System for coding speech information using an adaptive codebook with enhanced variable resolution scheme

• US 20020147583A1
• Filed: 02/12/2001
• Published: 10/10/2002
• Est. Priority Date: 09/15/2000
• Status: Active Grant

• Current Assignee: MACOM Technology Solutions Holdings, Inc.
• Original Assignee: Mindspeed Technologies Inc. (MACOM Technology Solutions Holdings, Inc.)
• Inventors: Gao, Yang
• Granted Patent: US 6,760,698 B2
• Time in Patent Office: Days
• US Class Current: 704/223
• CPC Class Codes: G10L 19/08 Determination or coding of the excitation function; Determination or coding of the long-term prediction parameters
• CPC Class Codes: G10L 2019/0011 Long term prediction filters, i.e. pitch estimation
• CPC Class Codes: G10L 21/0364 for improving intelligibility

Abstract

A speech coding system includes an adaptive codebook containing excitation vector data associated with corresponding adaptive codebook indices (e.g., pitch lags). Different excitation vectors in the adaptive codebook have distinct corresponding resolution levels. The resolution levels include a first resolution range of continuously variable or finely variable resolution levels. A gain adjuster scales a selected excitation vector data or preferential excitation vector data from the adaptive codebook. A synthesis filter synthesizes a synthesized speech signal in response to an input of the scaled excitation vector data. The speech coding system may be applied to an encoder, a decoder, or both.

Claims

1. A system for coding a speech signal, the system comprising:

an adaptive codebook containing excitation vector data associated with corresponding adaptive codebook indices, a resolution of the excitation vector data versus values of the adaptive codebook indices varying in accordance with multiple resolution levels, including a first resolution range having generally continuously variable resolution levels within a corresponding first pitch lag range;

a gain adjuster for scaling selected excitation vector data from the adaptive codebook; and

a synthesis filter for synthesizing a synthesized speech signal in response to an input of the scaled excitation vector data.

2. The system according to claim 1 wherein the generally continuously variable resolution levels vary from one another throughout at least a majority of a first pitch lag range.

3. The system according to claim 1 wherein the generally continuously variable resolution levels vary from one another throughout a substantial entirety of the first pitch lag range.

4. The system according to claim 1 further comprising:

a minimizer for minimizing a residual signal formed from a combination of the synthesized speech signal and a reference speech signal, where the system is organized to form an encoder.

5. The system according to claim 1 where the first pitch lag range comprises an intermediate pitch lag range associated with the adaptive codebook indices, the intermediate pitch lag range affiliated with a generally linear segment defining a resolution of the excitation vector data versus corresponding pitch lag values.

6. The system according to claim 5 where the generally linear segment is sloped to provide a higher resolution of the excitation vector data for lower pitch lag values and a lower resolution of the excitation vector data for higher pitch lag values.

7. The system according to claim 1 where the first pitch lag range is bounded by a second pitch lag range of lower pitch lag values than those of the first pitch lag range, the second pitch lag range having at least one resolution level equal to or higher than the generally continuously variable resolution levels of the first pitch lag range.

8. The system according to claim 1 where the first pitch lag range is bounded by a third pitch lag range of higher pitch lag values than those of the second pitch lag range, the third pitch lag range having at least one resolution level equal to or lower than the generally continuously variable resolution levels of the first pitch lag range.

9. The system according to claim 1 where the first pitch lag range is bounded by a second pitch lag range and a third pitch lag range, the second range and the third range having distinct, generally constant resolution levels of the excitation vector data.

10. The system according to claim 1 where adaptive codebook supports multiple ranges of pitch lags, including the first pitch lag range spanning intermediate pitch lag values, a second pitch lag range covering lower pitch lag values and a third pitch lag range covering higher pitch lag values, where the resolution level of excitation vectors affiliated with the second pitch lag range exceeds the resolution level s of excitation vectors affiliated with the third pitch lag range.

11. The system according to claim 1 where the first pitch lag range and the associated first resolution range collectively define a region that contains a generally linear segment of resolution of the excitation vector data versus pitch lag that conforms to the following equation:

R_L=ε

/(y+η

(L<sup>−

1</sup>−

k))

where R_L is the resolution at pitch lag L, L falls within the first resolution range, L<sup>−

1</sup> represents previous pitch lag value with respect to the pitch lag L;

ε

, η

, and y represent constants that are functions of a slope of the pitch lag versus resolution, and k represents a lower-bound value of the first resolution range.

12. The system according to claim 1 where the first pitch lag range and the associated first resolution range collectively define a region that contains a generally linear segment of granularity of the excitation vector data versus pitch lag that conforms to the following equation:

14. The system according to claim 13 wherein the generally continuously variable resolution levels vary from one another throughout at least a majority of a first pitch lag range.

15. The system according to claim 13 wherein the generally continuously variable resolution levels vary from one another throughout a substantial entirety of the first pitch lag range.

