Voice communication processing system

• US 5,448,680 A
• Filed: 02/12/1992
• Issued: 09/05/1995
• Est. Priority Date: 02/12/1992
• Status: Expired due to Fees

• Current Assignee: the united states of america as represented by the secretary of the navy
• Original Assignee: the united states of america as represented by the secretary of the navy
• Inventors: Kang, George S., Fransen, Lawrence J.
• Primary Examiner(s): Knepper, David D.
• Application Number:
  US07/839,159
• Time in Patent Office: 1,301 Days
• Field of Search: 395/2, 395/2.34, 381/41-43, 381/31, 381/36
• US Class Current: 704/225
• CPC Class Codes: G10L 19/07 Line spectrum pair [LSP] vocoders
• CPC Class Codes: G10L 19/125 Pitch excitation, e.g. pitch synchronous innovation CELP [PSI-CELP]
• CPC Class Codes: G10L 25/12 the extracted parameters being prediction coefficients

Abstract

A voice communication processing system and method for processing a speech waveform as a digital bit stream having a reduced number of bits representing speech parameters. The bit representation of amplitude parameters is reduced by storing only probable amplitude parameter transitions corresponding to amplitude parameter indices in an amplitude table and by joint encoding the amplitude parameter indices over multiple frames. The bit representation of the pitch period is reduced by storing a range of pitch periods in a pitch table and by joint encoding pitch period indices corresponding to an average pitch period over two frames. The bit representation of the vocal tract filter coefficients is reduced by storing only probable filter coefficient transitions corresponding to filter coefficient indices in a filter coefficient table and by joint encoding the filter coefficient indices over two frames. Voicing decisions are inferred by an associated vocal tract filter coefficient index obtained by searching the filter coefficient table where the table is divided according to the voicing decisions, and thus separate voicing decisions do not have to be transmitted. By providing a reduced bit representation of the various speech parameters as explained above, the present invention processes the speech waveform at a more efficient data rate. In addition, the present invention converts prediction coefficients (PCs) into line spectra pairs (LSPs) to be used as filter parameters when performing a linear predictive coder (LPC) analysis. Thus, by using LSPs, the present invention is able to more efficiently encode and decode speech.

Claims

1. A voice communication processing system for processing a speech waveform as a digital bit stream, comprising:

transmitting means for converting the speech waveform into the digital bit stream and transmitting the digital bit stream by encoding speech parameters from the speech waveform into reduced bit representation by joint encoding the speech parameters over frames in the digital bit stream; and

receiving means for receiving the digital bit stream and converting the digital bit stream into a reproduced speech waveform by decoding the reduced bit representation in the digital bit stream into reproduced speech parameters in the reproduced speech waveform;

wherein said transmitting means includes a parameter encoder encoding an amplitude parameter by joint encoding amplitude table indices of the frames in the digital bit stream.

2. A voice communication processing system for processing a speech waveform as a digital bit stream, comprising:

transmitting means for converting the speech waveform into the digital bit stream and transmitting the digital bit stream by encoding speech parameters from the speech waveform into a reduced bit representation by joint encoding the speech parameters over frames in the digital bit stream; and

receiving means for receiving the digital bit stream and converting the digital bit stream into a reproduced speech waveform by decoding the reduced bit representation in the digital bit stream into reproduced speech parameters in the reproduced speech waveform.;

wherein said transmitting means includes a parameter encoder encoding a pitch period by joint encoding pitch table indices being an average of the pitch period over the frames in the digital bit stream.

3. Encoding/decoding system in a voice communication processor converting a speech waveform into a digital bit stream, transmitting and receiving the digital bit stream, and converting the digital bit stream to a reproduced speech waveform, said encoding/decoding system comprising:

encoding means for encoding speech parameters from the speech waveform into a reduced bit representation by joint encoding the speech parameters over frames in the digital bit stream; and

decoding means for decoding the digital bit stream into reproduced speech parameters used for generating the reproduced speech waveform;

wherein said encoding means includes a parameter encoder encoding an amplitude parameter by joint encoding amplitude table indices of the frames in the digital bit stream.

4. Encoding/decoding system in a voice communication processor converting a speech waveform into a digital bit stream, transmitting and receiving the digital bit stream, and converting the digital bit stream to a reproduced speech waveform, said encoding/decoding system comprising:

encoding means for encoding speech parameters from the speech waveform into a reduced bit representation by joint encoding the speech parameters over frames in the digital bit stream; and

decoding means for decoding the digital bit stream into reproduced speech parameters used for generating the reproduced speech waveform;

wherein said encoding means includes a parameter encoder encoding a pitch period by joint encoding pitch table indices being an average of the pitch period over the frames in the digital bit stream.

5. A method of processing a speech waveform as a digital bit stream, comprising the steps of:

a) converting the Speech waveform into the digital bit stream and transmitting the digital bit stream by encoding speech parameters from the speech waveform into a reduced bit representation by joint encoding the speech parameters over frames in the digital bit stream; and

b) receiving the digital bit stream and converting the digital bit stream into a reproduced speech waveform by decoding the digital bit stream into reproduced speech parameters in the reproduced speech waveform;

wherein step a) includes;

a1) obtaining an amplitude parameter from the speech waveform for each of the frames;

a2) performing a look-up operation of an amplitude table to obtain an amplitude table index for each of the frames corresponding to the amplitude parameter; and

a3) joint encoding the amplitude table indices over the frames.

