aubio-ledfx 0.4.11

A collection of tools for music analysis

aubio-ledfx

Note: This is a maintained fork of the original aubio project by Paul Brossier, maintained by the LedFx team.

Why this fork exists:

• The original aubio project is no longer actively maintained with regular releases
• We provide Python 3.10-3.14 support with pre-built wheels on PyPI
• This fork includes the latest fixes and improvements from aubio's main branch
• LedFx and other projects depend on aubio and need a reliable, up-to-date package

All credit for aubio goes to the original author Paul Brossier and contributors.

Original project: https://github.com/aubio/aubio
This fork: https://github.com/LedFx/aubio-ledfx
PyPI package: https://pypi.org/project/aubio-ledfx/

---

About aubio

aubio is a collection of tools for music and audio analysis.

This package integrates the aubio library with NumPy to provide a set of efficient tools to process and analyse audio signals, including:

• read audio from any media file, including videos and remote streams
• high quality phase vocoder, spectral filterbanks, and linear filters
• Mel-Frequency Cepstrum Coefficients and standard spectral descriptors
• detection of note attacks (onset)
• pitch tracking (fundamental frequency estimation)
• beat detection and tempo tracking

This fork supports Python 3.8 through 3.14 on Linux (x86_64, ARM64), macOS (Intel, Apple Silicon), and Windows (AMD64).

Installation

Install from PyPI:

pip install aubio-ledfx

Pre-built wheels are available for:

• Linux: x86_64, ARM64 (manylinux)
• macOS: Intel (x86_64), Apple Silicon (ARM64)
• Windows: AMD64

Links

• PyPI package
• module documentation
• installation instructions
• aubio manual
• original aubio homepage
• issue tracker (this fork)

Demos

Some examples are available in the python/demos folder. Each script is a command line program which accepts one ore more argument.

Notes: installing additional modules is required to run some of the demos.

Analysis

• demo_source.py uses aubio to read audio samples from media files
• demo_onset_plot.py detects attacks in a sound file and plots the results using matplotlib
• demo_pitch.py looks for fundamental frequency in a sound file and plots the results using matplotlib
• demo_spectrogram.py, demo_specdesc.py, demo_mfcc.py for spectral analysis.

Real-time

• demo_pyaudio.py and demo_tapthebeat.py use pyaudio
• demo_pysoundcard_play.py, demo_pysoundcard.py use PySoundCard
• demo_alsa.py uses pyalsaaudio

Others

• demo_timestretch.py can change the duration of an input file and write the new sound to disk,
• demo_wav2midi.py detects the notes in a file and uses mido to write the results into a MIDI file

Example

Use demo_timestretch_online.py to slow down loop.wav, write the results in stretched_loop.wav:

$ python demo_timestretch_online.py loop.wav stretched_loop.wav 0.92

Building from Source

This fork uses the Meson build system and vcpkg for dependency management.

Quick Build

# Install build dependencies
pip install "meson>=1.9.0" meson-python ninja numpy

# Build and install
pip install .

For detailed build instructions, see the main README.

Built with

The core of aubio is written in C for portability and speed. The pre-built wheels on PyPI include the following optional features:

All platforms:

• NumPy integration for efficient array processing
• libsndfile for reading/writing uncompressed audio (WAV, AIFF, etc.)
• fftw3 for fast Fourier transforms
• libsamplerate for high-quality audio resampling
• Audio codec support: FLAC, Vorbis/Ogg
• Built-in WAV reader/writer

Platform-specific features:

• macOS: Accelerate framework (optimized FFT), CoreAudio (native media reading), ffmpeg, [rubberband] (time-stretching)
• Windows: ffmpeg, [rubberband] (time-stretching) - all DLLs bundled in wheel
• Linux: MP3 support (mpg123, lame), Opus codec - static linking for portability

Not included: JACK audio, Intel IPP, BLAS/Atlas (for custom builds, see building from source)

For a detailed breakdown of features by platform, see the Pre-built Wheel Features appendix below.

---

Pre-built Wheel Features

This appendix provides a complete breakdown of which optional features are included in the pre-built wheels distributed on PyPI.

Features by Platform

Feature	Linux	macOS	Windows	Description
Audio File I/O
libsndfile	✅	✅	✅	Read/write uncompressed audio (WAV, AIFF, AU, etc.)
ffmpeg/libav	❌	✅	✅	Decode almost any media format (MP4, MKV, WebM, etc.)
CoreAudio	—	✅	—	Native macOS/iOS audio file reading (all Apple formats)
Built-in WAV	✅	✅	✅	Simple WAV support without external libraries
Audio Codecs
FLAC	✅	✅	✅	FLAC lossless audio codec
Vorbis/Ogg	✅	✅	✅	Ogg Vorbis lossy audio codec
MP3 (mpg123)	✅	❌	❌	MP3 decoding (Linux only)
MP3 (lame)	✅	❌	❌	MP3 encoding (Linux only)
Opus	✅	❌	❌	Opus low-latency codec (Linux only)
Sample Rate Conversion
libsamplerate	✅	✅	✅	High-quality audio resampling (SRC)
Time Stretching
rubberband	❌	✅	✅	Audio time-stretching and pitch-shifting
FFT Implementation
fftw3f	✅	✅	✅	Fast Fourier Transform (single precision, recommended)
Accelerate	—	✅	—	Apple's optimized FFT and DSP framework
ooura	✅	✅	✅	Fallback FFT implementation (always included)

Platform-Specific Notes

Linux (x86_64, ARM64):

• All dependencies are statically linked for maximum portability
• No external .so files required - works on any manylinux-compatible system
• Excludes rubberband and ffmpeg due to static linking constraints
• Includes MP3 and Opus codecs as transitive dependencies of libsndfile

macOS (Intel x86_64, Apple Silicon ARM64):

• Uses native Accelerate framework for optimized FFT operations
• Uses CoreAudio for reading all macOS-supported media formats
• Includes rubberband and ffmpeg for maximum format compatibility
• Separate wheel builds for Intel and Apple Silicon architectures
• Minimum deployment target: macOS 10.15 (Intel), macOS 11.0 (Apple Silicon)

Windows (AMD64):

• All dependency DLLs are bundled inside the wheel via delvewheel
• Fully portable - no separate installation of dependencies required
• Includes rubberband and ffmpeg
• Works on Windows 10+ (x64)

Features NOT Included in Wheels

The following optional features are not included in pre-built wheels but can be enabled when building from source:

• JACK audio server: Real-time audio I/O (requires system JACK installation)
• Intel IPP: Intel's performance primitives (commercial license required)
• BLAS/Atlas: Linear algebra acceleration (minimal benefit for aubio's use cases)
• Double precision mode: Single precision (float32) is used by default

Python Version Support

Pre-built wheels are available for:

• Python 3.10, 3.11, 3.12, 3.13, 3.14
• All wheels include the same feature set per platform

---