Sub-sampled excitation waveform codebooks

• US 7,698,132 B2
• Filed: 12/17/2002
• Issued: 04/13/2010
• Est. Priority Date: 12/17/2002
• Status: Active Grant

• Current Assignee: Qualcomm, Inc.
• Original Assignee: Qualcomm, Inc.
• Inventors: El-Maleh, Khaled, Kandhadai, Ananthapadamanabhan A., Manjunath, Sharath
• Primary Examiner(s): Dorvil; Richemond
• Assistant Examiner(s): YEN, ERIC L
• Application Number:
  US10/322,245
• Publication Number:
  US 20040117176A1
• Time in Patent Office: 2,674 Days
• Field of Search: 704222-223, 704/220, 704/227, 704/230
• US Class Current: 704/222
• 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 19/20 using sound class specific coding, hybrid encoders or object based coding
• CPC Class Codes: G10L 19/22 Mode decision, i.e. based on audio signal content versus external parameters
• CPC Class Codes: G10L 2019/0004 Design or structure of the codebook

Abstract

Methods and apparatus are presented for reducing the number of bits needed to represent an excitation waveform. An acoustic signal in an analysis frame is analyzed to determine whether it is a band-limited signal. A sub-sampled sparse codebook is used to generate the excitation waveform if the acoustic signal is a band-limited signal. The sub-sampled sparse codebook is generated by decimating permissible pulse locations from the codebook track in accordance with the frequency characteristic of the acoustic signal.

Claims

1. A method for forming an excitation waveform in a speech coder, the method comprising:

determining whether an acoustic signal in an analysis frame is a band-limited signal;

if the acoustic signal is a band-limited signal, then using a sub-sampled sparse codebook to generate the excitation waveform, wherein the sub-sampled sparse codebook comprises either only even track positions or only odd track positions from the sparse codebook; and

if the acoustic signal is not a band-limited signal, then using a sparse codebook to generate the excitation waveform,

wherein the sparse codebook comprises a set of predetermined possible positions and the sub-sampled sparse code book comprises a subset of the predetermined positions, such that the excitation waveform is generated through placement of pulses within the predetermined positions or the subset; and

wherein using the sub-sampled sparse codebook to generate the excitation waveform comprises generating an initial excitation waveform, determining whether the initial excitation waveform comprises mostly odd track positions or mostly even track positions, and decimating the initial excitation waveform to generate the excitation waveform.

2. The method of claim 1, wherein determining whether an acoustic signal in an analysis frame is a band-limited signal comprises:

determining a voice activity level of the acoustic signal; and

using the voice activity level to determine whether the acoustic signal is a band-limited signal.

3. The method of claim 1, wherein determining whether an acoustic signal in an analysis frame is a band-limited signal comprises:

comparing an energy level of a low frequency band of the acoustic signal to an energy level of a high frequency band of the acoustic signal; and

if the energy level of the low frequency band of the acoustic signal is higher than the energy level of the high frequency band of the acoustic signal, then deciding that the acoustic signal is a band-limited signal.

4. The method of claim 1, wherein determining whether an acoustic signal in an analysis frame is a band-limited signal comprises:

determining a zero-crossing rate for the acoustic signal; and

if the zero-crossing rate is low, then deciding that the acoustic signal is a band-limited signal.

5. The method of claim 1, wherein determining whether an acoustic signal in an analysis frame is a band-limited signal comprises:

determining the periodicity of a low frequency band of the acoustic signal; and

if the periodicity of the low frequency band of the acoustic signal is high, then deciding that the acoustic signal is a band-limited signal.

6. The method of claim 1, wherein determining whether an acoustic signal in an analysis frame is a band-limited signal comprises:

analyzing the spectral content of the acoustic signal for a significant band-limited component.

7. The method of claim 1, further comprising:

determining at least one of a spectral content, voice activity and zero-crossing rate of the acoustic signal; and

based on determining at least one of the spectral content, voice activity and zero-crossing rate of the acoustic signal, generating the sub-sampled sparse codebook.

8. The method of claim 1, further comprising excluding certain candidate excitation waveforms from a search through a stochastic excitation waveform codebook.

9. The method of claim 1, if the acoustic signal is band-limited, further comprising reallocating bits, which would have been used to represent an excitation waveform from the sparse codebook, to represent another speech encoding parameter.

10. The method of claim 9, wherein the speech encoding parameter comprises a linear predictive coding (LPC) filter coefficient.

11. The method of claim 1, further comprising:

generating multiple candidate excitation waveforms based on different sub-sampled sparse codebooks; and

determining which of the multiple candidate excitation waveforms is better suited for acting as the excitation waveform.

