Xxii. Speech Communication A. Performance of the Articulatory Analog of the Speech Mechanism: a Report on the Status of Research

The objectives of our work are to further our understanding of: (a) the process whereby human listeners decode an acoustic speech signal into a sequence of discrete linguistic symbols such as phonemes; and (b) the process whereby human talkers encode a sequence of discrete linguistic symbols into an acoustic signal. Current research activities related to these objectives include experiments on the generation of speech by electrical analog speech synthesizers, development of means for controlling analog speech synthesizers by a digital computer, measurements of movements of the speech-generating structures during speech production, studies of methods of speech analysis, accumulation of data on the acoustic characteristics of utterances corresponding to phonemes in various linguistic contexts, and studies of the perception of speechlike sounds. On other occasions Rosen l ' 2 has described the various components of the dynamic transmission-line vocal-tract analog (familiarly, DAVO) in detail, and has also described experiments on vowel production and fricative-consonant production; Fujimura 3 has discussed stop-consonant production. More recently, Hecker 4 has reported on the development of the nasal components that have been added to the analog and has described experiments dealing with the production of nasal consonants. During the past several months the sound vocabulary of the synthesizer has been extended still further, and the results of the earlier studies have been incorporated into a general scheme for programming the analog. The analog of the vocal tract and nasal cavities which is under discussion is realized by two electrical transmission lines that represent the acoustic pathways of the speech mechanism above the level of the vocal folds. Sources are available for appropriately