Use of syllable-scale timing to discriminate words.

Assuming that only primitive, imperfect phonetic labeling of speech could be achieved by an automatic speech segmenter, words from several small vocabularies were identified using discriminant analysis, where the locations of prominent acoustic boundaries were combined in the optimum linear fashion. A set of two-syllable words that shared the same spelling in a weak alphabet (using categories like stop closure, fricative, vowel, etc.), but that differed in such features as which syllable was stressed, the tensity of the vowel, the identity of particular segments, etc., was selected. Discriminant analysis on the vector of six segmental boundaries achieved recognition accuracy 6.3 times better than chance. The words were produced by six talkers at two speaking tempos and measured by hand from sound spectrograms. Testing was performed on productions different from those used in training. When more confusable words (sharing the same stress pattern) were employed, performance was still five times better than chance. In a third experiment, the first two word sets were combined, and a subset of the variables (four of them) shared in common was employed for identification. This time recognition using temporal information was still able to do a reasonable job of discriminating the words. The results suggest that considerable information is available in segmental timing for identifying words even when the phonetic labeling of speech is weak.

Evidence for mora timing in Japanese.

Japanese has long been described as a "mora-timed" language by linguists. Japanese pedagogy has traditionally claimed that moras are constant in duration. Four experiments are reported investigating segmental timing in Japanese in order to test several straightforward hypotheses about mora timing. First, it is demonstrated that words with an increasing number of moras increase in duration by nearly constant increments. The next two experiments explored the mechanisms by which constant mora durations are achieved given that there are large universal differences in the inherent duration of various segment types (e.g., /u/vs/a/), and given that some syllables are supposed to be two mora long (such as, those with long vowels or final consonants). In each case, it was found that the duration of a word stays very close to a target duration that depends on the number of moras in it. This is achieved by stretching or compressing the duration of neighboring segments and adjacent moras. Thus increasing the number of segments in two-mora syllables results in lengthening, not the expected shortening, of other segments in the heavier syllable.

Linguistic timing factors in combination.

This study attempts to determine the ways in which linguistic timing factors combine with each other in the production of English and specifically, to test and explore aspects of the timing model of Klatt [J. Acoust. Soc. Am. 59, 1208-1221 (1976)]. In two experiments the tensity feature of a sentence stressed vowel and the voicing of the following stop were changed along with a variable that alters the length of the VB syllable. The duration of both the vowel, V, and the following stop closure, C were measured. In the first experiment either one or two unstressed syllables are added to the word and in the second speaking tempo is changed. The significant results of both experiments are accounted for with a formal timing model that separately specifies (1) the Vm + C duration (where Vm = mean duration across the two vowels /i/ and /I/) depending on tempo or word length, (2) the ratio Vm/C depending on the voicing of the stop (/b/ or /p/), and (3) as a final ordered step, a rule that adjusts the vowel length by a constant ratio depending on the identity of the vowel. This integrated model provides for (1) the "incompressiblity" effects, (2) prosodic properties, and (3) the constant V/C ratios for values of the voicing feature found in perception experiments.

Relation between voice-onset time and vowel duration.

As part of an investigation of the temporal implementation rules of English, measurements were made of voice-onset time for initial English stops and the duration of the following voiced vowel in monosyllabic words for New York City speakers. It was found that the VOT of a word-initial consonant was longer before a voiceless final cluster than before a single nasal, and longer before tense vowels than lax vowels. The vowels were also longer in environments where VOT was longer, but VOT did not maintain a constant ratio with the vowel duration, even for a single place of articulation. VOT was changed by a smaller proportion than the following voiced vowel in both cases. VOT changes associated with the vowel were consistent across place of articulation of the stop. In the final experiment, when vowel tensity and final consonant effects were combined, it was found that the proportion of vowel duration change that carried over to the preceding VOT is different for the two phonetic changes. These results imply that temporal implementation rules simultaneously influence several acoustic intervals including both VOT and the "inherent" interval corresponding to a segment, either by independent control of the relevant articulatory variables or by some unknown common mechanism.