<< Chapter < Page Chapter >> Page >

In order to compare songs, we can generate match scores for them using a matched filter. We wanted a filter capable of taking the spectral peaks information generated by the fingerprint finding algorithm for two different songs and produce a single number that would tell us how much the two songs being compared look alike. We wanted this filter to be as insensitive as possible to noise and produce a score that is independent of the length of each recording.

Our approach to this was completely different from that used by the creators of Shazam, as we did not use Hash tables at all and did not combine the peaks into pairs limited by certain regions, as they did. In the end we still managed to get very good accuracy and decent performance by using a matched filter.

Get Jobilize Job Search Mobile App in your pocket Now!

Get it on Google Play Download on the App Store Now




Source:  OpenStax, Digital song analysis using frequency analysis. OpenStax CNX. Dec 19, 2009 Download for free at http://cnx.org/content/col11148/1.1
Google Play and the Google Play logo are trademarks of Google Inc.

Notification Switch

Would you like to follow the 'Digital song analysis using frequency analysis' conversation and receive update notifications?

Ask