Our Research Lab

Voice Quality Methods

New Methods for Evaluating Voice Quality

VQ

Voice quality refers simply to the sound quality of a person's voice.  Each person has a unique voice quality but characteristics common with other voices.  Often, laryngeal pathology and neurological disorders lead to physical changes in the voice production system that influence voice quality.  For example, if one vocal fold is weakened, lesioned, or paralyzed, then voice quality will be more "breathy" than the voice otherwise would be.  An index of voice quality, therefore, may serve as important clinical metric, potentially providing initial symptoms of the underlying problem and often used as a measure of treatment outcome. 

In general, measurement methods for voice quality lack the rigor used in the study of broader sound quality. The use of voice quality measures clinically would benefit from methods co-registered or validated with data from large numbers of "listeners."  Furthermore, scales of voice quality would allow one to quantify how much voice quality changes due to a pathology or to treatment.  Finally, if we can accurately predict voice quality perception based on computational models, we may be able to automate certain aspects of voice quality assessment.

VQ

Our work in voice quality has involved listening experiments using a variety of measurement methods in naïve and expert listener panels. It also has involved the development of computational methods for estimating voice quality perception based on the use of pre-processing of the vocal acoustic signal with simple models of sound processing by the auditory periphery (outer, middle, inner ear, and auditory nerve).

We have shown that measurement of voice quality can be highly precise, repeatable, and predictable. Based on these results, we have developed novel scales for evaluation of voice quality and prototype software for use in clinical settings. We continue to make advancements in this area and work towards valid, reliable, precise, and practically useful measures of voice quality.

VQ

VQ track

Please consider becoming a research participant!