Method and apparatus for speech digitizing

Abstract

Abstract Speech digitising is carried out using both signal form coding and source coding with a coder for digitisation and a decoder for reconstruction of the speech signal. The speech signal is divided up into segments in the coder and processed in one part of the segments with an approximation of the sampling values which is as accurate as possible, where an estimated value is calculated for pending sampling values using known sampling values. In the other part of the segments, parameters for speech simulation in terms of the source coding only are derived. The individual signal segments are processed at variable bit rates which are allocated to different operating modes, and each signal segment is categorised as one of the operating modes. The individual speech segments are thus coded according to requirements with a larger or smaller number of bits, providing a hybrid coding method which unifies the source coding and the signal form coding. Along with the signal quantisation of preceding and subsequent signal processing stages, this produces an average bit rate of 6 kbit/s and a speech quality which corresponds totally to that of telephony transmission. <IMAGE>

Inventors

• 👤 Arthur Schaub

Bibliographic Details