Audio encoding apparatus and audio decoding apparatus for encoding in multiple stages a multi-pulse signal

• US 6,192,334 B1
• Filed: 04/01/1998
• Issued: 02/20/2001
• Est. Priority Date: 04/04/1997
• Status: Expired due to Term

• Current Assignee: NEC Corporation
• Original Assignee: NEC Corporation
• Inventors: Nomura, Toshiyuki
• Primary Examiner(s): Hudspeth, David R.
• Assistant Examiner(s): Chawan, Vijay B
• Application Number:
  US09/053,606
• Time in Patent Office: 1,056 Days
• Field of Search: 704/207, 704/222, 704/219, 704/231, 704/226, 704/220, 704/200
• US Class Current: 704/219
• CPC Class Codes: G10L 19/10 the excitation function being a multipulse excitation
• CPC Class Codes: G10L 19/107 Sparse pulse excitation, e.g. by using algebraic codebook
• CPC Class Codes: G10L 19/18 Vocoders using multiple modes

Abstract

Auxiliary multi-pulse setting circuit 130 set candidates of pulse positions so that the pulse positions to which no pulse is located are selected in auxiliary multi-pulse searching circuit 131 prior to the pulse positions at which pulses have already been encoded in multi-pulse searching circuit 110. Auxiliary multi-pulse searching circuit 131 generates an auxiliary multi-pulse signal according to the candidates of pulse positions set in auxiliary multi-pulse setting circuit 130 and encodes the auxiliary multi-pulse signal so that difference between the reproduced audio signal which is obtained by driving a linear predictive synthesis filter with the auxiliary multi-pulse signal and an input audio signal is minimized similarly to multi-pulse searching circuit 110.

Claims

1. An audio encoding apparatus for encoding in multiple stages a multi-pulse signal representing an excitation signal of a reproduced audio signal by plural pulses so that a difference between said reproduced audio signal and an input audio signal is minimized, said reproduced audio signal being obtained by driving a linear predictive synthesis filter with said excitation signal which corresponds to said multi-pulse signal, said audio encoding apparatus comprising:

a multi-pulse setting circuit, between said stages which sets pulse position candidates so that position candidates in which no pulse is located in a preceding stage or stages are selected prior to position candidates in which pulses have been already encoded in preceding stages; and

a multi-pulse searching circuit, in each of said multiple stages, which encodes said multi-pulse signal by searching a pulse or pulses among said position candidates set by said multi-pulse setting circuit.

2. An audio encoding apparatus for encoding in multiple stages a multi-pulse signal representing an excitation signal of a reproduced audio signal by plural pulses so that a difference between said reproduced audio signal and an input audio signal is minimized, said reproduced audio signal being obtained by driving a linear predictive synthesis filter with said excitation signal, which comprises between said stages a multi-pulse setting circuit which sets pulse positions so that positions in which no pulse is located are selected prior to positions in which pulses have been already encoded in preceding stages,

wherein each of said multiple stages encodes pulses of said multi-pulsc signal which are in the positions set by said multi-pulse setting circuit;

and

wherein said multi-pulse setting circuit divides each sub-frame into plural sub-areas, selects a limited number of said sub-areas from the top of the ascending order of the number of pulses already encoded therein, and outputs the indices of the selected sub-areas to a next stage.

3. The audio encoding apparatus as set forth in claim 2, wherein each of said multiple stages encodes pulses of said multi-pulse signal only in said sub-areas corresponding to said indices from said multi-pulse setting circuit.

4. An audio decoding apparatus for reproducing an audio signal by driving a linear predictive synthesis filter with an excitation signal, coefficients of said linear predictive synthesis filter being reproduced from data encoded apparatus in an encoding apparatus, said excitation signal being represented by a multi-pulse signal reproduced in multiple stages from data encoded in corresponding multiple stages in said encoding apparatus, said audio decoding apparatus comprising:

a multi-pulse setting circuit, between said stages, which sets pulse position candidates so that position candidates in which no pulse is located in a preceding stage or stages are selected prior to position candidates in which pulses have been already decoded in preceding stages; and

a multi-pulse decoding circuit, in each of said multiple stages, which decodes said multi-pulse signal based on said pulse position candidates set by said multi-pulse setting circuit.

