Studying fricatives with Ken Stevens and ever since

I began studying fricatives as Ken Stevens’ Ph.D. student at MIT, following John Heinz’ experiments and Ken’s 1971 JASA paper. Experiments with simple mechanical models allowed source location and spectra to be measured. Using these sources in transmission-line analogues showed predicted and measured far-field spectra matched, validating the use of such models for voiceless fricatives. Parameters that captured the acoustic effects of changes in model dimensions and airflow distinguished sibilant and non-sibilant cases, and the place of the constriction. In the years since, I have used aerodynamic measurements (e.g., Rothenberg mask, hot-wires) to explore the interaction of convection and acoustic velocities in the vocal tract. These phenomena are important in voiced fricatives and whistling and are ignored in classic models. Signal processing techniques (ensemble averaging, multitaper analysis, spectral decomposition) have, e.g., demonstrated the interaction of phonation and turbulence sources in voiced fricatives. Spectral parameters allowed comparison of fricatives of different languages, or phonetic contexts; of speakers of different ages and genders, and with speech disorders. My goal has shifted from improving articulatory synthesis to defining acoustic parameters that allow production parameters to be inferred. Open questions and the many links to Ken’s lab and collaborators will be described.