Year: 1963
Type: conference-abstract
Abstract: A polymodal vocoder was developed, based on a concept of flexibility for present and future data channels and data-handling equipment for digital-speech communication. The polymodal feature provides a user with a choice of three primary modes of vocoder operation, consisting of three digital-transmission rates with speech intelligibility, quality, and ability to recognize voices scaled in accordance with the data rate utilized. The minimal rate (1200 bits per second) is a 14-channel vocoder configuration; the conventional mode (2400 bits per second) is an 18-channel configuration, while the voice-excited vocoder mode (9600 bits per second) has two arrangements, one utilizing an 800-cps baseband and another with a 950-cps baseband. Spectrum-pattern data are normalized with respect to speech amplitude, for compatibility of the polymodal vocoder with an auxiliary processor for spectrum-pattern matching to further reduce the data-rate requirement and provide additional data rates. The polymodal system also lends itself to a concept of a fully adaptive system, self-adjusting to match available channel capacity. Design features of the polymodal vocoder are discussed, and recordings of the various modes presented. [This program was supported by Air Force Cambridge Research Laboratories (OAR) under contract AF19(628)-302].
DOI: 10.1121/1.2142545
Authors: L. A. Yaggi, A. E. Mason
Institution: Texas Instruments (United States)
Language: English
Topic: Speech and Audio Processing
Subfield: Signal Processing
Field: Computer Science
Domain: Physical Sciences
Open Access status: bronze
OpenAlex ID: W2009200731