I’ve been working on a small C library of DSP functions called miniDSP. It was born out of a need for simple, reusable lightweight components for audio processing tasks, without the overhead of massive frameworks. The library has grown quite a bit since I first wrote about it — here’s an updated overview of what’s in there now.

What is miniDSP?

miniDSP is a C23 library of essential digital signal processing tools, easy to integrate into any C project. It handles everything from spectral analysis and signal generation to DTMF decoding and live audio I/O. Full API reference and tutorials are at wooters.github.io/miniDSP.

What’s in the box?

The library is organized into a few logical areas. Each section links to its tutorial and API docs on the documentation site.

Frequency-domain analysis

The spectral analysis functions cover the fundamentals of frequency-domain work:

• Magnitude spectrum — |X(k)| from a real signal using FFT
• Power spectral density — periodogram showing how signal power distributes across frequencies
• Phase spectrum — arg(X(k)) in radians; prerequisite for phase-vocoder effects
• Spectrogram (STFT) — sliding-window FFT producing a time-frequency magnitude matrix
• Mel filterbank & MFCCs — perceptually-scaled features for speech and audio ML
• Window functions — Hanning, Hamming, Blackman, and rectangular windows

Signal generation

Generators for building test signals and demos: sine wave, white noise (Gaussian), impulse, chirp (linear and logarithmic), square wave, and sawtooth.

Filtering

Two filter families are included:

• Biquad filters (biquad.h) — seven classic types based on Robert Bristow-Johnson’s Audio EQ Cookbook: low-pass, high-pass, band-pass, notch, peaking EQ, low shelf, and high shelf.
• FIR filters & convolution — time-domain convolution, moving-average, general FIR, and FFT overlap-add for longer kernels.

Signal analysis & measurement

Functions for understanding what a signal is doing:

• RMS, zero-crossing rate, autocorrelation, peak detection, and signal mixing
• F0 estimation via autocorrelation and FFT peak-picking
• Energy, power (dB), normalized entropy, scaling, and automatic gain control

Effects

A handful of simple audio effects implemented as self-contained C functions: delay/echo, tremolo, and comb-filter reverb.

DTMF tone detection and generation

DTMF (Dual-Tone Multi-Frequency) is the signalling system used by touch-tone telephones. miniDSP includes:

• Multi-digit tone synthesis — generate DTMF sequences with configurable tone and pause durations
• Sliding-window FFT detection — decode digits from audio using a frame-based detector with an ITU-T Q.24-compliant state machine that enforces minimum 40 ms tone-on and 40 ms inter-digit pause timing
• Decoded digit output — each detected tone includes the digit character and onset/offset timestamps in seconds

Delay estimation

GCC-PHAT — estimate the time delay between two microphone signals using Generalized Cross-Correlation with Phase Transform. The core of acoustic source localization.

I/O

For reading, writing, and capturing audio:

• File I/O (fileio.h) — read any format libsndfile supports (WAV, FLAC, AIFF, OGG, etc.); write WAV, NumPy .npy, or safetensors.
• Live audio (liveio.h) — record from the microphone and play back to speakers via PortAudio, with a non-blocking callback API.

Example

Generate a 440 Hz sine wave and find its peak frequency bin:

#include "minidsp.h"
#include <stdio.h>

int main(void) {
    double signal[1024];
    MD_sine_wave(signal, 1024, 1.0, 440.0, 16000.0);

    double mag[1024 / 2 + 1];
    MD_magnitude_spectrum(signal, 1024, mag);

    unsigned peak = 0;
    for (unsigned k = 1; k < 1024 / 2 + 1; k++)
        if (mag[k] > mag[peak]) peak = k;
    printf("Peak bin: %u (%.1f Hz)\n", peak, peak * 16000.0 / 1024);

    MD_shutdown();
}

There are 16 runnable examples in the examples/ directory, many of which produce interactive HTML plots (Plotly.js + D3.js) for exploring the output visually.

Documentation & Tutorials

The documentation site includes a Doxygen API reference and step-by-step tutorials covering:

• Signal generators — sine waves, white noise, impulse, chirp, and more
• Window functions — Hanning, Hamming, Blackman, and rectangular
• Magnitude spectrum — decomposing signals into frequency components via the DFT
• Power spectral density — measuring how power distributes across frequencies
• Spectrogram (STFT) — visualising frequency content over time
• Mel filterbank & MFCCs — perceptual features for speech and audio ML
• Phase spectrum — per-frequency phase angles
• Basic signal operations — RMS, zero-crossing rate, autocorrelation, peak detection, and signal mixing
• Simple effects — delay/echo, tremolo, and comb-filter reverb
• FIR convolution — time-domain and FFT overlap-add convolution
• Pitch detection — F0 estimation with autocorrelation and FFT peak picking
• DTMF — tone detection and generation with ITU-T Q.24 timing

Get It

Clone and build:

git clone https://github.com/wooters/miniDSP.git
cd miniDSP
make

This produces libminidsp.a in the repo root. Link against it with -lminidsp -lfftw3 -lm (add -lsndfile for fileio.h, -lportaudio for liveio.h).

Dependencies

Library	Purpose	Debian/Ubuntu	macOS (Homebrew)
FFTW3	Fast Fourier Transform	apt install libfftw3-dev	brew install fftw
PortAudio	Live audio I/O	apt install portaudio19-dev	brew install portaudio
libsndfile	Audio file reading/writing	apt install libsndfile1-dev	brew install libsndfile

The Makefiles auto-detect Homebrew paths on macOS (both Apple Silicon and Intel). On Ubuntu, GCC 14 or later is required for C23 support.

Running tests

make test             # build and run all tests
make container-test   # run tests inside an Ubuntu 24.04 container (macOS only)
make docs             # generate API docs in docs/html

Full source, build instructions, and examples on GitHub: https://github.com/wooters/miniDSP