Automatic musical key detection
Name
Anna Aljanaki
Abstract
In this thesis we have proposed a model for tonality estimation, which is capable of handling music coming from various musical traditions and does not require their thorough analysis. In our model we have employed an assumption, that most musical traditions use duration to maintain pitch salience. Proceeding from this assumption, we have proposed an algorithm for automatic key detection, based on a distributional approach. The proposed method was evaluated on both symbolic and acoustic datasets.
Graduation Thesis language
English
Graduation Thesis type
Master - Information Technology
Supervisor(s)
Konstantin Tretjakov
Defence year
2011