A small Python applet that captures audio input from a microphone and visualizes dissonance curves based on the detected overtone frequencies. This is a quick demo that works ok with my Macbook and guitar. Results may vary with microphone quality and environment noise. I'm curious to see if it works with other instruments as well!
Using uv to run the script:
uv run dissonance_curves.py # Use default microphone
uv run dissonance_curves.py -d 1 # Use specific device
uv run dissonance_curves.py --list-devicesDepending on your microphone quality and environment noise, you may need to adjust
the noise threshold multiplier in the compute_noise_threshold function in the
dissonance_curves.py file. I also implemented exponential smoothing of nearby
FFT frames to help stabilize the analysis, but it reduces latency.
- Minute physics video on dissonance curves: https://youtu.be/tCsl6ZcY9ag?t=464
- Dissonance model: https://gist.github.com/endolith/3066664
- Frequency estimation: https://gist.github.com/endolith/255291
- Parabolic interpolation: https://ccrma.stanford.edu/~jos/sasp/Quadratic_Interpolation_Spectral_Peaks.html
- Audio input with sounddevice: https://python-sounddevice.readthedocs.io/en/0.4.6/