Laryngeal and aerodynamic adjustments for voicing versus devoicing of /h/: a within-speaker study

The aim of this within-speaker case study was to explore how effectively a phonetically trained speaker could alter the likelihood of voicing around abduction, and what changes he made to do so. An American English-speaking male produced intervocalic /h/ in varying loudness and vowel contexts. When given no specific instructions about voicing (block 1), he produced almost entirely voiced /h/. He was then asked to devoice /h/ (block 2). Measures of voicing, baseline airflow, pulse amplitudes, fundamental frequency (F0), open quotient, and speed quotient were made from oral airflow signals. Subglottal pressure was estimated from intraoral pressures during /p/. In block 2, the speaker produced 70% devoiced /h/. He achieved this by making several changes associated with higher phonation threshold pressures: greater abduction degrees, lower subglottal pressures, greater longitudinal tension of the vocal folds, and altered laryngeal settings. Qualitative inspection of the DC flow contours along with correlational and principal components analyses indicated widespread changes in respiratory, laryngeal, and supralaryngeal settings, and differing interrelationships among variables. Our speaker showed tacit knowledge of the range of parameters affecting voicing. Differing relationships among variables across the two blocks support a view of phonation as a dynamic process, where speakers adjust multiple parameters, simultaneously.