The Effect of Breathy and Strained Vocal Quality on Vowel Perception

INTRODUCTION: Research into speech intelligibility in dysarthria historically focuses on articulation deficits. However, voice quality deficits associated with motor speech disorders may also impact speech perception. This study investigates how breathy and strained vocal quality affects vowel identification and ratings of vowel goodness. METHODS: A healthy speaker recorded vowels with normal, simulated breathy and simulated strained voice quality. Acoustic, physiologic, and perceptual measures confirmed the presence of the desired voice deficits. 16 volunteer listeners participated in three perceptual tasks: vowel identification, vowel goodness ratings, and voice quality ratings. RESULTS: In the voice quality rating task, listeners detected voice quality deficits with ease. Breathy and strained stimuli were rated as significantly poorer in voice quality than normal stimuli. The voice quality deficits did not appear to impact vowel identification: identification accuracy for all three sets was high (95% and above) and scores did not differ significantly across the three sets of vowels. Listener judgments of vowel goodness, however, were affected by voice quality. Breathy and strained vowels were rated as significantly poorer than normal vowels. In addition, listeners needed more time to rate the articulatory goodness of the disordered stimuli and replayed them more often while making their goodness judgments. CONCLUSION: Simulation of voice quality deficits appears to be a valid way of assessing the impact of speech factors beyond articulation on the perception of disordered speech. Stimuli with simulated breathiness and strain were rated as poorer in voice quality than normally voiced vowels, indicating that voice quality is salient to listeners. Although identification accuracy was not affected by voice quality deficits, breathy and strained vowels were judged as poorer in articulatory goodness than normally voiced vowels. Abnormal voice quality appeared to interfere with listener judgments of the articulatory goodness of vowels. Voice quality deficits associated with dysarthria may affect speech perception by causing increased listener effort even if speech intelligibility is not directly affected. Further study of the effect of voice quality in more realistic listening conditions (e.g., in noise) with more complex speech stimuli (e.g., sentences or conversation) will help determine the need for phonatory treatment of dysarthric speech