Adaptive tilt compensation for synthesized speech

• US 9,401,156 B2
• Filed: 06/27/2008
• Issued: 07/26/2016
• Est. Priority Date: 09/18/1998
• Status: Expired due to Fees

• Current Assignee: Samsung Electronics Co. Ltd.
• Original Assignee: Samsung Electronics Co. Ltd.
• Inventors: Su, Huan-Yu, Gao, Yang
• Primary Examiner(s): Opsasnick, Michael N
• Application Number:
  US12/215,649
• Publication Number:
  US 20080294429A1
• Time in Patent Office: 2,951 Days
• Field of Search: 704219-222
• US Class Current: 1/1
• CPC Class Codes: G10L 19/0204 using subband decomposition
• CPC Class Codes: G10L 19/09 Long term prediction, i.e. removing periodical redundancies, e.g. by using adaptive codebook or pitch predictor
• CPC Class Codes: G10L 19/12 the excitation function being a code excitation, e.g. in code excited linear prediction [CELP] vocoders
• CPC Class Codes: G10L 19/18 Vocoders using multiple modes
• CPC Class Codes: G10L 19/20 using sound class specific coding, hybrid encoders or object based coding
• CPC Class Codes: G10L 2019/0002 Codebook adaptations
• CPC Class Codes: G10L 2019/0016 Codebook for LPC parameters
• CPC Class Codes: G10L 25/90 Pitch determination of speech signals

Abstract

There is provided a method of using an adaptive tilt compensation by a speech decoder. The method comprises receiving a bit stream including a plurality of parameters representative of a speech signal; identifying an adaptive code vector and a fixed code vector using the plurality of parameters; scaling the adaptive code vector and the fixed code vector to generate a scaled adaptive code vector and a scaled fixed code vector; summing the scaled adaptive code vector and the scaled fixed code vector to generate a synthesized output; calculating a first reflection coefficient based on the plurality of parameters representative of the speech signal; multiplying the first reflection coefficient by a factor to generate a tilt factor; and applying the tilt factor to the synthesized output based on an encoding bit rate.

Claims

1. A method of using an adaptive tilt compensation by a speech decoder and generating a speech signal, the method comprising:

receiving a bit stream including a plurality of parameters representative of the speech signal;

identifying an adaptive code vector and a fixed code vector using the plurality of parameters;

scaling the adaptive code vector and the fixed code vector to generate a scaled adaptive code vector and a scaled fixed code vector;

generating a first synthesized output using the scaled adaptive code vector and the scaled fixed code vector;

generating a tilt factor based on the plurality of parameters representative of the speech signal and an encoding rate, wherein the tilt factor is generated by calculating a first reflection coefficient and multiplying the first reflection coefficient by a factor;

applying the tilt factor to the first synthesized output to generate a second synthesized output; and

converting the second synthesized output into the speech signal;

wherein the converting of the second synthesized output into the speech signal comprises applying a synthesis filter to the second synthesized output, and

wherein the generating a tilt factor comprises increasing the tilt factor as the encoding rate decreases while decoding the speech signal.

2. The method of claim 1, wherein the speech decoder includes a first encoding bit rate and a second encoding bit rate, wherein the first encoding bit rate is higher than the second encoding bit rate, and wherein the method further comprises:

determining whether the encoding bit rate is the first encoding bit rate or the second encoding bit rate;

applying the tilt factor to the first synthesized output if the encoding bit rate is the second encoding bit rate; and

deciding to not apply the tilt factor to the first synthesized output if the encoding bit rate is the first encoding bit rate.

3. The method of claim 1, wherein generating the first synthesized output is by summing the scaled adaptive code vector and the scaled fixed code vector.

4. The method of claim 1, wherein the first synthesized output is a synthesized residual.

5. The method of claim 1, wherein the first synthesized output is a weighted synthesized residual.

6. The method of claim 1, wherein the first synthesized output is a signal in a residual domain.

7. The method of claim 1, wherein the first synthesized output is a weighted signal in a residual domain.

8. A speech decoder comprising:

a receiver configured to receive a bit stream including a plurality of parameters representative of a speech signal;

an adaptive codebook; and

a fixed codebook;

wherein the speech decoder is configured to identify an adaptive code vector and a fixed code vector using the plurality of parameters from the adaptive codebook and the fixed codebook, respectively;

wherein the speech decoder is further configured to scale the adaptive code vector and the fixed code vector to generate a scaled adaptive code vector and a scaled fixed code vector;

wherein the speech decoder is further configured to generate a first synthesized output using the scaled adaptive code vector and the scaled fixed code vector;

wherein the speech decoder is further configured to generate a tilt factor based on the plurality of parameters representative of the speech signal and an encoding rate and wherein the tilt factor is generated by calculating a first reflection coefficient and multiplying the first reflection coefficient by a factor;

wherein the speech decoder is further configured to apply the tilt factor to the first synthesized output to generate a second synthesized output;

wherein the speech decoder is further configured to apply a synthesis filter to the second synthesized output; and

wherein the speech decoder generates the tilt factor by increasing the tilt factor as the encoding rate decreases while decoding the speech signal.