5. An audio decoding apparatus for reproducing an audio signal by driving a linear predictive synthesis filter with an excitation signal, coefficients of said linear predictive synthesis filter being reproduced from data encoded in an encoding apparatus, said excitation signal being represented by plural pulses reproduced in multiple stages from data encoded in corresponding multiple stages in said encoding apparatus, which comprises between said stages a multi-pulse setting circuit which sets pulse positions so that positions in which no pulse is located are selected prior to positions in which pulses have been already decoded in preceding stages, wherein each of said multiple stages decodes pulses of said multi-pulse signal which are in the pulse positions set by said multi-pulse setting circuit;

wherein said multi-pulse setting circuit divides each sub-frame into plural sub-areas, selects a limited number of said sub-areas from the top of the ascending order of the number of pulses already encoded therein, and outputs the indices of the selected sub-areas to a next stage.

6. The audio decoding apparatus as set forth in claim 5, wherein each of said multiple stages decodes pulses of said multi-pulse signal only in said sub-areas corresponding to said indices from said multi-pulse setting circuit.

7. An audio encoding apparatus for encoding in multiple stages an excitation signal of an audio signal by selecting pulse positions of a multi-pulse signal which minimize distortion between an input audio signal and a reproduced audio signal, said excitation signal being expressed by said multi-pulse signal consisting of a plurality of pulses, said reproduced audio signal being obtained by exciting a linear predictive synthesis filter by said excitation signal, said apparatus comprising:

main means for searching for a multi-pulse, wherein said main means encodes positions of pulses of said multi-pulse signal in a first stage by using said input audio signal on the basis of predetermined first pulse-position-candidate information; and

at least one auxiliary means for searching for a multi-pulse;

wherein said auxiliary means for searching for a multi-pulse comprises;

an auxiliary multi-pulse setting circuit which sets second pulse-position-candidate information which will be used in a self-stage, on the basis of said multi-pulse signal which has been set in a preceding stage or stages; and

an auxiliary multi-pulse searching circuit which encodes pulse positions of said multi-pulse signal in said self-stage by using said input audio signal on the basis of said second pulse-position-candidate information.

8. The audio encoding apparatus as set forth in claim 7, wherein said auxiliary multi-pulse setting circuit sets said second pulse-position-candidate information which gives priority to positions where no pulse has been located rather than positions which already have been encoded in said preceding stage or stages.

9. An audio decoding apparatus for decoding, from encoded data, an excitation signal which has been encoded into an expression by a multi-pulse signal consisting of a plurality of pulses in multiple stages;

decoding linear predictor coefficients from said encoded data;

exciting a linear predictive synthesis filter having said linear predictor coefficients by said excitation signal, and thereby reproducing a reproduction of an audio signal, said apparatus comprising;

main means for creating a reproduced signal, wherein said main means creates a reproduced signal of a first stage from an excitation signal of said first stage and said linear predictor coefficients, said excitation signal of said first stage being reproduced from predetermined first pulse-position-candidate information; and

at least one auxiliary means for creating a reproduced signal;

wherein said auxiliary means for creating a reproduced signal comprises;

an auxiliary multi-pulse setting circuit which sets second pulse-position-candidate information which will be used in a self-stage, on the basis of an excitation signal which has been decoded in a preceding stage or stages; and

an auxiliary multi-pulse decoding circuit which decodes an excitation signal of said self-stage on the basis of said second pulse-position-candidate information; and

wherein said auxiliary means for creating a reproduced signal creates an auxiliary reproduced signal by using said excitation signal of said self-stage and said linear predictor coefficients.

10. The audio decoding apparatus as set forth in claim 9, wherein said auxiliary multi-pulse setting circuit sets said second pulse-position-candidate information which gives priority to positions where no pulse has been located rather than positions which already have been set by decoding in said preceding stage or stages.

11. An audio encoding/decoding apparatus comprising:

an audio encoding apparatus for encoding in multiple stages an excitation signal of an audio signal by selecting pulse positions of a multi-pulse signal which minimize distortion between an input audio signal and a reproduced audio signal, said excitation signal being expressed by said multi-pulse signal consisting of a plurality of pulses, said reproduced audio signal being obtained by exciting a linear predictive synthesis filter by said excitation signal, said apparatus comprising;

main means for searching for a multi-pulse, wherein said main means encodes positions of pulses of said multi-pulse signal in a first stage by using said input audio signal on the basis of predetermined first pulse-position-candidate information; and

at least one auxiliary means for searching for a multi-pulse;

wherein said auxiliary means for searching for a multi-pulse comprises;

an auxiliary multi-pulse setting circuit which sets second pulse-position-candidate information which will be used in a self-stage, on the basis of said multi-pulse signal which has been set in a preceding stage or stages; and

an auxiliary multi-pulse encoding circuit which encodes pulse positions of said multi-pulse signal in said self-stage by using said input audio signal on the basis of said second pulse-position-candidate information; and

