| Summary: | 碩士 === 國立臺灣大學 === 電機工程學研究所 === 101 === Measuring rhythmic similarity in traditional methods often focus only on the position and length of onsets, but lack of the concept of dynamics. Dynamics are means of expression, it denotes the volume (degree of loudness or softness) at which music is played, and it can affect audiences’ emotional response [1].
When considering different position of onset, we also introduce the metrical salience, which is, the perception of onsets at different positions in a music measure are different. In general, events in positions that are perceived as salient are memorized and recalled easier, attract primary attention, are more expected to occur, and, when they are verified, the audience will sense more difference. In our method, we take both dynamics and metrical salience into consider and build a new algorithm to measure the similarity between two rhythms.
|
|---|
