotkeys and popup notifications for the ease of use in the background
• Pass Through

Platforms

All desktops.

Linux is the priority.

Goal

Make voice changing more developer-friendly by creating

• a voice conversion library
• a simple voice changer desktop application
• a command-line voice changer program

Open Source and Free for modification.

Installation

For Arch-based Linux Distributions - from AUR

No cloning of this repo needed.

yay -S avoc

or for Manjaro

pamac build avoc

Launch from the menu or by running:

gio launch /usr/share/applications/AVoc.desktop

For other Linuxes

Requires pyenv, update-desktop-database and some build tools to be installed.

After that, install the voice changer into a local directory:

git clone https://github.com/develOseven/avoc
cd avoc
pyenv local 3.12.3
python -m venv .venv
source .venv/bin/activate
CMAKE_ARGS="-DCMAKE_POLICY_VERSION_MINIMUM=3.5" pip install . -r requirements-3.12.3.txt
mkdir -p ~/.local/share/applications/
cp -t ~/.local/share/applications/ src/avoc/AVoc.desktop
echo "Path=$PWD" >> ~/.local/share/applications/AVoc.desktop
mkdir -p ~/.local/share/icons/hicolor/scalable/apps/
cp -t ~/.local/share/icons/hicolor/scalable/apps/ src/avoc/AVoc.svg
update-desktop-database

Launch from the menu or by running:

gio launch ~/.local/share/applications/AVoc.desktop

To uninstall:

rm ~/.local/share/applications/AVoc.desktop ~/.local/share/icons/hicolor/scalable/apps/AVoc.svg

(Optional) Virtual Microphone

To make a game take audio from the voice changer, the operating system needs to be configured to create a virtual microphone.

Linux with PulseAudio

Add this to ~/.config/pulse/default.pa:

load-module module-null-sink sink_name=voice-sink sink_properties=device.description=Voice_Sink
load-module module-remap-source master=voice-sink.monitor source_name=voice-mic source_properties=device.description=Voice_Microphone

And re-login or restart PulseAudio with pulseaudio -k

The Voice_Sink and Voice_Microphone devices will appear. Use the Voice_Sink as voice changer output, and use the Voice_Microphone as input for the game.

Development

Python Environment

Assign a compatible Python version to this directory using pyenv:

pyenv local 3.12.3

Create an environment using venv:

python -m venv .venv

or through VSCode with ~/.pyenv/shims/python as the Python interpreter.

Install the dependencies:

source .venv/bin/activate
pip install -r requirements-3.12.3.txt

Run:

python -m main

(Optional) Get sources of the voice conversion library and install it in developer mode:

(cd .. && git clone https://github.com/develOseven/voiceconversion)
source .venv/bin/activate
pip uninstall voiceconversion
pip install -e ../voiceconversion --config-settings editable_mode=strict

It allows to work on the voice conversion library.

(Optional) Add to the "configurations" in the VSCode's launch.json:

{
    "name": "Python Debugger: Module",
    "type": "debugpy",
    "request": "launch",
    "module": "main",
}