an audio decoding apparatus for decoding, from encoded data, an excitation signal which has been encoded into an expression by a multi-pulse signal consisting of a plurality of pulses in multiple stages;

decoding linear predictor coefficients from said encoded data;

exciting a linear predictive synthesis filter having said linear predictor coefficients by said excitation signal, and thereby reproducing a reproduction of an audio signal, said apparatus comprising;

main means for creating a reproduced signal, wherein said main means creates a reproduced signal of a first stage from an excitation signal of said first stage and said linear predictor coefficients, said excitation signal of said first stage being reproduced from predetermined first pulse-position-candidate information; and

at least one auxiliary means for creating a reproduced signal;

wherein said auxiliary means for creating a reproduced signal comprises;

an auxiliary multi-pulse setting circuit which sets second pulse-position-candidate information which will be used in a self-stage, on the basis of an excitation signal which has been decoded in a preceding stage or stages; and

an auxiliary multi-pulse decoding circuit which decodes an excitation signal of said self-stage on the basis of said second pulse-position-candidate information; and

wherein said auxiliary means for creating a reproduced signal creates an auxiliary reproduced signal by using said excitation signal of said self-stage and said linear predictor coefficients.

12. The audio encoding/decoding apparatus as set forth in claim 11, wherein said auxiliary multi-pulse setting circuit sets said second pulse-position-candidate information which gives priority to positions where no pulse has been located rather than positions which already have been encoded in said preceding stage or stages.

13. The audio encoding/decoding apparatus as set forth in claim 11, wherein said auxiliary multi-pulse setting circuit sets said second pulse-position-candidate information which gives priority to positions where no pulse has been located rather than positions which already have been set by decoding in said preceding stage or stages.

14. The audio encoding/decoding apparatus as set forth in claim 13, wherein said auxiliary multi-pulse setting circuit sets said second pulse-position-candidate information which gives priority to positions where no pulse has been located rather than positions which already have been encoded in said preceding stage or stages.

15. An audio encoding method for encoding in multiple stages an excitation signal of an audio signal by selecting pulse positions of a multi-pulse signal which minimize distortion between an input audio signal and a reproduced audio signal, said excitation signal being expressed by said multi-pulse signal consisting of a plurality of pulses, said reproduced audio signal being obtained by exciting a linear predictive synthesis filter by said excitation signal, said method comprising:

a main step of searching for a multi-pulse, wherein said main step encodes positions of pulses of said multi-pulse signal in a first stage by using said input audio signal on the basis of predetermined first pulse-position-candidate information; and

at least one auxiliary step of searching for a multi-pulse, wherein said auxiliary step comprises;

an auxiliary multi-pulse setting step which sets second pulse-position-candidate information which will be used in a self-stage, on the basis of said multi-pulse signal which has been set in a preceding stage or stages; and

an auxiliary multi-pulse searching step which encodes pulse positions of said multi-pulse signal in said self-stage by using said input audio signal on the basis of said second pulse-position-candidate information.

16. The audio encoding method as set forth in claim 15, wherein said auxiliary multi-pulse setting step sets said second pulse-position-candidate information which gives priority to positions where no pulse has been located rather than positions which already have been encoded in said preceding stage or stages.

17. An audio decoding method for decoding, from encoded data, an excitation signal which has been encoded into an expression by a multi-pulse signal consisting of a plurality of pulses in multiple stages;

decoding linear predictor coefficients from said encoded data;

exciting a linear predictive synthesis filter having said linear predictor coefficients by said excitation signal, and thereby reproducing a reproduction of an audio signal, said method comprising;

a main step of creating a reproduced signal, wherein said main step creates a reproduced signal of a first stage from an excitation signal of said first stage and said linear predictor coefficients, said excitation signal of said first stage being reproduced from predetermined first pulse-position-candidate information; and

at least one auxiliary step of creating a reproduced signal, wherein said auxiliary step comprises;

an auxiliary multi-pulse setting step which sets second pulse-position-candidate information which will be used in a self-stage, on the basis of an excitation signal which has been decoded in a preceding stage or stages; and

an auxiliary multi-pulse decoding step which decodes an excitation signal of said self-stage on the basis of said second pulse-position-candidate information; and

wherein said auxiliary step of creating a reproduced signal creates an auxiliary reproduced signal by using said excitation signal of said self-stage and said linear predictor coefficients.

18. The audio decoding method as set forth in claim 17, wherein said auxiliary multi-pulse setting step sets said second pulse-position-candidate information which gives priority to positions where no pulse has been located rather than positions which already have been set by decoding in said preceding stage or stages.