6. A method of processing a speech waveform as a digital bit stream, comprising the steps of:

a) converting the speech waveform into the digital bit stream and transmitting the digital bit stream by encoding speech parameters from the speech waveform into a reduced bit representation by joint encoding the speech parameters over frames in the digital bit stream; and

b) receiving the digital bit stream and converting the digital bit stream into a reproduced speech waveform by decoding the digital bit stream into reproduced speech parameters in the reproduced speech waveform;

wherein step a) includes;

a1) obtaining a pitch period from the speech waveform for each of the frames;

a2) performing a look-up operation of a pitch table to obtain a pitch table index for each of the frames corresponding to an average of the pitch period over the frames, and

a3) joint encoding the pitch table indices over the frames.

7. A voice communication processing system for processing a speech waveform as a digital bit stream, comprising:

transmitting means for converting the speech waveform into the digital bit stream and transmitting the digital bit stream by encoding speech parameters from the speech waveform into a reduced bit representation by joint encoding the speech parameters over frames in the digital bit stream; and

receiving means for receiving the digital bit stream and converting the digital bit stream into a reproduced speech waveform by decoding the reduced bit representation in the digital bit stream into reproduced speech parameters in the reproduced speech waveform;

wherein said transmitting means further comprises;

prediction coefficient generating means for receiving the speech waveform and the generating prediction coefficients responsive to the speech waveform;

coefficient generating means for generating coefficients of real-root removed sum and difference filters responsive to the prediction coefficients using polynomial division and for generating sine and cosine coefficients;

a storage table connected to said transforming means and storing the sine and cosine coefficients as stored sine and cosine coefficients; and

spectrum generating means for generating spectrum coefficients by transforming the coefficients using the stored sine and cosine coefficients and for determining line spectrum pairs for generating the reproduced speech waveform by determining which of the spectrum coefficients have a null frequency using a parabolic fitting.

8. A voice communication processing system according to claim 7, wherein said coefficient generating means decomposes a linear predictive coefficient analysis filter used to represent the speech waveform into sum and difference filters and removes extraneous roots of each of said sum and difference filters to generate the coefficients of the real-root removed sum and difference filters.

9. A voice communication processing system according to claim 7, further comprising a formula register connected to said coefficient generating means, and wherein said coefficient generating means generates coefficient formulas which are stored in said formula register, the coefficients determined by the coefficient formulas.

10. A method of processing a speech waveform as a digital bit stream, comprising the steps of:

a) converting the speech waveform into the digital bit stream and transmitting the digital bit stream by encoding speech parameters from the speech waveform into a reduced bit representation by joint encoding the speech parameters over frames in the digital bit stream; and

b) receiving the digital bit stream and converting the digital bit stream into a reproduced speech waveform by decoding the digital bit stream into reproduced speech parameters in the reproduced speech waveform;

wherein step a) includes

a1) receiving the speech waveform and generating prediction coefficients responsive to the speech waveform;

a2) generating coefficients of real-root removed sum and difference filters responsive to the prediction coefficients using polynomial division and generating sine and cosine coefficients;

a3) storing the sine and cosine coefficients in a storage table as stored sine and cosine coefficients;

a4) generating spectrum coefficients by transforming the coefficients using the stored sine and cosine coefficients; and

a5) determining line spectrum pairs for generating the reproduced speech waveform by determining which of the spectrum coefficients have a null frequency using a parabolic fitting.

11. A method according to claim 10, wherein step a) further includes before said generating step a2) , the steps of:

(1) decomposing a linear predictive coefficient analysis filter used to represent the speech waveform into sum and difference filters; and

(2) removing extraneous roots of each of said sum and difference filters to generate the coefficients of the real-root removed sum and difference filters.

12. A method according to claim 10, wherein step a2) further comprises the step of generating coefficient formulas which are stored in a formula storage table, the coefficients determined by the coefficient formulas.

13. A method for transforming prediction coefficients to line spectrum pairs, comprising the steps of:

a) generating prediction coefficients responsive to a speech waveform;

b) generating coefficients of real-root removed sum and difference filters responsive to the prediction coefficients using polynomial division and generating sine and cosine coefficients;

c) storing the sine and cosine coefficients in a storage table as stored sine and cosine coefficients;

d) generating spectrum coefficients by transforming the coefficients using the stored sine and cosine coefficients; and

e) determining line spectrum pairs for generating a reproduced speech waveform by determining which of the spectrum coefficients have a null frequency using a parabolic fitting.

14. A method according to claim 13, further including before said generating step b), the steps of:

(1) decomposing the linear predictive coefficient analysis filter into sum and difference filters; and

(2) removing extraneous roots of each of said sum and difference filters to generate the coefficients of the real-root removed sum and difference filters.

15. A method according to claim 13, wherein step a) further comprises the step of generating coefficient formulas which are stored in a formula storage table, the coefficients determined by the coefficient formulas.

16. A converter transforming prediction coefficients to line spectrum pairs, comprising:

prediction coefficient generating means for receiving a speech waveform and for generating prediction coefficients responsive to the speech waveform;

coefficient generating means for generating coefficients of real-root removed sum and difference filters responsive to the prediction coefficients using polynomial division and for generating sine and cosine coefficients;

a storage table connected to said transforming means storing the sine and cosine coefficients as stored sine and cosine coefficients; and

spectrum generating means for generating spectrum coefficients by transforming the coefficients using the stored sine and cosine coefficients and for determining line spectrum pairs for generating a reproduced speech waveform by determining which of the spectrum coefficients have a null frequency using a parabolic fitting.

17. A converter according to claim 16, wherein said coefficient generating means decomposes a linear predictive coefficient analysis filter used to represent the speech waveform into sum and difference filters and removes extraneous roots of each of said sum and difference filters to generate the coefficients of the real-root removed sum and difference filters.

18. A converter according to claim 16, further comprising a formula register connected to said coefficient generating means, and wherein said coefficient generating means generates coefficient formulas which are stored in said formula register, the coefficients determined by the coefficient formulas.