9. The speech decoder of claim 8, wherein the speech decoder includes a first encoding bit rate and a second encoding bit rate, wherein the first encoding bit rate is higher than the second encoding bit rate, and wherein the speech decoder is further configured to determine whether the encoding bit rate is the first encoding bit rate or the second encoding bit rate, apply the tilt factor to the first synthesized output if the encoding bit rate is the second encoding bit rate, and decide to not apply the tilt factor to the first synthesized output if the encoding bit rate is the first encoding bit rate.

10. The speech decoder of claim 8, wherein the speech decoder is further configured to generate the first synthesized output by summing the scaled adaptive code vector and the scaled fixed code vector.

11. The speech decoder of claim 8, wherein the first synthesized output is a synthesized residual.

12. The speech decoder of claim 8, wherein the first synthesized output is a weighted synthesized residual.

13. The speech decoder of claim 8, wherein the first synthesized output is a signal in a residual domain.

14. The speech decoder of claim 8, wherein the first synthesized output is a weighted signal in a residual domain.

15. A method of using an adaptive tilt compensation by a multi-rate speech decoder and generating a speech signal, the method comprising:

receiving a bit stream including a plurality of parameters representative of the speech signal;

identifying an adaptive code vector and a fixed code vector using the plurality of parameters;

scaling the adaptive code vector and the fixed code vector to generate a scaled adaptive code vector and a scaled fixed code vector;

generating a first synthesized output using the scaled adaptive code vector and the scaled fixed code vector;

determining an amount based on an encoding bit rate for a tilt factor by calculating a first reflection coefficient based on the plurality of parameters representative of the speech signal and multiplying the first reflection coefficient by a factor;

applying the tilt factor to the first synthesized output to generate a second synthesized output; and

converting the second synthesized output into the speech signal;

wherein the converting of the second synthesized output into the speech signal comprises applying a synthesis filter to the second synthesized output, and

wherein the determining the amount based on the encoding bit rate for the tilt factor increases the tilt factor as the encoding rate decreases while decoding the speech signal.

16. The method of claim 15, wherein the first synthesized output is a synthesized residual.

17. The method of claim 15, wherein the first synthesized output is a weighted synthesized residual.

18. The method of claim 15, wherein the first synthesized output is a signal in a residual domain.

19. The method of claim 15, wherein the first synthesized output is a weighted signal in a residual domain.

20. A multi-rate speech decoder comprising:

a receiver configured to receive a bit stream including a plurality of parameters representative of a speech signal;

an adaptive codebook; and

a fixed codebook;

wherein the multi-rate speech decoder is configured to identify an adaptive code vector and a fixed code vector using the plurality of parameters from the adaptive codebook and the fixed codebook;

wherein the multi-rate speech decoder is further configured to scale the adaptive code vector and the fixed code vector to generate a scaled adaptive code vector and a scaled fixed code vector;

wherein the multi-rate speech decoder is further configured to generate a first synthesized output using the scaled adaptive code vector and the scaled fixed code vector;

wherein the multi-rate speech decoder is further configured to determine an amount based on an encoding bit rate for a tilt factor by calculating a first reflection coefficient based on the plurality of parameters representative of the speech signal and multiplying the first reflection coefficient by a factor;

wherein the multi-rate speech decoder is further configured to apply the tilt factor to the first synthesized output to generate a second synthesized output;

wherein the multi-rate speech decoder is further configured to convert the second synthesized output into the speech signal; and

wherein the multi-rate speech decoder is further configured to convert the second synthesized output into the speech signal by applying a synthesis filter to the second synthesized output; and

wherein the multi-rate speech decoder determines the amount based on the encoding bit rate for the tilt factor by increasing the tilt factor as the encoding rate decreases while decoding the speech signal.

21. The multi-rate speech decoder of claim 20, wherein the first synthesized output is a synthesized residual.

22. The multi-rate speech decoder of claim 20, wherein the first synthesized output is a weighted synthesized residual.

23. The multi-rate speech decoder of claim 20, wherein the first synthesized output is a signal in a residual domain.

24. The multi-rate speech decoder of claim 20, wherein the first synthesized output is a weighted signal in a residual domain.

