Each singer maintained her characteristic kinematic pattern regardless of F0 or SPL range, although these did influence aspects of RC and AB behavior. Given the essential role of breathing in classical singing, further work is needed to understand how singers develop their highly individual respiratory strategies and the principles by which each singer's breathing strategy can be optimized.

The results confirm earlier studies that the vented pneumotachograph facemask does not affect breathing behavior in speech research studies and extends the finding to the study of breathing behavior in singing and to the use of a standard respiratory facemask.

The influence of fundamental frequency (F0) on the relationship between sound pressure level (SPL) and spectral balance (SB) has been largely unexplored in the female singing voice. Five classically trained females performed a messa di voce across their musical F0 range. Average maximum SB rose with F0 by 0.27 dB/semitone (ST) to B4 and then decreased, while average minimum SB fell by 0.5 dB/ST to E5 and then generally rose. Of 318 tokens, 208 showed a linear SPL:SB relationship (R(2)>or=0.5), but F0 affected SPL:SB slope and intercept and their interaction above and below B4. The possibility that this reflects a change from subglottal inertance to compliance is discussed. Consistency of SB behavior change at B4 and E5 contrasted with variability in first-formant frequency. Nonlinear SPL:SB relationships did not arise from SB saturation. The presence of low-SPL "tails" may reflect the challenge in modifying vocal-fold adduction during crescendo and decrescendo. The results show that analysis of the SPL:SB relationship must take F0 into consideration.

Breathing instruction for classical singing is becoming more physiologically focused, yet the effect of chest-wall kinematic directives on breathing behaviour is largely unexplored. Five female classical singers sang Caccini's Ave Maria without directive and under two directives: 'steadily pull the abdomen inward' and 'steadily expand the abdomen' through each phrase. The directives had a statistically significant effect on chest-wall dimension at initiation of phrase and on excursion, but dimension at termination of each phrase reverted to habitual behaviour. Rib-cage dimensional change counteracted abdominal change so that lung volume measures were consistent within singer across all breathing conditions. The results have implications for the distinction between consciously controlled and innate respiratory behaviours in singing. Implications for singing pedagogy are discussed.

The messa di voce, in its pure form a crescendo and decrescendo on one note, has been revered for centuries in classical singing, but the pedagogical assumptions of linearity and symmetry have received little critical assessment, especially across a wide fundamental frequency (F0) range. Five trained female classical singers performed a total of 318 messe di voce across their musical F0 range to identify its acoustical characteristics and the influence of F0. Sound pressure level (SPL) range was generally greater during crescendo at higher F0's and during decrescendo at lower FO's. Change in SPL during the messa di voce was predominantly nonlinear, and the shape of the SPL traces differed greatly between crescendo and decrescendo. Nonlinearity in SPL change was not related to SPL range but did show a F0 influence in decrescendo. Change in spectral balance (0-2 vs. 2-4 kHz) with respect to SPL change showed markedly more symmetry than linearity, so that changes in the mode of phonation during the messa di voce were dependent upon SPL regardless of whether the singer was in crescendo or decrescendo. Perceptual and physiological implications are discussed.