daveboling / applyaudiosr Goto Github PK

A simple CLI script that automates the processes necessary to apply audio super resolution to a full-length WAV file

NOTE: Currently, this project only supports mono WAV files. Stereo support will be added in the future. Please keep this in mind if you plan to use this project for any purpose.

Steps to Run:

# Create a new conda environment
conda create -n applyaudiosr python=3.9
conda activate applyaudiosr

# Install poetry
pip3 install poetry

# Install packages
poetry install

# Run
poetry run applyaudiosr --waveform-path lofi_sample.wav --guidance-scale 0.5 --seed 12345 --ddim-steps 25 --model-name basic

After the process is complete, the outputs will be stored in a folder with the same name as the audio file. Please note that these outputs will be located in the directory where the command is executed.

Test:

poetry run python -m unittest discover tests

Tested on

• macOS (M1)

AudioSR processing was made possible by https://github.com/haoheliu/versatile_audio_super_resolution 🙇