25. A method of using an adaptive tilt compensation by a multi-rate speech decoder and generating a speech signal, the method comprising:

receiving a bit stream including a plurality of parameters representative of the speech signal, wherein the plurality of parameters include a first parameter and a second parameter;

identifying an adaptive code vector using the first parameter;

identifying a fixed code vector using the second parameter;

scaling the adaptive code vector to generate a scaled adaptive code vector;

scaling the fixed code vector to generate a scaled fixed code vector;

generating a first synthesized output using the scaled adaptive code vector and the scaled fixed code vector;

generating a tilt factor based on a bit rate of the multi-rate speech decoder;

applying the tilt factor to the first synthesized output to generate a second synthesized output; and

converting the second synthesized output into the speech signal;

wherein the converting of the second synthesized output into the speech signal comprises applying a synthesis filter to the second synthesized output, and

wherein the generating the tilt factor comprises increasing the tilt factor as the bit rate decreases while decoding the speech signal.

26. The method of claim 25, wherein the plurality of parameters further include a third parameter and generating the tilt factor comprises:

calculating a first reflection coefficient based on the third parameter; and

multiplying the first reflection coefficient by a factor to generate the tilt factor.

27. The method of claim 26, wherein generating the first synthesized output comprises summing the scaled adaptive code vector and the scaled fixed code vector.

28. The method of claim 26, wherein the first synthesized output is a synthesized residual.

29. The method of claim 26, wherein the first synthesized output is a weighted synthesized residual.

30. The method of claim 26, wherein the first synthesized output is a signal in a residual domain.

31. The method of claim 26, wherein the first synthesized output is a weighted signal in a residual domain.

32. A multi-rate speech decoder comprising:

a receiver configured to receive a bit stream including a plurality of parameters representative of a speech signal, wherein the plurality of parameters include a first parameter and a second parameter;

an adaptive codebook; and

a fixed codebook;

wherein the multi-rate speech decoder is configured to identify an adaptive code vector from the adaptive codebook using the first parameter;

wherein the multi-rate speech decoder is configured to identify a fixed code vector from the fixed codebook using the second parameter;

wherein the multi-rate speech decoder is further configured to scale the adaptive code vector to generate a scaled adaptive code vector;

wherein the multi-rate speech decoder is further configured to scale the fixed code vector to generate a scaled fixed code vector;

wherein the multi-rate speech decoder is further configured to generate a first synthesized output using the scaled adaptive code vector and the scaled fixed code vector;

wherein the multi-rate speech decoder is further configured to generate a tilt factor based on a bit rate of the multi-rate speech decoder;

wherein the multi-rate speech decoder is further configured to apply the tilt factor to the first synthesized output to generate a second synthesized output; and

wherein the multi-rate speech decoder is further configured to convert the second synthesized output into the speech signal by applying a synthesis filter to the second synthesized output, and wherein the multi-rate speech decoder is configured to generate the tilt factor by increasing the tilt factor as the bit rate decreases while decoding the speech signal.

33. The multi-rate speech decoder of claim 32, wherein the plurality of parameters further include a third parameter and wherein the multi-rate speech decoder is configured to generate the tilt factor by:

calculating a first reflection coefficient based on the third parameter; and

multiplying the first reflection coefficient by a factor to generate the tilt factor.

34. The multi-rate speech decoder of claim 33, wherein the first synthesized output is a synthesized residual.

35. The multi-rate speech decoder of claim 33, wherein the first synthesized output is a weighted synthesized residual.

36. The multi-rate speech decoder of claim 33, wherein the first synthesized output is a signal in a residual domain.

37. The multi-rate speech decoder of claim 33, wherein the first synthesized output is'"'"' a weighted signal in a residual domain.

38. The multi-rate speech decoder of claim 33, wherein the multi-rate speech decoder is configured to generate the first synthesized output by summing the scaled adaptive code vector and the scaled fixed code vector.

39. A multi-rate speech decoder comprising:

a receiver configured to receive a bit stream including a plurality of parameters representative of a speech signal, wherein the plurality of parameters include a first parameter and a second parameter;

an adaptive code vector generator configured to generate an adaptive code vector using the first parameter;

a fixed code vector generator configured to generate a fixed code vector using the second parameter;

an adaptive codebook gain configured to scale the adaptive code vector to generate a scaled adaptive code vector;

a fixed codebook gain configured to scale the fixed code vector to generate a scaled fixed code vector;

a first synthesized output generator configured to generate a first synthesized output using the scaled adaptive code vector and the scaled fixed code vector;

a tilt factor generator configured to generate a tilt factor based on a bit rate of the multi-rate speech decoder;

a tilt compensator configured to apply the tilt factor to the first synthesized output to generate a second synthesized output; and

a speech converter configured to apply a synthesis filter to the second synthesized output to generate the speech signal, and

wherein the tilt factor generator is configured to increase the tilt factor as the bit rate decreases while decoding the speech signal.

40. The multi-rate speech decoder of claim 39, wherein the plurality of parameters further include a third parameter and wherein the tilt factor generator comprises:

a first reflection coefficient calculator configured to calculate a first reflection coefficient based on the third parameter; and

a multiplier configured to multiply the first reflection coefficient by a factor to generate the tilt factor.

