Realtime_PyAudio_FFT

A simple package to do realtime audio analysis in native Python, using PyAudio and Numpy to extract and visualize FFT features from a live audio stream.

Demo Video

The basic pipeline:

• Starts a stream_reader that pulls live audio data from any source using PyAudio (soundcard, microphone, ...)
• Reads data from this stream many times per second (eg 1000 updates per second) and stores that data in a fifo buffer
• When triggered by .get_audio_features(), the stream_analyzer, applies a Fast-Fourier-Transform to the most recent audio window in the buffer
• When visualize is enabled, the visualizer displays these FFT features in realtime using a PyGame GUI (I made two display modes: 2D and 3D)

Requirements:

pip install -r requirements.txt

You also might have to sudo apt install libasound-dev portaudio19-dev libportaudio2 libportaudiocpp0 (tested on Ubuntu)

I developped this code on my local machine --> it has not been properly tested on other setups.. If something doesn't work, please first try to fix it yourself and post an issue/solution when appropriate!

• Tested on Ubuntu 18.04
• Other platforms like Mac/Windows should work if PyGame can find your display and Python finds your audio card (these can be tricky with WSL)
• For Mac OSX (tested on Catalina 10.15.4), please make sure you run with Python downloaded from Python.org (pygame doesn't work well with the default/Homebrew Python)

Tested with:

• Python 3.6.3
• pygame --> Version: 1.9.6 &
• pyaudio --> Version: 0.2.11
• scipy --> Version: 1.4.1

Alternatively to pyaudio, you can use sounddevice which might be more compatible with Windows/Mac

• just run python3 -m pip install sounddevice
• Tested on Ubuntu 18.04 with sounddevice version 0.3.15
• The code to switch between the two sound interfaces is in the __init__ function of the Stream_Analyzer class

Usage:

just run python run_FFT_analyzer.py and play a sound on your machine!

• I have personally learned A LOT about sound by watching this realtime visualization while listening to music
• You can run the stream_analyzer in headless mode and use the FFT features in any Python Application that requires live musical features

ToDo:

• Implement realtime beat detection / melody extraction on top of FFT features (eg using Harmonic/Percussive decomposition)
• The pygame.transform operations sometimes cause weird visual artifacts (boxes) for some resolution settings --> fix??
• Remove the matplotlib dependency since it's only needed for the colormap of the vis..
• Slow bars decay speed currently depends on how often .get_audio_features() is called --> fix