著者

ラチンスキスタニスワヴアンジェイ 小野 順貴 嵯峨山 茂樹

出版者

一般社団法人情報処理学会

雑誌

情報処理学会研究報告音楽情報科学(MUS) (ISSN:09196072)

巻号頁・発行日

vol.2008, no.12, pp.193-198, 2008-02-09

参考文献数

13

---

We propose a new approach for dealing with multipitch analysis of musical signals that makes use of the fact that such signals are highly structured. This structure comes from the many musicological rules of the western tonal music and we model it by using the recently developed method of Hierarchical Hidden Markov Models. We propose a model with four layers: song key chord and note combination layer. One of the big advantage of this approach is that besides from information about pitches we get higher level musical information about chord progression and key modulation.We propose a new approach for dealing with multipitch analysis of musical signals that makes use of the fact that such signals are highly structured. This structure comes from the many musicological rules of the western tonal music, and we model it by using the recently developed method of Hierarchical Hidden Markov Models. We propose a model with four layers: song, key, chord, and note combination layer. One of the big advantage of this approach is that, besides from information about pitches, we get higher level musical information about chord progression and key modulation.