Audio landscapes

3D visualization of audio files (currently .wav only).

Computes FFT on windows of .wav chunks to define vertex heights on 3D triangle mesh, with customizable color map and other options. X axis is time; y axis is frequency.

Hobby project for:

• familiarization w/ audio formats (buffering, channels, underruns)
• exploration of 3D visualization options

Examples:

python terrain.py -f data/audio_file.wav

Visualizer options:

• audio_filename: filepath of .wav file to play (and plot)
• cmap: matplotlib colormap to use for plotting (options)
• cmap_agger: method to use to reduce a set of three vertices (defining a face) into a single value (by default, np.mean)
• refresh_ms: time between each plot refresh, in milliseconds
• ignore_threshold: rounds down to zero for any values below threshold (crude 'denoising')
• other parameters defined in terrain.py

Useful links:

• Audio buffer explanation
• PyAudio docs

Attribution/credits:

• Baseline 3D mesh code from Audio-Spectrum-Analyzer-in-Python: heavily refactored for efficiency and modified to accept .wav files and allow customization/parameters