/piv-audio-convolution

A Python script to convolve a given audio signal with a given noise signal at a requested signal-to-noise ratio (SNR).

Primary LanguagePython

This is a quick script to convolve a given audio signal with a given noise signal at a requested signal-to-noise ratio (SNR).

Usage

Note

This project has only been tested on Python version 3.12. Your milage may vary when using other (especially lower) Python versions.

  1. Install Dependencies

    cd ./piv-audio-convolution
pip install -r requirements.txt

  2. Run the command utility

    python ./main.py <IMPULSE_RESPONSE_MAT_FILEPATH> <INPUT_SIGNAL_FILEPATH> [--snr <SNR>] [--noise <NOISE_FILEPATH>] [--output <OUTPUT_FILEPATH>]

    impulse_response_mat is the .mat file generated by the ScanIR program encoding the room impulse response.

    input_signal is the 'pure' input speech signal to convolve with the impulse response.

    --snr option allows you to specify the signal-to-noise ratio. Defaults to ∞ db (no noise).

    --noise option allows you to specify the noise file used. Defaults to ./resources/babble.wav

    --output option allows you to specify the output filepath.