Implementing a Sound Identifier in Python
Offered By: EuroPython Conference via YouTube
Course Description
Overview
Explore the implementation of a Shazam-style sound recognizer in Python during this 22-minute conference talk from EuroPython 2016. Dive into Digital Signal Processing (DSP) techniques and powerful libraries used to create an audio identification system. Learn about the project's structure, including a classifier for fingerprinting and storing audio, and a recognizer for matching smaller audio chunks. Discover the journey from concept to implementation, addressing challenges and future improvements. Gain insights into music information retrieval, normalization, fingerprinting, spectrograms, nearest neighbor algorithms, and hash storage. No prior DSP knowledge required – only Python experience needed to follow along. Access the project's code on GitHub and understand how it was inspired by a FOSDEM 2016 talk on over-the-air audio identification.
Syllabus
Introduction
Music Information Retrieval
Why Python
Demo
Normalizer
Fingerprint
Diagram
Spectrogram
Nearest Neighbor
Anchor Points
Hash
Storage
Deja Vu
Shazam
Genius
Notebook
MusicBrainz
Taught by
EuroPython Conference
Related Courses
Fundamentals of Electrical EngineeringRice University via Coursera Digital Signal Processing
École Polytechnique Fédérale de Lausanne via Coursera Fundamentals of Electrical Engineering Laboratory
Rice University via Coursera Processamento Digital de Sinais - Amostragem
Universidade Estadual de Campinas via Coursera Physics-Based Sound Synthesis for Games and Interactive Systems
Stanford University via Kadenze