Target zones for synthetic vowels
James D. Miller, John W. Hawks · The Journal of the Acoustical Society of America · 1989
Using algorithms developed by the second author (JWH), he synthesized 7703 vowel tokens such that the values of F0, F1, F2, and F3 uniformly sampled locations in a plane in the middle of the vowel slab of the auditory-perceptual space [J. D. Miller, “Auditory-perceptual interpretation of the vowel,” J. Acoust. Soc. Am. (in press)]. Each token was 400 ms and had a rise-fall pitch contour. The tokens were arranged in random order and presented in groups of 100 to one listener (JDM), who identified each token as [IY, IH, EH, EY, AE, AA, AH, AO, OW, UH, or UW] and rated the clarity of each token on a scale from 1 (poor) to 5 (excellent). The sizes and shapes of the target zones derived from these experiments will be compared to those based on measurements of natural speech. The data are, in general, consistent with the concept of vowel target zones as being large with irregular and abutting boundaries. [Work supported by NINCDS.]