19. An audio encoding/decoding method comprising the steps of:

an audio encoding method for encoding in multiple stages an excitation signal of an audio signal by selecting pulse positions of a multi-pulse signal which minimize distortion between an input audio signal and a reproduced audio signal, said excitation signal being expressed by said multi-pulse signal consisting of a plurality of pulses, said reproduced audio signal being obtained by exciting a linear predictive synthesis filter by said excitation signal, said method comprising;

a main step of searching for a multi-pulse, wherein said main step encodes positions of pulses of said multi-pulse signal in a first stage by using said input audio signal on the basis of predetermined first pulse-position-candidate information; and

at least one auxiliary step of searching for a multi-pulse, wherein said auxiliary step comprises;

an auxiliary multi-pulse setting step which sets second pulse-position-candidate information which will be used in a self-stage, on the basis of said multi-pulse signal which has been set in a preceding stage or stages; and

an auxiliary multi-pulse encoding step which encodes pulse positions of said multi-pulse signal in said self-stage by using said input audio signal on the basis of said second pulse-position-candidate information; and

an audio decoding method for decoding, from encoded data, an excitation signal which has been encoded into an expression by a multi-pulse signal consisting of a plurality of pulses in multiple stages;

decoding linear predictor coefficients from said encoded data;

exciting a linear predictive synthesis filter having said linear predictor coefficients by said excitation signal, and thereby reproducing a reproduction of an audio signal, said method comprising;

a main step of creating a reproduced signal, wherein said main step creates a reproduced signal of a first stage from an excitation signal of said first stage and said linear predictor coefficients, said excitation signal of said first stage being reproduced from predetermined first pulse-position-candidate information; and

at least one auxiliary step of creating a reproduced signal, wherein said auxiliary step comprises;

an auxiliary multi-pulse setting step which sets second pulse-position-candidate information which will be used in a self-stage, on the basis of an excitation signal which has been decoded in a preceding stage or stages; and

an auxiliary multi-pulse decoding step which decodes an excitation signal of said self-stage on the basis of said second pulse-position-candidate information; and

wherein said auxiliary step of creating a reproduced signal creates an auxiliary reproduced signal by using said excitation signal of said self-stage and said linear predictor coefficients.

20. The audio encoding/decoding method as set forth in claim 19, wherein said auxiliary multi-pulse setting step sets said second pulse-position-candidate information which gives priority to positions where no pulse has been located rather than positions which already have been encoded in said preceding stage or stages.

21. The audio encoding/decoding method as set forth in claim 19, wherein said auxiliary multi-pulse setting step sets said second pulse-position-candidate information which gives priority to positions where no pulse has been located rather than positions which already have been set by decoding in said preceding stage or stages.

22. The audio encoding/decoding method as set forth in claim 21, wherein said auxiliary multi-pulse setting step sets said second pulse-position-candidate information which gives priority to positions where no pulse has been located rather than positions which already have been encoded in said preceding stage or stages.

23. An audio encoding apparatus for encoding in multiple stages an excitation signal of an audio signal by selecting pulse positions of a multi-pulse signal which minimize distortion between an input audio signal and a reproduced audio signal, said excitation signal being expressed by said multi-pulse signal consisting of a plurality of pulses, said reproduced audio signal being obtained by exciting a linear predictive synthesis filter by said excitation signal, said apparatus comprising:

at least one auxiliary means for searching for a multi-pulse, wherein said auxiliary means encodes pulse positions of a multi-pulse signal of a self-stage from said input audio signal on the basis of pulse-position-candidate information which gives priority to pulse positions where no pulse has been located rather than pulse positions which already have been encoded in a preceding stage or stages.

24. The audio encoding apparatus as set forth in claim 23, wherein said auxiliary means for searching for a multi-pulse comprises:

an auxiliary multi-pulse setting circuit which sets said pulse-position-candidate information which gives priority to pulse positions where no pulse has been located rather than pulse positions which already have been encoded in said preceding stage or stages; and

an auxiliary multi-pulse encoding circuit which encodes pulse positions of said multi-pulse signal in said self-stage from said input audio signal on the basis of said pulse-position-candidate information set in said auxiliary multi-pulse setting circuit.

25. An audio decoding apparatus for decoding, from encoded data, an excitation signal which has been encoded into an expression by a multi-pulse signal consisting of a plurality of pulses in multiple stages;

decoding linear predictor coefficients from said encoded data;

exciting a linear predictive synthesis filter having said linear predictor coefficients by said excitation signal, and thereby reproducing a reproduction of an audio signal, said apparatus comprising;

at least one auxiliary means for creating a reproduced signal, wherein said auxiliary means decodes an excitation signal of a self-stage on the basis of pulse-position-candidate information which gives priority to pulse positions where no pulse has been located rather than pulse positions which already have been set by decoding in a preceding stage or stages, and creates an auxiliary reproduced signal by using said excitation signal of said self-stage and said linear predictor coefficients.

