A resonance-vocoder and baseband complement: A hybrid systems for speech transmission

The production and perception of speech are characterized by certain constraints. The efficiency with which speech information can be communicated is considerably dependent upon how successfully these constraints can be incorporated into the transmission apparatus. This paper undertakes to review some of these relations and to indicate their relevance to communication. It attempts to point out the extent to which vocal and auditory constraints are applied in existing speech compression systems such as the Vocoder and the resonance-Vocoder. The theory underlying the latter suggests savings in channel capacity greater than tenfold. To date, however, only moderate success has been achieved in transmitting speech of acceptable intelligibility and quality over such systems. The last part of the paper describes a transmission scheme which is part conventional and part resonance-Vocoder. Through this compromise, some of the band-saving features of the latter can be retained while its intelligibility and quality can be improved upon. The hybrid system occupies a total bandwidth of approximately 600 cps and requires a signal-to-noise ratio comparable to that of conventional voice channels. Articulation tests performed on the system yield mean scores ranging from 74 to 84 per cent for monosyllabic words.