Crate audio_lab 

audio-lab

audio-lab is a Rust library for offline audio research and algorithm testing. It provides a compact set of tools for generating signals, analyzing and manipulating them, plotting results, playing them back, and saving them to audio files or NumPy arrays. It is designed for experiments, validation, and measurement workflows, not for real-time processing.

Features

• Generate test signals such as sweeps, sine waves, impulses, noise, and pulsed noise
• Work with time-domain signals, frequency-domain signals, and spectrograms
• Manipulate signals with trimming, gain, shifting, resampling, convolution, and deconvolution
• Plot waveforms, spectra, and spectrograms in native windows
• Play signals back for quick listening checks
• Save and load audio files and NumPy arrays

Installation

Add audio-lab to your Cargo.toml:

[dependencies]
audio-lab = "0.1.0"

Usage

Generating a test signal

use audio_lab::signal;

fn main() -> Result<(), Box<dyn std::error::Error>> {
    let sweep = signal::generate_sweep(
        48_000,
        20.0..20_000.0,
        signal::SweepConfig::default(),
    )?;

    let noise = signal::generate_noise(
        48_000,
        signal::NoiseConfig {
            spectrum: signal::Spectrum::Pink,
            amplitude: 0.2,
            ..Default::default()
        },
    )?;

    let test_signal = audio_lab::signal::mix_time!(sweep, noise)?;
    Ok(())
}

Plotting

use audio_lab::{plot, signal};

fn main() -> Result<(), Box<dyn std::error::Error>> {
    let signal = signal::generate_sine(48_000, 440.0, signal::SineConfig::default())?;

    plot::show_time("time", &[&signal])?;
    plot::show_freq("spectrum", &[&signal.into_freq()])?;
    Ok(())
}

Playback

use audio_lab::signal::{self, playback};

fn main() -> Result<(), Box<dyn std::error::Error>> {
    let signal = signal::generate_sine(48_000, 440.0, signal::SineConfig::default())?;
    playback::play(&signal)?;
    Ok(())
}

Saving to audio or NumPy

use audio_lab::signal::{
    self,
    fs::{audio, npy},
};

fn main() -> Result<(), Box<dyn std::error::Error>> {
    let signal = signal::generate_sweep(
        48_000,
        20.0..20_000.0,
        signal::SweepConfig::default(),
    )?;

    audio::write_audio(&signal, "test.wav", audio::SampleFormat::Float32)?;
    npy::write_npy_time(&signal, "test.npy")?;
    Ok(())
}

Computing a spectrogram

use audio_lab::{dsp, plot, signal};

fn main() -> Result<(), Box<dyn std::error::Error>> {
    let signal = signal::generate_sweep(
        48_000,
        20.0..20_000.0,
        signal::SweepConfig::default(),
    )?;

    let spec = dsp::stft(&signal, &dsp::StftConfig::default())?;
    plot::show_spectrogram("spectrogram", &spec, Default::default())?;
    Ok(())
}

Estimating an impulse response

use audio_lab::{dsp, signal};

fn main() -> Result<(), Box<dyn std::error::Error>> {
    let excitation = signal::generate_sweep(
        48_000,
        20.0..20_000.0,
        signal::SweepConfig::default(),
    )?;

    let recorded_output = excitation.clone();

    let impulse_response = dsp::deconvolve(
        &recorded_output,
        &excitation,
        &dsp::DeconvolutionConfig::default(),
    )?;

    println!("{}", impulse_response);
    Ok(())
}

Modules

• signal: signal types, generators, I/O, playback
• dsp: DSP and measurement primitives
• plot: plotting and visualization

Cargo features

All of these are enabled by default, except serde:

• plot: the plot module. Pulls in egui, so turn it off if you don’t need windows.
• playback: signal::playback, plays signals on the default audio device.
• audio-file: signal::fs::audio, reads and writes audio files.
• numpy: signal::fs::npy, reads and writes NumPy .npy files.
• serde: Serialize and Deserialize for the signal types.

Modules

dsp

Generic DSP and measurement primitives built on top of signal.

plot

Plot waveforms, spectra and spectrograms in native windows.

signal

Signal types, test signal generators, file I/O and playback.

Macros

join_freq

join_time

mix_freq

mix_time