26. The audio decoding apparatus as set forth in claim 25, wherein said auxiliary means for searching for a multi-pulse comprises:

an auxiliary multi-pulse setting circuit which sets said pulse-position-candidate information which gives priority to pulse positions where no pulse has been located rather than pulse positions which already have been set by decoding in said preceding stage or stages; and

an auxiliary multi-pulse decoding circuit which decodes an excitation signal in said self-stage on the basis of said pulse-position-candidate information set in said auxiliary multi-pulse setting circuit.

27. An audio encoding/decoding apparatus comprising:

an audio encoding apparatus for encoding in multiple stages an excitation signal of an audio signal by selecting pulse positions of a multi-pulse signal which minimize distortion between an input audio signal and a reproduced audio signal, said excitation signal being expressed by said multi-pulse signal consisting of a plurality of pulses, said reproduced audio signal being obtained by exciting a linear predictive synthesis filter by said excitation signal, said apparatus comprising;

at least one auxiliary means for searching for a multi-pulse, wherein said auxiliary means encodes pulse positions of a multi-pulse signal of a self-stage from said input audio signal on the basis of pulse-position-candidate information which gives priority to pulse positions where no pulse has been located rather than pulse positions which already have been encoded in a preceding stage or stages; and

an audio decoding apparatus for decoding, from encoded data, an excitation signal which has been encoded into an expression by a multi-pulse signal consisting of a plurality of pulses in multiple stages;

decoding linear predictor coefficients from said encoded data;

exciting a linear predictive synthesis filter having said linear predictor coefficients by said excitation signal, and thereby reproducing a reproduction of an audio signal, said apparatus comprising;

at least one auxiliary means for creating a reproduced signal, wherein said auxiliary means decodes an excitation signal of a self-stage on the basis of pulse-position-candidate information which gives priority to pulse positions where no pulse has been located rather than pulse positions which already have been set by decoding in a preceding stage or stages, and creates an auxiliary reproduced signal by using said excitation signal of said self-stage and said linear predictor coefficients.

28. The audio encoding/decoding apparatus as set forth in claim 27, wherein said auxiliary means for searching for a multi-pulse comprises:

an auxiliary multi-pulse setting circuit which sets said pulse-position-candidate information which gives priority to pulse positions where no pulse has been located rather than pulse positions which already have been encoded in said preceding stage or stages; and

an auxiliary multi-pulse encoding circuit which encodes pulse positions of said multi-pulse signal in said self-stage from said input audio signal on the basis of said pulse-position-candidate information set in said auxiliary multi-pulse setting circuit.

29. The audio encoding/decoding apparatus as set forth in claim 27, wherein said auxiliary means for searching for a multi-pulse comprises:

an auxiliary multi-pulse setting circuit which sets said pulse-position-candidate information which gives priority to pulse positions where no pulse has been located rather than pulse positions which already have been set by decoding in said preceding stage or stages; and

an auxiliary multi-pulse decoding circuit which decodes an excitation signal in said self-stage on the basis of said pulse-position-candidate information set in said auxiliary multi-pulse setting circuit.

30. The audio encoding/decoding apparatus as set forth in claim 29, wherein said auxiliary means for searching for a multi-pulse comprises:

an auxiliary multi-pulse setting circuit which sets said pulse-position-candidate information which gives priority to pulse positions where no pulse has been located rather than pulse positions which already have been encoded in said preceding stage or stages; and

an auxiliary multi-pulse encoding circuit which encodes pulse positions of said multi-pulse signal in said self-stage from said input audio signal on the basis of said pulse-position-candidate information set in said auxiliary multi-pulse setting circuit.

31. An audio encoding method for encoding in multiple stages an excitation signal of an audio signal by selecting pulse positions of a multi-pulse signal which minimize distortion between an input audio signal and a reproduced audio signal, said excitation signal being expressed by said multi-pulse signal consisting of a plurality of pulses, said reproduced audio signal being obtained by exciting a linear predictive synthesis filter by said excitation signal, said method comprising:

a first step of setting pulse-position-candidate information which gives priority to pulse positions where no pulse has been located rather than pulse positions which already have been encoded in a preceding stage or stages; and

a second step of encoding pulse positions of said multi-pulse signal in a self-stage by using said input audio signal on the basis of said pulse-position-candidate information set at said first step.

