Pitch-dependent adjustments of the vocal tract in the singing of Japanese traditional music

The adjustment of vocal-tract resonance can enhance the perceptual effect of singing. For example, resonance tuning involves aligning vocal-tract resonance frequencies near the fundamental or harmonic frequencies, a technique well studied in operatic singing. This study focuses on satsumabiwa, a musical genre of Japanese traditional narrative singing that is accompanied by the biwa, a pear-shaped short-necked lute. We measured the vocal tract of a professional female singer in 3-D using magnetic resonance imaging as she sang three vowels (/a/, /i/, and /u/) at pitches of 174, 233, 349, and 466 Hz. Notably, the highest pitch was produced using the falsetto register, whereas the others utilized the modal register. We obtained the cross-sectional area function from volumetric images and estimated the transfer characteristic of the vocal tract using an acoustic tube model. Our results clearly showed adjustments in the vocal tract, as evidenced by morphological and acoustic data, which varied according to both vowel and pitch conditions. Specifically, as the pitch increased, the frequency of the first formant also increased for the /i/ and /u/ vowels, aligning with the resonance tuning strategy.