PREDICTIVE CODING OF SPEECH SIGNALS

• US 3,631,520 A
• Filed: 08/19/1968
• Issued: 12/28/1971
• Est. Priority Date: 08/19/1968
• Status: Expired due to Term

• Current Assignee: Bishnu S. Atal
• Original Assignee: Bishnu S. Atal
• Inventors: Atal, Bishnu S.
• Application Number:
  US04/753,408
• Time in Patent Office: 1,226 Days
• US Class Current: 704/219
• CPC Class Codes: G10L 19/04 using predictive techniques
• CPC Class Codes: H03M 3/022 with adaptable step size, e.g. adaptive delta modulation [ADM]

Abstract

Predictive coding of signals, i.e., the reduction or redundancy in a signal by subtracting from it that part which can be predicted from its past, is a well-known technique for reducing the channel capacity required to transmit a signal with specified fidelity. It has been widely applied to signals, such as television signals which have regularly repeating intervals of information, but has not been satisfactorily applied to signals, such as speech, which exhibit characteristics that vary from speaker to speaker and from time to time for one speaker. According to this invention, an adaptive predictor is employed which is readjusted periodically to match the time-varying characteristics of a speech signal.

Claims

1. Speech signal processing apparatus, which comprises:

means, adjusted in accordance with parameters representative of identifying characteristics of selected pitch periods of an applied speech signal, for predicting the present value of said speech signal on the basis of signals in selected past intervals thereof;

means for coding the differences between the predicted value and the present value of said signal for transmission;

means for analyzing selected pitch periods of said speech signal to develop a plurality of parameter signals which represent vocal tract transmission and source characteristics of said speech signal within said periods; and

means for periodically adjusting said predicting means in accordance with said parameter signals.

2. Speech signal processing apparatus as defined in claim 1, wherein, said characteristics of said speech signal represented by said parameter signals include the extent of selected past pitch periods and the magnitude of signals within said pitch periods.

3. Speech signal processing apparatus as defined in claim 1, wherein new parameter signals are developed every 5 milliseconds.

4. Speech signal processing apparatus as defined in claim 1, wherein said means for predicting the present value of said applied speech signal comprises, a linear predictor characterized by a z-transform given by where b is a factor representative of signal values during consecutive selected signal intervals, K is a number representative of the duration of consecutive pitch periods of said applied signal, am are amplitude factors representative of the short time spectral envelope of said speech signal, and N represents a selected number of said factors am.

5. A communication system for conveying the information content of a speech signal over a channel of relatively small capacity which comprises, in combination:

at a transmitter station;

means for reducing the redundancy in a speech signal by subtracting from it a predicted value of the signal derived from past pitch period intervals thereof selected in response to parameter signals developed from an analysis of selected pitch period intervals, means for analyzing selected pitch periods of said speech signal to develop a plurality of parameter signals which denote selected time varying characteristics of said speech signal within said intervals, means for periodically adjusting said predicting means in accordance with said parameter signals, and means for transmitting both the difference between said predicted value and the present value of a speech signal and said parameter signals to a receiver station, and at said receiver station;

means, adjusted in response to received parameter signals, for predicting the value of said speech signal in response to previously reconstructed speech signals, and means for adding received difference signals to said predicted value signals to develop a replica of said speech signal.

6. A communication system as defined in claim 5 in further combination with, means at said transmitter station for encoding said difference signal and said parameter signals for transmission as a composite signal, and means at said receiver station for decoding said received signals to recover said difference signals and said parameter signals.

7. A communication system as defined in claim 5 wherein said difference signals and said parameter signals are transmitted to said receiver station via diverse transmission facilities.

8. A communication system as defined in claim 5 wherein said parameter signals are scrambled according to a prescribed code for transmission.

9. Apparatus for predicting the present value of a speech signal from its past, which comprises:

means supplied with reconstructed samples of a predictively coded speech signal and with parameter signals which denote, respectively, the values during each of a selected number of consecutive intervals of said speech signal of the duration K of a pitch period of said speech signal, the relative amplitudes b of correlated signals in a number of said selected signal intervals, and amplitude factors am representative of the short time spectral envelope of said speech signal during said selected intervals, for developing signal samples that closely represent the present value of said speech signal; and

means for periodically adjusting the values of b, K, and am in accordance with current speech signal values.

10. Apparatus for developing parameter signals for use in the predictive coding of speech signals, which comprises, in combination:

means for developing from past samples of an applied speech signal a first signal which denotes the duration of a pitch period of said applied speech signal;

means for developing from past samples of said applied signal a second signal which specifies the relative amplitudes of correlated signals in a number of selected consecutive intervals of said applied speech signal;

means for developing from past samples of said applied signal a set of signals which represents the short time spectral envelope of said applied signal during said selected signal intervals; and

means for periodically selecting a number of consecutive intervals of said applied speech signal to represent past samples of said applied speech signal.

11. Apparatus for reconstructing a speech signal from signals representative of the difference between the present value of said speech signal and a predicted value derived from past pitch period intervals thereof, which comprises, means, adjusted in accordance with received parameter signals representative of vocal tract transmission and source characteristics of a speech signal, for predicting the value of said speech signal in response to previously reconstructed speech signals, and MEANS for adding said received difference signal to said predicted value signal to develop a replica of said speech signal.