Use of Spectral/Cepstral Analyses for Differentiating Normal From Hypofunctional Voices in Sustained Vowel and Continuous Speech Contexts PurposeIn this study, the authors evaluated the diagnostic value of spectral/cepstral measures to differentiate dysphonic from nondysphonic voices using sustained vowels and continuous speech samples.MethodologyThirty-two age- and gender-matched individuals (16 participants with dysphonia and 16 controls) were recorded reading a standard passage (The Rainbow Passage; Fairbanks, 1960) and sustaining the ... Article
Article  |   December 01, 2011
Use of Spectral/Cepstral Analyses for Differentiating Normal From Hypofunctional Voices in Sustained Vowel and Continuous Speech Contexts
 
Author Affiliations & Notes
  • Christopher R. Watts
    Texas Christian University, Fort Worth, TX
  • Shaheen N. Awan
    Bloomsburg University of Pennsylvania
  • Correspondence to Christopher R. Watts: c.watts@tcu.edu
  • Editor: Anne Smith
    Editor: Anne Smith×
  • Associate Editor: Mahalakshmi Sivasankar
    Associate Editor: Mahalakshmi Sivasankar×
Article Information
Speech, Voice & Prosody / Speech
Article   |   December 01, 2011
Use of Spectral/Cepstral Analyses for Differentiating Normal From Hypofunctional Voices in Sustained Vowel and Continuous Speech Contexts
Journal of Speech, Language, and Hearing Research, December 2011, Vol. 54, 1525-1537. doi:10.1044/1092-4388(2011/10-0209)
History: Received July 29, 2010 , Revised December 20, 2010 , Accepted April 25, 2011
 
Journal of Speech, Language, and Hearing Research, December 2011, Vol. 54, 1525-1537. doi:10.1044/1092-4388(2011/10-0209)
History: Received July 29, 2010; Revised December 20, 2010; Accepted April 25, 2011
Web of Science® Times Cited: 32

PurposeIn this study, the authors evaluated the diagnostic value of spectral/cepstral measures to differentiate dysphonic from nondysphonic voices using sustained vowels and continuous speech samples.

MethodologyThirty-two age- and gender-matched individuals (16 participants with dysphonia and 16 controls) were recorded reading a standard passage (The Rainbow Passage; Fairbanks, 1960) and sustaining the vowel /ɑ/. Recorded voices were analyzed with custom software that calculated 4 spectral/cepstral measures.

ResultsMeasures of cepstral peak prominence (CPP) and low–high spectral ratio (L/H ratio) were significantly different between groups in both speaking conditions; the standard deviation of the CPP was significantly different between groups in continuous speech only. In differentiating dysphonic individuals with a hypofunctional etiology from nondysphonic individuals, receiver operating characteristic (ROC) analyses demonstrated (a) high sensitivity and high specificity for the CPP in the sustained vowel condition and (b) high sensitivity and moderate specificity for the CPP in the speech condition.

ConclusionsIn a sample of dysphonic speakers (hypofunctional etiologies) versus typical speakers, spectral/cepstral measures of CPP and L/H ratio were able to differentiate these groups from one another in both vowel prolongation and continuous speech contexts with high sensitivity and specificity. The results of this study support the growing body of literature documenting the significant value of cepstral and other spectral-based acoustic measures to the clinical evaluation and management processes.

Acknowledgments
We would like to acknowledge Emily Lambert, Sara Miesemer, and Teresa Nicolia for their assistance on this project. We would also like to acknowledge Kimberly Coker and Shelby Diviney of the D. Wayne Tidwell Voice and Swallowing Center in Fort Worth, Texas, for their contributions to this project.
Order a Subscription
Pay Per View
Entire Journal of Speech, Language, and Hearing Research content & archive
24-hour access
This Article
24-hour access