For various reasons, audio recordings made of speakers with dysarthria or other communicative disorders are often made using head mounted microphones. The close proximity of the mouth to the microphone may result in an elimination of the typical reverberant quality of speech. Naturally occurring reverberation can reduce speech intelligibility while the speech may sound more natural to listeners. Thus, using head mounted microphones to record speech samples may result in evaluations of speech abilities that do not reflect a patient's abilities in typical communicative situations. This study examined the speech of four speakers with dysarthria whose speech was varied from a non-reverberant condition to two reverberant conditions: a natural level of reverberation and an excessive level. Measures of Acceptability ("How would you rate this person's speaking skills?"), a subjective measure, and Intelligibility (percentage of correct words identified), an objective measure, were obtained using 30 normal young adults. Acceptability ratings showed differences between speakers but not across reverberant conditions. Intelligibility ratings showed differences between speakers and differences between reverberant conditions for two of the four speakers. The non-reverberant condition resulted in the highest intelligibility ratings. Moderate significant correlations were found between Acceptability and Intelligibility ratings for the two mid-range speakers with dysarthria. Suggestions are made for recording techniques used to assess the speech status of persons with dysarthria.
展开▼