12. Apparatus for forming an excitation waveform, comprising:

a memory element; and

a processing element configured to execute a set of instructions stored on the memory element, the set of instructions for;

determining whether an acoustic signal in an analysis frame is a band-limited signal;

using a sub-sampled sparse codebook to generate the excitation waveform if the acoustic signal is a band-limited signal, wherein the sub-sampled sparse codebook comprises either only even track positions or only odd track positions from a sparse codebook; and

using the sparse codebook to generate the excitation waveform if the acoustic signal is not a band-limited signal, wherein the sparse codebook comprises a set of predetermined possible positions and the sub-sampled sparse code book comprises a subset of the predetermined positions, such that the excitation waveform is generated through placement of pulses within the predetermined positions or the subset; and

wherein using the sub-sampled sparse codebook to generate the excitation waveform comprises generating an initial excitation waveform, determining whether the initial excitation waveform comprises mostly odd track positions or mostly even track positions, and decimating the initial excitation waveform to generate the excitation waveform.

13. The apparatus of claim 12, wherein the apparatus is a wideband vocoder.

14. The apparatus of claim 12, wherein the apparatus is a narrowband vocoder.

15. The apparatus of claim 12, wherein the apparatus is a variable rate vocoder.

16. The apparatus of claim 12, wherein the apparatus is a fixed rate vocoder.

17. An apparatus for forming an excitation waveform, comprising:

means for determining whether an acoustic signal in an analysis frame is a band-limited signal;

means for using a sub-sampled sparse codebook to generate the excitation waveform if the acoustic signal is a band-limited signal, wherein the sub-sampled sparse codebook comprises either only even track positions or only odd track positions from a sparse codebook; and

means for using the sparse codebook to generate the excitation waveform if die acoustic signal is not a band-limited signal, wherein the sparse codebook comprises a set of predetermined possible positions and the sub-sampled sparse codebook comprises a subset of the predetermined positions, such that die excitation waveform is generated through placement of pulses within the predetermined positions or the subset;

wherein using the sub-sampled sparse codebook to generate the excitation waveform comprises generating an initial excitation waveform, determining whether the initial excitation waveform comprises mostly odd track positions or mostly even track positions, and decimating the initial excitation waveform to generate the excitation waveform.

18. The apparatus of claim 17, wherein the apparatus is a wideband vocoder.

19. A method for a signal coder to reduce the number of bits used to represent an excitation waveform, comprising:

determining a frequency characteristic of an acoustic signal;

generating a sub-sampled sparse codebook waveform from a sparse codebook if the frequency characteristic indicates that sub-sampling does not impair the perceptual quality of the acoustic signal, wherein the sparse codebook comprises a set of predetermined possible positions and the sub-sampled sparse code book comprises a subset of the predetermined positions, such that the excitation waveform is generated through placement of pulses within the predetermined positions or the subset, wherein the sub-sampled sparse codebook comprises either only even track positions or only odd track positions from the sparse codebook; and

using the sub-sampled sparse codebook waveform to represent the excitation waveform rather than a waveform from the sparse codebook;

wherein using the sub-sampled sparse codebook to represent the excitation waveform comprises generating an initial excitation waveform, determining whether the initial excitation waveform comprises mostly odd track positions or mostly even track positions, and decimating the initial excitation waveform.

20. Apparatus for reducing the number of bits used to represent an excitation waveform, comprising:

a memory element; and

a processing element configured to execute a set of instructions stored on the memory element, the set of instructions for;

determining a frequency characteristic of an acoustic signal;

generating a sub-sampled sparse codebook waveform from a sparse codebook if the frequency characteristic indicates that sub-sampling does not impair the perceptual quality of the acoustic signal, wherein the sparse codebook comprises a set of predetermined possible positions and the sub-sampled sparse code book comprises a subset of the predetermined positions, such that the excitation waveform is generated through placement of pulses within the predetermined positions or the subset, wherein the sub-sampled sparse codebook comprises either only even track positions or only odd track positions from the sparse codebook; and

using the sub-sampled sparse codebook waveform to represent the excitation waveform rather than a waveform from the sparse codebook;

wherein using the sub-sampled sparse codebook to represent the excitation waveform comprises generating an initial excitation waveform, determining whether the initial excitation waveform comprises mostly odd track positions or mostly even track positions, and decimating the initial excitation waveform.

21. An apparatus for reducing the number of bits used to represent an excitation waveform, comprising:

means for determining a frequency characteristic of an acoustic signal;

means for generating a sub-sampled sparse codebook waveform from a sparse codebook if the frequency characteristic indicates that sub-sampling does not impair the perceptual quality of the acoustic signal, wherein the sparse codebook comprises a set of predetermined possible positions and the sub-sampled sparse code book comprises a subset of the predetermined positions, such that the excitation waveform is generated through placement of pulses within the predetermined positions or the subset, wherein the sub-sampled sparse codebook comprises either only even track positions or only odd track positions from the sparse codebook; and

means for using the sub-sampled sparse codebook waveform to represent the excitation waveform rather than a waveform from the sparse codebook;

wherein using the sub-sampled sparse codebook waveform to represent the excitation waveform comprises generating an initial excitation waveform, determining whether the initial excitation waveform comprises mostly odd track positions or mostly even track positions, and decimating the initial excitation waveform.

