The Reliability of Acoustic Measures of Speech Analyzed Using PRAAT
Heidi Berens, Allison Greene, Jennifer Ireland, Ann Stommer · Cornerstone (Minnesota State University, Mankato) · 2008
In clinical practice, speech-language pathologists regularly perceptually analyze certain aspects of prosody (e.g., pitch, stress), voice quality (e.g., hoarseness, nasality), and phonetics (i.e., speech sounds) that might more reliably be measured using instrumental (i.e., acoustic) tools. Two barriers to the application of acoustic analysis are (1) the cost and size of the instruments and (2) the fact that the most commonly used instruments only analyze short segments of speech. Recently, PRAAT, free computer software for the acoustic analysis of speech, has become available and members of the Computer Science Department have developed tools to facilitate the use PRAA T. The purpose of this research was to explore the reliability/stabilityof selected acoustic measures generated using the modified version of PRAA T on samples of speech from adolescents who are typically developing and who have Williams syndrome. (Williams syndrome is a genetic disorder that results from a deletion on chromosome 7.)