Acoustic and Perceptual Characteristics of the Speech of Deaf Children

Abstract

The speech of deaf children differs from that of normal children in many ways. There are articulatory errors such as omission of consonants, insertion of adventitious sounds (typically between syllables and in consonant clusters), substitutions of one phoneme for another, moderate distortions in which the intended phoneme is still recognizable, and severe distortions in which there is no recognizable phoneme. In addition to articulatory errors involving individual phonemes, there are suprasegmental errors such as incorrect rhythm, stress, intonation, phrasing, and errors involving other prosodic features. There are also problems of voice quality such as excessive harshness, breathiness, nasality, and problems of pitch, i.e., average pitch too high, too low, or excessively variable. Acoustic measurements correlate reasonably well with these perceptual observations. On average, the fundamental frequency contours of deaf children tend to be higher than for normal children of the same age and sex, but there are notable exceptions in which the fundamental frequency contour is very low, or exhibits sudden changes from one extreme to the other. Formant frequencies during vowel sounds are similar to those for normal children, but are biased towards values typical of the schwa vowel. Substitution of the schwa for other vowels is a common error. Experiments involving measurements on other acoustic parameters, such as formant transitions and fricative spectra are currently in progress.