Narrowband voice system.

Abstract

The problem of how to accommodate more simultaneous users in a given bandwidth communications channel was examined. It was determined that voice signals could be reduced in bandwidth by 50 percent. Two speech formants, Fl and F3 , were filtered from the speech signal and frequency translation performed to form a compact audio signal. At the receiver this process is reversed and the formants returned to their normal spectral positions. Intelligibility scores of 94 percent were obtained in a listening test conducted using the two formants Fl (250-650 Hz) and F3 (1950-2450 Hz).