22. The apparatus of claim 21, wherein the apparatus is a wideband vocoder.

23. The apparatus of claim 21, wherein the apparatus is a narrowband vocoder.

24. The apparatus of claim 21, wherein the apparatus is a variable rate vocoder.

25. The apparatus of claim 21, wherein the apparatus is a fixed rate vocoder.

26. A method for execution by a suitably programmed processor to generate a sub-sampled sparse codebook from a sparse codebook, wherein the sparse codebook comprises pulses at a set of permissible pulse locations, the method comprising:

analyzing a frequency characteristic of an acoustic signal;

determining whether an initial excitation waveform corresponding to the acoustic signal comprises mostly odd track positions or mostly even track positions; and

decimating a subset of permissible pulse locations from the set of permissible pulse locations of the sparse codebook in accordance with the frequency characteristic of the acoustic signal to generate the sub-sampled sparse codebook, wherein the sub-sampled sparse codebook comprises either only even track positions or only odd track positions from the sparse codebook.

27. Apparatus for generating a sub-sampled sparse codebook from a sparse codebook, wherein the sparse codebook comprises pulses at a set of permissible pulse locations, the apparatus comprising:

a memory element; and

a processing element configured to execute a set of instructions stored on the memory element, the set of instructions for;

analyzing a frequency characteristic of an acoustic signal;

determining whether an initial excitation waveform corresponding to the acoustic signal comprises mostly odd track positions or mostly even track positions; and

decimating a subset of permissible pulse locations from the set of permissible pulse locations of the sparse codebook in accordance with the frequency characteristic of the acoustic signal to generate the sub-sampled sparse codebook, wherein the sub-sampled sparse codebook comprises either only even track positions or only odd track positions from the sparse codebook.

28. Apparatus for generating a sub-sampled sparse codebook from a sparse codebook, wherein the sparse codebook comprises pulses at a set of permissible pulse locations, the apparatus comprising:

means for analyzing a frequency characteristic of an acoustic signal;

means for determining whether an initial excitation waveform corresponding to the acoustic signal comprises mostly odd track positions or mostly even track positions; and

means for decimating a subset of permissible pulse locations from the set of permissible pulse locations of the sparse codebook in accordance with the frequency characteristic of the acoustic signal to generate the sub-sampled sparse codebook, wherein the sub-sampled sparse codebook comprises either only even track positions or only odd track positions from the sparse codebook.

29. The apparatus of claim 28, wherein the apparatus is a wideband vocoder.

30. The apparatus of claim 28, wherein the apparatus is a narrowband vocoder.

31. The apparatus of claim 28, wherein the apparatus is a variable rate vocoder.

32. The apparatus of claim 28, wherein the apparatus is a fixed rate vocoder.

33. A speech coder, comprising:

a linear predictive coding (LPC) unit configured to determine LPC coefficients of an acoustic signal;

a frequency analysis unit configured to determine whether the acoustic signal is band-limited;

a quantizer unit configured to receive the LPC coefficients to and quantize the LPC coefficients; and

an excitation parameter generator configured to receive a determination from the frequency analysis unit regarding whether the acoustic signal is band-limited and to implement a sub-sampled sparse codebook, the sparse codebook comprising a set of predetermined possible positions and the sub-sampled sparse code book comprising a subset of the predetermined positions, wherein the sub-sampled sparse codebook comprises either only even track positions or only odd track positions from the sparse codebook, and wherein implementing the sub-sampled sparse codebook comprises determining whether an initial excitation waveform comprises mostly odd track positions or mostly even track positions.

34. The speech coder of claim 33, wherein the quantizer unit is further configured to receive the determination from the frequency analysis unit regarding whether the acoustic signal is band-limited and to update the quantization scheme accordingly.

35. The speech coder of claim 33, wherein the quantizer unit is further configured to receive information from the excitation parameter generator regarding the implementation of the sub-sampled sparse codebook and to update the quantization scheme accordingly.

36. A computer-program product comprising a computer-readable medium having instructions thereon, the instructions comprising:

code for determining whether an acoustic signal in an analysis frame is a band-limited signal;

code for using a sub-sampled sparse codebook to generate an excitation waveform if the acoustic signal is a band-limited signal, wherein the sub-sampled sparse codebook comprises either only even track positions or only odd track positions from a sparse codebook; and

code for using the sparse codebook to generate the excitation waveform if the acoustic signal is not a band-limited signal, wherein the sparse codebook comprises a set of predetermined possible positions and the sub-sampled sparse code book comprises a subset of the predetermined positions, such that the excitation waveform is generated through placement of pulses within the predetermined positions or the subset;

wherein using the sub-sampled sparse codebook to generate the excitation waveform comprises generating an initial excitation waveform, determining whether the initial excitation waveform comprises mostly odd track positions or mostly even track positions, and decimating the initial excitation waveform.