Adaptive predictive speech signal coding system

Claim 2

A band compression system for the transmission of speech signals, which comprises: a linear prediction vocoder analyzer supplied with a speech signal for developing a signal replica of the spectral envelope of said speech signal, synthesizer means responsive to said spectral envelope signal for developing a frequency compressed spectral envelope signal, means for transmitting said compressed envelope signal to a receiver, and a linear prediction vocoder synthesizer at said receiver for developing a replica of said speech signal.

Claim 3

A band compression system for the transmission of speech signals, which comprises: a linear prediction vocoder analyzer supplied with a speech signal for developing a signal replica of the spectral envelope of said speech signal, synthesizer means responsive to said spectral envelope signal for developing a frequency compressed spectral envelope signal, means for transmitting said compressed envelope signal to a receiver, and a linear prediction vocoder synthesizer at said receiver for developing a replica of said speech signal.

Claim 4

A band compression system for the transmission of speech signals, which comprises: a linear prediction vocoder analyzer supplied with a speech signal for developing a signal replica of the spectral envelope of said speech signal, synthesizer means responsive to said spectral envelope signal for developing a frequency compressed spectral envelope signal, means for transmitting said compressed envelope signal to a receiver, and a linear prediction vocoder synthesizer at said receiver for developing a replica of said speech signal.

Claim 5

A band compression system for the transmission of speech signals, which comprises: a linear prediction vocoder analyzer supplied with a speech signal for developing a signal representation of the spectral envelope of said speech signal, means for generating a train of low rate pulsive signals, means for shaping the spectral envelope of said pulsive signals in accordance with said spectral envelope signal, means for transmitting said spectral envelope signal to a receiver, means at said receiver for recovering said signal representation of the spectral envelope of said speech signal, and means responsive to said recovered signal representation for synthesizing a replica of said speech signals.

Claim 6

A band compression system for the transmission of speech signals, which comprises: a linear prediction vocoder analyzer supplied with a speech signal for developing a signal representation of the spectral envelope of said speech signal, means for generating a train of low rate pulsive signals, means for shaping the spectral envelope of said pulsive signals in accordance with said spectral envelope signal, means for transmitting said spectral envelope signal to a receiver, means at said receiver for recovering said signal representation of the spectral envelope of said speech signal, and means responsive to said recovered signal representation for synthesizing a replica of said speech signals.

Claim 7

A speech signal coder for developing reduced bandwidth signals for transmission, which comprises: MEANS for sampling an applied speech signal at a rate to preserve the characteristics of said speech signal, means for analyzing said speech samples to develop parameter signals which represent the coefficients of a linear prediction of the present value of said speech signal on the basis of past values of said speech signal, the pitch, and the voicing character thereof, means for generating a train of pulses at a rate lower than the rate of sampling of said speech signal, a linear prediction synthesizer for shaping the spectrum of said pulses in accordance with said prediction coefficient values to produce a spectral envelope signal, and means for combining said spectral envelope signal and said pitch and voicing character parameter signals for transmission to a synthesizer.

Claim 8

A speech signal coder for developing reduced bandwidth signals for transmission, which comprises: MEANS for sampling an applied speech signal at a rate to preserve the characteristics of said speech signal, means for analyzing said speech samples to develop parameter signals which represent the coefficients of a linear prediction of the present value of said speech signal on the basis of past values of said speech signal, the pitch, and the voicing character thereof, means for generating a train of pulses at a rate lower than the rate of sampling of said speech signal, a linear prediction synthesizer for shaping the spectrum of said pulses in accordance with said prediction coefficient values to produce a spectral envelope signal, and means for combining said spectral envelope signal and said pitch and voicing character parameter signals for transmission to a synthesizer.

Claim 9

A speech signal coder as defined in claim 3, wherein, said train of pulses is generated at a rate commensurate with the rate at which said parameter values of said linear predictor are revalued.

Claim 10

A speech signal coder as defined in claim 3, wherein, the spectrum of said pulses is shaped in accordance with said prediction coefficient values to produce spectrally shaped pulses with the same rms value as said speech signal.

Claim 11

A speech signal coder as defined in claim 3, wherein said synthesizer comprises: means for recovering said coefficients from said transmitted spectral envelope signal and for recovering said parameter signals from said transmitted combined signal, means for adjusting a linear predictor in accordance with said recovered coefficients, means for predicting the present value of said received signal on the basis of said selected past values of it, and means for developing a replica of said applied speech signal from said predicted values and from signals generated under control of said pitch and voicing character parameter signals.

Claim 12

In combination with the speech coder as defined in claim 3, a speech decoder, which comprises, means for recovering said spectral envelope signal, said pitch, and said voicing character parameter signals, means for analyzing said spectral envelope signal at said lower rate to develop parameter signals which represent linear prediction coefficients and a signal representative of the rms value of said spectral envelope signal, means responsive to said pitch signal for developing a train of pitch rate pulses, means for generating a train of noise rate pulses, means responsive to said parameter signals for predicting the present value of a reconstituted speech signal on the basis of past values of said reconstituted speech signal, means responsive to said parameter signals, said predicted value signals, said rms signal, and said reconstituted pitch signal for respectively shaping said train of pitch pulses and said train of noise pulses, and means for selectively combining said shaped pulses and said reconstituted speech signal to produce a replica of said applied speech signal.