The relative contribution to speech intelligibility from consonants and vowels using synthesized and naturally-spoken sentences Free

Several studies have attempted to understand why intelligibility is maintained in sentences or words in which all consonants were replaced by silence or noise. Studies have traditionally used naturally spoken stimuli to study these effects. Cole et al. (1996) found that when sentences were presented with only vowels or only consonants intact, participants performed significantly better in the vowel only condition. Others [Kewley‐Port et al. (2007); Fogerty and Kewley‐Port (2009); Lee and Kewley‐Port (2009)] have attempted to assess the source of intelligibility improvements for the vowel-only conditions without conclusive results. Naturally spoken stimuli have multiple imbued cues that are difficult to control and thus present challenges in teasing out the relative contributions of each cue. It has been proposed that synthesized speech may provide a method to assess the specific contribution of the cues included in consonants and vowels to speech intelligibility. The goals of this study were to determine if the intelligibility scores were the same for similarly processed synthesized and naturally produced sentences, and to define the extremes of intelligibility using different synthesis techniques. Participants performed similarly in synthesized and naturally produced speech conditions, supporting the use of synthesized speech in determining the relative contribution of vowel cues.