Shazam-style audio fingerprinting and song recognition, built for a BWSI/CogWorks capstone.
A from-scratch implementation of audio fingerprinting and song recognition (in the style of Shazam): record a short clip via microphone, extract spectrogram peaks, generate fingerprints, and match against a local song database.
The included notebook, Audio_Capstone_Project_Scaffolding.ipynb,
walks through the whole workflow — populating the database, recording a clip, getting a match, and
plotting the spectrogram/fingerprints — and is the best starting point for understanding how the
pieces fit together.
Uses the course-provided Microphone library and sturdy-garbanzo
(a small sample-song library) as git submodules.
- Python, numpy/scipy (spectrogram + peak-finding), microphone I/O
Built during a Beginning Workshop in Science and Innovation (BWSI) / CogWorks summer program, in collaboration with teammate Hunter Baker.
See the walkthrough notebook, Audio_Capstone_Project_Scaffolding.ipynb, for the full workflow.