High-gain digital speech interpolation with adaptive differential PCM encoding

A generalization of digital speech interpolation (DSI) combined with adaptive differential pulse-code modulation (ADPCM) is described. The general configuration of the DSI-ADPCM system is detailed, and the speech encoding strategy in a heavily loaded environment is described. Advantages obtained by incorporating a voiceband data discriminator into the DSI-ADPCM are discussed. The ADPCM encoding with voiceband shifts and sampling rate conversions is analyzed and the simulated performance of the ADPCM encoding is described in terms of a long-time averaged signal-to-noise ratio and subjective quality of the processed speech. For voiceband data, long-time averaged signal-to-noise ratios relevant to 4.8 kbit/s eight-phase PSK and 9.6 kbit/s 16-point QAM voiceband data are given.