Abstract:
Various factors such as listeners’ familiarity & experience with disordered speech, listener’s comprehension, ability to predict and the cues provided by the context are considered to be crucial in the the assessment of speech intelligibility in persons with dysarthria. This study addressed the issue of listening competency in normal hearing listeners and its effects on predictability of target words embedded in sentences in naturally degraded speech samples of persons with dysarthria & artificially degraded sample (where distortion was added to the speech of a model speaker). In Experiment 1, thirty normal hearing adults rated the 38 High predictable (HP) and 38 Low predictable (LP) sentences in non degraded and artificially degraded conditions of stimuli produced by a ‘model speaker’. In Experiment 2, normal hearing adults rated the intelligibility of the naturally degraded speech samples of 3 persons with dysarthria. Experiment 1 revealed that the mean scores of HP sentences were better & statistically significant compared to LP sentences and the overall combined mean scores of non degraded and artificially degraded stimuli of HP sentences were better compared to the LP sentences. Experiment 2 revealed that the mean scores of HP and LP sentences produced by persons with dysarthria was significantly different. The scores in the HP context of the ‘model speaker’ (degraded condition) was similar to HP score of the first and third sample of the persons with dysarthria. The LP sentence of the ‘model speaker’ was similar to the LP score of the third sample of the person with dysarthria. The listening competence amongst the listeners varied across degraded and non degraded HP and LP sentence contexts and the degraded LP sentences were sensitive in evaluating the listening competence of normal listeners’ as it was devoid of all the contextual cues for the assessment of speech intelligibility, rendering the task difficult, thus having good potential in tapping the listeners competence.