Use of ambiguous visual stimuli to demonstrate the value of acoustic cues in speech perception

Abstract

The purpose of this study was to evaluate a set of English phrases and sentences in a minimal-pairs syllables-test format for assessing use of acoustic cues in the audiovisual perception of speech; in particular, by persons with severe or profound hearing loss. Test items that were not rejected were categorized as “acceptable” or “ideal.” An acceptable item was visually confusable but auditorally distinct compared to its foil (analogous to /p/ vs/m/). An ideal item was, in addition, tactilely distinct. Forty-eight items, spoken by a male and a female talker, were presented in a videotaped two-alternative forced-choice format under three conditions to three young-adult subject groups. In a lipreading- without-sound condition, 39 items were visually confusable; of these, 16 items were reliably identifiable when sound was added (the “acceptable” items). Seven of the 16 acceptable items were also identifiable in a tactile-pattern-alone condition (the “ideal” items). Performance on the 16 acceptable items was analyzed for an additional 27 young adults with moderate, severe, or profound hearing loss. Results obtained with the new protocol were consistent with conventional assessment results (CID Everyday Sentences, with and without sound) in all but 4 cases. Communication characteristics of the exceptions were examined to explain lack of agreement.