A phonetically relevant auditory-perceptual space
James D. Miller · The Journal of the Acoustical Society of America · 1982
The vowels, based on averages of Peterson and Barney [J. Acoust. Soc. Am. 24, 175–184 (1952)], can be described independently of the age and sex of the talkers in a space with axes: x = log(F3/F2), y = log(F1/F0′), and z = log(F2/F1), where F0′ = aF0, F1, F2, and F3 are the center frequencies of the first three formants and F0 is the fundamental. The constant a varies across speakers, but the average value for men is 1.478 times that for children and women. So described, the vowels fall within a tetrahedron, whose properties will be described, and their spatial organization appears to reflect their perceptual and articulatory organization and to suggest rules for the enlargement or compression of their spectra. Distances between vowels are in log frequency and are easily converted to cents, semitones, or octaves. As examples, the neighboring vowels, /ɪ/ and /ɛ/ are separated by 0.76 octaves while the distant vowels, /ɑ/ and /i/ are separated by 3.2 octaves. New vowel data and possibilities for locating the consonants in the space will be mentioned. [Supported by NINCDS Grant NS 03856.]