32. An audio decoding method for decoding, from encoded data, an excitation signal which has been encoded into an expression by a multi-pulse signal consisting of a plurality of pulses in multiple stages;

decoding linear predictor coefficients from said encoded data;

exciting a linear predictive synthesis filter having said linear predictor coefficients by said excitation signal, and thereby reproducing a reproduction of an audio signal, said method comprising;

a first step of decoding an excitation signal of a self-stage on the basis of pulse-position-candidate information which gives priority to pulse locations where no pulse has been located rather than pulse positions which have been set in decoding in a preceding stage or stages; and

a second step of creating an auxiliary reproduced signal by using said excitation signal of said self-stage reproduced at said first step and said linear predictor coefficients.

33. The audio decoding method as set forth in claim 32, wherein said first step comprises:

an auxiliary multi-pulse setting step of setting said pulse-position-candidate information which gives priority to pulse positions where no pulse has been located rather than pulse positions which already have been set in decoding in said preceding stage or stages; and

an auxiliary multi-pulse decoding step of decoding said excitation signal of said self-stage on the basis of said pulse-position-candidate information which has been set in said auxiliary multi-pulse setting step.

34. An audio encoding/decoding method comprising an audio encoding method for encoding in multiple stages an excitation signal of an audio signal by selecting pulse positions of a multi-pulse signal which minimize distortion between an input audio signal and a reproduced audio signal, said excitation signal being expressed by said multi-pulse signal consisting of a plurality of pulses, said reproduced audio signal being obtained by exciting a linear predictive synthesis filter by said excitation signal, said method comprising:

a first step of setting pulse-position-candidate information which gives priority to pulse positions where no pulse has been located rather than pulse positions which already have been encoded in a preceding stage or stages; and

a second step of encoding pulse positions of said multi-pulse signal in a self-stage by using said input audio signal on the basis of said pulse-position-candidate information set at said first step; and

an audio decoding method for decoding, from encoded data, an excitation signal which has been encoded into an expression by a multi-pulse signal consisting of a plurality of pulses in multiple stages;

decoding linear predictor coefficients from said encoded data;

exciting a linear predictive synthesis filter having said linear predictor coefficients by said excitation signal, and thereby reproducing a reproduction of an audio signal, said method comprising;

a first step of decoding an excitation signal of a self-stage on the basis of pulse-position-candidate information which gives priority to pulse locations where no pulse has been located rather than pulse positions which have been set in decoding in a preceding stage or stages; and

a second step of creating an auxiliary reproduced signal by using said excitation signal of said self-stage reproduced at said first step and said linear predictor coefficients.

35. The audio encoding/decoding method as set forth in claim 34, wherein said first step comprises:

an auxiliary multi-pulse setting step of setting said pulse-position-candidate information which gives priority to pulse positions where no pulse has been located rather than pulse positions which already have been set in decoding in said preceding stage or stages; and

an auxiliary multi-pulse decoding step of decoding said excitation signal of said self-stage on the basis of said pulse-position-candidate information which has been set in said auxiliary multi-pulse setting step.

36. A recordable medium on which a computer program is recorded, said computer program comprising instructions for executing an audio encoding method for encoding in multiple stages an excitation signal of an audio signal, said excitation signal being expressed by a multi-pulse signal consisting of a plurality of pulses, by selecting pulse positions of said multi-pulse signal which minimize distortion between an input audio signal and a reproduced audio signal, said reproduced audio signal being obtained by exciting a linear predictive synthesis filter by said excitation signal, said method comprising:

a first step of setting pulse-position-candidate information which gives priority to pulse positions where no pulse has been located rather than pulse positions which already have been encoded in a preceding stage or stages; and

a second step of encoding pulse positions of said multi-pulse signal in a self-stage by using said input audio signal on the basis of said pulse-position-candidate information set at said first step.

37. A recordable medium on which a computer program is recorded, said computer program comprising instructions for executing an audio decoding method for decoding, from encoded data, an excitation signal which has been encoded into an expression by a multi-pulse signal consisting of a plurality of pulses in multiple stages;

decoding linear predictor coefficients from said encoded data;

exciting a linear predictive synthesis filter having said linear predictor coefficients by said excitation signal, and thereby reproducing a reproduction of an audio signal, said method comprising;

a first step of decoding an excitation signal of a self-stage on the basis of pulse-position-candidate information which gives priority to pulse locations where no pulse has been located rather than pulse positions which have been set in decoding in a preceding stage or stages; and

a second step of creating an auxiliary reproduced signal by using said excitation signal of said self-stage reproduced at said first step and said linear predictor coefficients.

