Basic research for computer-based musical analysis of utai
Leader: ITŌ Katsunobu
Member: CAROREPETTO Rafael, YAMANAKA Reiko, MIYAMOTO Keizō
Aims: There has been rapid development in the techniques of music information processing (MIR), using computers to process, analyze, and generate acoustic signals. The techniques of MIR, however, are implicitly based on Western music, which causes problems when dealing with non-Western music such as utai (the vocal music of Noh). CompMusic (Computational Models for the Discovery of the World’s Music) is an attempt to develop a new range of information modeling techniques that are not based on characteristics particular to Western music, but rather respect the cultural specificity of various traditional musics. In order to build a research base utilizing new insights from the fields of musicology and MIR, we will: 1. establish a method for the analysis of musical pieces employing Western staff notation and the notation of acoustic signals, together with the traditional notation of utaibon scores; and, 2. survey the musical terminology of utai, and its translation into English.
Methods: 1. By establishing a method for analysis employing traditional notation in addition to staff notation and acoustic signals, we will establish a base for facilitating the utilization of earlier insights from musicology and MIR. 2. In order to facilitate the utilization of new insights from these fields, it will be necessary to grasp clearly what the musical phenomena of utai have in common with those of other musics, and where they differ. We will expand existing lists of translations of Noh music terminology based on a survey of musicological literature on the subject.
For the analysis of Noh singing, we observed comparatively among score information, a commentary of the Noh singing, and the acoustic signals extracted from the commercial compact discs. As a result, we confirmed vibrato depth in Noh singing is greater than that in Western music, and the pitches of scale notes are not absolute and are changeable even within a single phrase. In addition, an onset of a note in Noh singing is more flexible and not as important as it is in Western music. Based on this knowledge, we proposed a continuous notation that the notation used in the previous commentary fit to an acoustic signal smoothly.
We also developed an automatic melody estimation method using our proposed notation. We confirmed a merit of the proposed method, because such different onsets correspond to different scores in the discrete notation, similar to a score of Western music. In addition, we evaluated the resultant melody by listening vocal melodies synthesized by the vocaloid using the translated Western music style score from our notation.
We presented this study in the international conference for contents processing and won the best paper award.
For the survey of English translation of terms for Noh singing, we could not find any definitive translation. Even in musicology, there are not so many terms for the ornaments and melisma, which is the singing of a single syllable of text while moving between several different notes in succession, but these phenomena are salient and important in Noh singing. We have to survey explanations for these phenomena among the previous research. In addition, we found some explanation using Western music score notation for ornaments of Noh singing with literal translated terms.
2016Katunobu Itou, Rafael Caro Repetto and Xavier Serra, “Melody Transcription Framework using Score Information for Noh Singing,” CONTENT 2016, Rome, (2016/3) (Best Paper Award)