Shazam: Music Recognition using FFT

University of Washington, Seattle

01/2020 - 03/2020

Keywords: Digital Signal Processing | Software Programming | Fast-Fourier Transform | Python

Role

  • DSP Engineer
  • Software Engineer

    • Impacts

  • Implemented DSP theories in a real-world application.

    • Skills

  • Python
  • Digital signal processing
  • Jupitor Notebook
  • Fast fourier transform
  • Data structures

    • Descriptions

    We designed a small program in Python that identifies songs in our database. We collected some songs, used Fast-Fourier Transform to find their frequency spectra and encoded them as fingerprints, and when given audio clips, we can match their frequency spectra with the fingerprints of the songs in our hashmap database and recognize the songs correctly. During the encoding of fingerprints, we plotted spectrograms, ran filterings on the matrices, and binarize them into 0 and 1 signals.

    Back to top Back to Portfolio Back to Home