38. The recordable medium as set forth in claim 37, wherein said first step comprises:

an auxiliary multi-pulse setting step of setting said pulse-position-candidate information which gives priority to pulse positions where no pulse has been located rather than pulse positions which already have been set in decoding in said preceding stage or stages; and

an auxiliary multi-pulse decoding step of decoding said excitation signal of said self-stage on he basis of said pulse-position-candidate information which has been set in said auxiliary multi-pulse setting step.

39. A recordable medium on which a computer program is recorded, said computer program comprising instructions for executing an audio encoding method for encoding in multiple stages an excitation signal of an audio signal, said excitation signal being expressed by a multi-pulse signal consisting of a plurality of pulses, by selecting pulse positions of said multi-pulse signal which minimize distortion between an input audio signal and a reproduced audio signal, said reproduced audio signal being obtained by exciting a linear predictive synthesis filter by said excitation signal, said method comprising:

a first step of setting pulse-position-candidate information which gives priority to pulse positions where no pulse has been located rather than pulse positions which already have been encoded in a preceding stage or stages; and

a second step of encoding pulse positions of said multi-pulse signal in a self-stage by using said input audio signal on the basis of said pulse-position-candidate information set at said first step; and

said computer program comprising instructions for executing an audio decoding method for decoding, from encoded data, an excitation signal which has been encoded into an expression by a multi-pulse signal consisting of a plurality of pulses in multiple stages;

decoding linear predictor coefficients from said encoded data;

exciting a linear predictive synthesis filter having said linear predictor coefficients by said excitation signal, and thereby reproducing a reproduction of an audio signal, said method comprising;

a first step of decoding an excitation signal of a self-stage on the basis of pulse-position-candidate information which gives priority to pulse locations where no pulse has been located rather than pulse positions which have been set in decoding in a preceding stage or stages; and

a second step of creating an auxiliary reproduced signal by using said excitation signal of said self-stage reproduced at said first step and said linear predictor coefficients.

40. The recordable medium as set forth in claim 39, wherein said first step comprises:

an auxiliary multi-pulse setting step of setting said pulse-position-candidate information which gives priority to pulse positions where no pulse has been located rather than pulse positions which already have been set in decoding in said preceding stage or stages; and

an auxiliary multi-pulse decoding step of decoding said excitation signal of said self-stage on he basis of said pulse-position-candidate information which has been set in said auxiliary multi-pulse setting step.

41. A recordable medium on which a computer program is recorded, said computer program comprising instructions for executing an audio encoding method for encoding in multiple stages an excitation signal of an audio signal, said excitation signal being expressed by a multi-pulse signal consisting of a plurality of pulse, by selecting pulse positions of said multi-pulse signal which minimize distortion between an input audio signal and a reproduced audio signal, said reproduced audio signal being obtained by exciting a linear predictive synthesis filter by said excitation signal, said method comprising:

a main step of searching for a multi-pulse, wherein said main step encodes positions of pulses of said multi-pulse signal in a first stage by using said input audio signal on the basis of first pulse-position-candidate information which already has been determined; and

at least one auxiliary step of searching for a multi-pulse, wherein said auxiliary step comprises;

an auxiliary multi-pulse setting step which sets second pulse-position-candidate information, which will be used in a self-stage, on the basis of said multi-pulse signal which has been set in a preceding stage or stages; and

an auxiliary multi-pulse encoding step which encodes pulse positions of said multi-pulse signal in said self-stage by using said input audio signal on the basis of said second pulse-position-candidate information.

42. The recordable medium as set forth in claim 41, wherein said auxiliary multi-pulse setting step sets said second pulse-position-candidate information which gives priority to positions where no pulse has been located rather than positions which already have been encoded in said preceding stage or stages.

43. A recordable medium on which a computer program is recorded, said computer program comprising instructions for executing an audio decoding method for decoding, from encoded data, an excitation signal which has been encoded into an expression by a multi-pulse signal consisting of a plurality of pulses in multiple stages;

decoding linear predictor coefficients from said encoded data;

exciting a linear predictive synthesis filter having said linear predictor coefficients by said excitation signal, and thereby reproducing a reproduction of an audio signal, said method comprising;

a main step of creating a reproduced signal, wherein said main step creates a reproduced signal of a first stage from an excitation signal of said first stage and said linear predictor coefficients, said excitation signal of said first stage being reproduced from predetermined first pulse-position-candidate information; and

at least one auxiliary step of creating a reproduced signal, wherein said auxiliary step comprises;

an auxiliary multi-pulse setting step which sets second pulse-position-candidate information, which will be used in a self-stage, on the basis of an excitation signal which has been decoded in a preceding stage or stages; and

an auxiliary multi-pulse decoding step which decodes an excitation signal of said self-stage on the basis of said second pulse-position-candidate information; and

wherein said auxiliary step of creating a reproduced signal creates an auxiliary reproduced signal by using said excitation signal of said self-stage and said linear predictor coefficients.