41. The multi-rate speech decoder of claim 40, wherein the first synthesized output is a synthesized residual.

42. The multi-rate speech decoder of claim 40, wherein the first synthesized output is a weighted synthesized residual.

43. The multi-rate speech decoder of claim 40, wherein the first synthesized output is a signal in a residual domain.

44. The multi-rate speech decoder of claim 40, wherein the first synthesized output is a weighted signal in a residual domain.

45. The multi-rate speech decoder of claim 40, wherein a first synthesized output generator is configured to add the scaled adaptive code vector and the scaled fixed code vector to generate the first synthesized output.

46. A method of using an adaptive tilt compensation by a multi-rate speech decoder and generating a speech signal, the method comprising:

receiving a bit stream including a plurality of parameters representative of the speech signal, wherein the plurality of parameters include at least an adaptive codebook index, an adaptive codebook gain, a fixed codebook index, a fixed codebook gain and synthesis filter parameters;

identifying an adaptive code vector using at least one of the plurality of parameters;

identifying a fixed code vector using at least one of the plurality of parameters;

scaling the adaptive code vector to generate a scaled adaptive code vector;

scaling the fixed code vector to generate a scaled fixed code vector;

generating a first synthesized output using the scaled adaptive code vector and the scaled fixed code vector;

generating a tilt factor based on a bit rate of the multi-rate speech decoder;

generating a second synthesized output from the first synthesized output by applying the tilt factor; and

converting the second synthesized output into the speech signal;

wherein the converting of the second synthesized output into the speech signal comprises applying a synthesis filter to the second synthesized output, and

wherein the generating the tilt factor comprises increasing tilt factor as the bit rate decreases while decoding the speech signal.

47. The method of claim 46, wherein the generating the tilt factor comprises:

calculating a first reflection coefficient based on at least one of the plurality of parameters; and

multiplying the first reflection coefficient by a factor to generate the tilt factor.

48. The method of claim 47, wherein generating the first synthesized output comprises summing the scaled adaptive code vector and the scaled fixed code vector.

49. The method of claim 47, wherein the first synthesized output is a synthesized residual.

50. The method of claim 47, wherein the first synthesized output is a weighted synthesized residual.

51. The method of claim 47, wherein the first synthesized output is a signal in a residual domain.

52. The method of claim 47, wherein the first synthesized output is a weighted signal in a residual domain.

53. A multi-rate speech decoder comprising:

a receiver configured to receive a bit stream including a plurality of parameters representative of the speech signal, wherein the plurality of parameters include at least an adaptive codebook index, an adaptive codebook gain, a fixed codebook index, a fixed codebook gain and synthesis filter parameters;

an adaptive codebook; and

a fixed codebook;

wherein the multi-rate speech decoder is configured to identify an adaptive code vector from the adaptive codebook using at least one of the plurality of parameters;

wherein the multi-rate speech decoder is configured to identify a fixed code vector from the fixed codebook using at least one of the plurality of parameters;

wherein the multi-rate speech decoder is further configured to scale the adaptive code vector to generate a scaled adaptive code vector;

wherein the multi-rate speech decoder is further configured to scale the fixed code vector to generate a scaled fixed code vector;

wherein the multi-rate speech decoder is further configured to generate a first synthesized output using the scaled adaptive code vector and the scaled fixed code vector;

wherein the multi-rate speech decoder is further configured to generate a tilt factor based on a bit rate of the multi-rate speech decoder;

wherein the multi-rate speech decoder is further configured to generate a second synthesized output from the first synthesized output by applying the tilt factor;

wherein the multi-rate speech decoder is further configured to convert the second synthesized output into the speech signal by applying a synthesis filter to the second synthesized output; and

wherein the multi-rate speech decoder is further configured to increase the tilt factor as the bit rate decreases while decoding the speech signal.

54. The multi-rate speech decoder of claim 53, wherein the multi-rate speech decoder is configured to generate the tilt factor by:

calculating a first reflection coefficient based on at least one of the plurality of parameters; and

multiplying the first reflection coefficient by a factor to generate the tilt factor.

55. The multi-rate speech decoder of claim 54, wherein the first synthesized output is a synthesized residual.

56. The multi-rate speech decoder of claim 54, wherein the first synthesized output is a weighted synthesized residual.

57. The multi-rate speech decoder of claim 54, wherein the first synthesized output is a signal in a residual domain.

58. The multi-rate speech decoder of claim 54, wherein the first synthesized output is a weighted signal in a residual domain.

59. The multi-rate speech decoder of claim 54, wherein the multi-rate speech decoder is configured to generate the first synthesized output by summing the scaled adaptive code vector and the scaled fixed code vector.