User Tools
simple_chord_recognition
====== Differences ====== This shows you the differences between two versions of the page.
|
simple_chord_recognition [2014/01/26 16:58] 160.39.250.229 |
simple_chord_recognition [2014/01/28 17:03] (current) camelcase |
||
|---|---|---|---|
| Line 1: | Line 1: | ||
| - | [[image:templates.png]] | + | ====== Chord Recognition ====== |
| + | |||
| + | | Authors | Jeff Scott and Alex Cannon | | ||
| + | | Affiliation | Drexel University and Swathmore College | | ||
| + | | Code | [[https://github.com/jeffscott/HAMR_2014|Github Link]] | | ||
| + | |||
| + | Templates are constructed by taking the median spectrum of sections of audio with chord labels | ||
| + | {{:templates.png?400|templates.png}} | ||
| + | |||
| + | * Each column is a template for a specific chord | ||
| + | * For test songs, compute euclidian distance between all templates and current spectral frame | ||
| + | * Choose best match | ||
| + | * 24.59% accuracy for root only | ||
| + | * Trained on 40 songs, tested on 70 songs | ||
| + | |||
| + | Here is a better view of some of the example templates | ||
| + | |||
| + | {{:small1.png?400|small1.png}} | ||
| + | {{:small2.png?400|small2.png}} | ||
simple_chord_recognition.1390773538.txt.gz · Last modified: 2014/01/26 16:58 by 160.39.250.229
Page Tools
Except where otherwise noted, content on this wiki is licensed under the following license: CC Attribution-Share Alike 3.0 Unported