44. The recordable medium as set forth in claim 43, wherein said auxiliary multi-pulse setting step sets said second pulse-position-candidate information which gives priority to positions where no pulse has been located rather than positions which already have been set by decoding in said preceding stage or stages.

45. A recordable medium on which a computer program is recorded, said computer program comprising instructions for executing an audio encoding method for encoding in multiple stages an excitation signal of an audio signal, said excitation signal being expressed by a multi-pulse signal consisting of a plurality of pulse, by selecting pulse positions of said multi-pulse signal which minimize distortion between an input audio signal and a reproduced audio signal, said reproduced audio signal being obtained by exciting a linear predictive synthesis filter by said excitation signal, said method comprising:

a main step of searching for a multi-pulse, wherein said main step encodes positions of pulses of said multi-pulse signal in a first stage by using said input audio signal on the basis of first pulse-position-candidate information which already has been determined; and

at least one auxiliary step of searching for a multi-pulse, wherein said auxiliary step comprises;

an auxiliary multi-pulse setting step which sets second pulse-position-candidate information, which will be used in a self-stage, on the basis of said multi-pulse signal which has been set in a preceding stage or stages; and

an auxiliary multi-pulse encoding step which encodes pulse positions of said multi-pulse signal in said self-stage by using said input audio signal on the basis of said second pulse-position-candidate information; and

said computer program comprising instructions for executing an audio decoding method for decoding, from encoded data, an excitation signal which has been encoded into an expression by a multi-pulse signal consisting of a plurality of pulses in multiple stages;

decoding linear predictor coefficients from said encoded data;

exciting a linear predictive synthesis filter having said linear predictor coefficients by said excitation signal, and thereby reproducing a reproduction of an audio signal, said method comprising;

a main step of creating a reproduced signal, wherein said main step creates a reproduced signal of a first stage from an excitation signal of said first stage and said linear predictor coefficients, said excitation signal of said first stage being reproduced from predetermined first pulse-position-candidate information; and

at least one auxiliary step of creating a reproduced signal, wherein said auxiliary step comprises;

an auxiliary multi-pulse setting step which sets second pulse-position-candidate information, which will be used in a self-stage, on the basis of an excitation signal which has been decoded in a preceding stage or stages; and

an auxiliary multi-pulse decoding step which decodes an excitation signal of said self-stage on the basis of said second pulse-position-candidate information; and

wherein said auxiliary step of creating a reproduced signal creates an auxiliary reproduced signal by using said excitation signal of said self-stage and said linear predictor coefficients.

46. The recordable medium as set forth in claim 45, wherein said auxiliary multi-pulse setting step sets said second pulse-position-candidate information which gives priority to positions where no pulse has been located rather than positions which already have been encoded in said preceding stage or stages.

47. The recordable medium as set forth in claim 45, wherein said auxiliary multi-pulse setting step sets said second pulse-position-candidate information which gives priority to positions where no pulse has been located rather than positions which already have been set by decoding in said preceding stage or stages.

48. The recordable medium as set forth in claim 47, wherein said auxiliary multi-pulse setting step sets said second pulse-position-candidate information which gives priority to positions where no pulse has been located rather than positions which already have been encoded in said preceding stage or stages.

49. A bitstream indicative of results of setting positions of a multi-pulse in each stage on the basis of a audio signal and an excitation signal which has been set in a stage or stages preceding said each stage and encoding said positions, said bitstream comprising:

a first bitstream indicative of positions of a multi-pulse in a first stage, said first bitstream being generated by obtaining and encoding said positions of said multi-pulse in said first stage by using said audio signal; and

a second bitstream indicative of positions of an auxiliary multi-pulse which are set by setting pulses at positions to which no pulse has been located in first through (n−

1)-th stages when encoding an excitation signal at an n-th stage on the basis of an excitation signal encoded in first through (n−

1)-th stages and said audio signal, wherein n is an integer greater than one.

50. The bitstream as set forth in claim 49, wherein said second bitstream is generated by encoding pulse positions which are set in a self-stage on the basis of pulse positions which are set by encoding in said preceding stage or stages.