Statistical representation of word-initial obstruents: General considerations
Gary G. Weismer, Karen M. Forrest, Paul Milenkovic · The Journal of the Acoustical Society of America · 1987
The description of obstruent spectra has typically been done in categorical terms [e.g., S. E. Blumstein and K. N. Stevens, J. Acoust. Soc. Am. 66, 1001–1017 (1979)]. Whereas this may be satisfactory for some purposes, the categorical approach is not helpful in cases where obstruent sounds are produced that are not easily described according to the normal set of segmental contrasts. In particular, many dysarthric and apraxic subjects produce stops and fricatives that give the perceptual impression of being “between” the normal segmental categories. The purpose of this paper is to describe some of the general considerations that led us to develop measures capable of indexing spectra associated with such indeterminate obstruent productions. A statistical representation of obstruent spectra will be described that is similar to certain approaches to describing a more general class of impulse noises [J. Erdreich, J. Acoust. Soc. Am. 79, 990–998 (1986)]. [Work supported by NIH Awards NS13274 and NS20976.]