16. The system according to claim 13 where the excitation vector data affiliated with the first pitch lag range has a higher resolution for lower pitch lag values and a lower resolution for higher pitch lag values.

17. The system according to claim 13 where the pitch lag values include a first pitch lag range, a second pitch lag range, and a third pitch lag range that collectively extend from a lower pitch lag value to an upper pitch lag value, where the lower pitch lag values is equal to or greater than approximately 15 samples and where the upper pitch lag value is less than or equal to approximately 175 samples of an input speech signal.

18. The system according to claim 13 where the first resolution range is associated with a corresponding first pitch lag range, the first pitch lag range extending from a pitch lag range of approximately 34 to approximately 90 samples of the input signal, a second pitch lag range extending from a pitch lag value range of approximately 17 samples to approximately 33 samples and a third pitch lag range extending from a pitch lag value of approximately 91 samples to approximately 148 samples of the input speech signal.

19. The system according to claim 13 where the pitch lag values include a second pitch lag range associated with a corresponding generally constant resolution of approximately 5.

20. The system according to claim 13 where the pitch lag values include a third pitch lag range associated with a corresponding generally constant resolution of approximately one.

13. An encoder for encoding a speech signal, the encoder comprising:

an adaptive codebook containing excitation vector data associated with corresponding pitch lag values, a resolution of the excitation vector data versus values of the pitch lag values varying in accordance with multiple ranges of resolution levels, including a first resolution range of continuously variable resolution levels of the excitation vector data;

a gain adjuster for scaling selected excitation vector data from the adaptive codebook;

a synthesis filter for synthesizing a synthesized speech signal in response to an input of the scaled excitation vector data; and

a minimizer for minimizing a residual signal formed from a combination of the synthesized speech signal and a reference speech signal.

21. An decoder for decoding a speech signal, the decoder comprising:

an adaptive codebook containing excitation vector data associated with corresponding pitch lag values, a resolution of the excitation vector data versus values of the pitch lag values varying in accordance with multiple ranges of resolution levels, including a first resolution range of continuously variable resolution levels of the excitation vector data;

a gain adjuster for scaling selected excitation vector data from the adaptive codebook; and

a synthesis filter for synthesizing a synthesized speech signal in response to an input of the scaled excitation vector data.

22. The system according to claim 21 wherein the generally continuously variable resolution levels vary from one another throughout at least a majority of a first pitch lag range.

23. The system according to claim 21 wherein the generally continuously variable resolution levels vary from one another throughout a substantial entirety of the first pitch lag range.

24. The system according to claim 21 where the excitation vector data affiliated with the first pitch lag range has a higher resolution for lower pitch lag values and a lower resolution for higher pitch lag values.

26. The method according to claim 25 further comprising:

minimizing a residual signal formed from a combination of the synthesized speech signal and a reference speech signal to select the selected excitation vector from the adaptive codebook.

27. The method according to claim 25 where the establishing step includes establishing the first pitch lag range as an intermediate pitch lag range associated with the adaptive codebook indices.

28. The method according to claim 25 where the establishing step includes establishing the first pitch lag range bounded by a second pitch lag region of one generally constant resolution level and a third region pitch lag region of another generally constant resolution level.

29. The method according to claim 25 where the establishing step includes establishing a generally linear segment of resolution versus pitch lag values in a region defined by the collective combination of the first pitch lag range and the first resolution range.

30. The method according to claim 25 where the first pitch is associated with intermediate pitch lag values and bounds a second pitch lag range associated with higher pitch lag values an d a third pitch lag range associated with lower pitch lag values, where the resolution level of the of the lower pitch lag values in the second pitch lag range exceeds the resolution levels of the higher pitch lag values in the third pitch lag range.

31. The method according to claim 25 where the first pitch lag range and the first resolution collectively define a region containing a generally linear segment of resolution versus pitch lag that conforms to the following equation:

R_L=ε

/(y+η

(L<sup>−

1</sup>−

k))

where R_L is the resolution at pitch lag L, L falls within the first resolution range, L<sup>−

1</sup> represents previous pitch lag value with respect to the pitch lag L;

ε

, η

, and y represent constants that are functions of a slope of the pitch lag versus resolution, and k represents a lower bound value of the first resolution range.

32. The method according to claim 25 where the first pitch lag range and the first resolution collectively define a region containing a generally linear segment of granularity versus pitch lag that conforms to the following equation:

25. A method for coding a speech signal, the coding method comprising the following steps:

establishing an adaptive codebook containing excitation vector data associated with corresponding adaptive codebook indices, a resolution of the excitation vector data versus values of the adaptive codebook indices varying in accordance with multiple resolution levels, including a first resolution range of continuously variable resolution levels associated with a corresponding first pitch lag range;

scaling selected excitation vector data from the adaptive codebook; and

synthesizing a synthesized speech signal in response to an input of the scaled excitation vector data.