Apparatus and method for speech encoding based on short-term prediction valves

• US 5,950,155 A
• Filed: 09/18/1996
• Issued: 09/07/1999
• Est. Priority Date: 12/21/1994
• Status: Expired due to Term

• Current Assignee: Sony Corporation (Sony Group Corp.)
• Original Assignee: Sony Corporation (Sony Group Corp.)
• Inventors: Nishiguchi, Masayuki
• Primary Examiner(s): Hudspeth, David R.
• Assistant Examiner(s): SOFOCLEOUS, MICHAEL D
• Application Number:
  US08/676,226
• Time in Patent Office: 1,084 Days
• Field of Search: 395/2.19, 395/2.2, 395/2.21, 395/2.37, 395/2.28, 395/2.32, 395/2.17, 704/219, 704/222, 704/230, 704/262
• US Class Current: 704/230
• CPC Class Codes: G10L 19/07 Line spectrum pair [LSP] vocoders
• 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 2019/0005 Multi-stage vector quantisation
• CPC Class Codes: G10L 25/24 the extracted parameters being the cepstrum

Abstract

Foe executing the code excitation linear prediction (CELP) coding, for example, α-parameters are taken out from the input speech signal by a linear prediction coding (LPC) analysis circuit 12. The α-parameters are then converted by an α-parameter to LSP converting circuit 13 into linear spectral pair (LSP) parameters and a vector of these line spectral pair (LSP) parameters is vector-quantized by a quantizer 14. The changeover switch 16 is controlled depending upon the pitch value detected by a pitch detection circuit 22 for selecting and using one of the codebook 15M for male voice and the codebook 15F for female voice for improving quantization characteristics without increasing the transmission bit rate.

Claims

1. A speech encoding method comprising the steps of:

generating short-term prediction coefficients based on an input speech signal;

providing first and second codebooks formed of assorted parameters representing said short-term prediction coefficients, said first and second codebooks relating to at least one of a plurality of characteristic parameters of said input speech signal;

selecting one of said first and second codebooks based on a pitch value of said input speech signal; and

quantizing said short-term prediction coefficients using said selected codebook.

2. The speech encoding method as claimed in claim 1, wherein said plurality of characteristic parameters includes said pitch value, a pitch strength, a frame power, a voiced/unvoiced discrimination flag, and a gradient of a signal spectrum.

3. The speech encoding method as claimed in claim 1, wherein said step of quantizing includes vector-quantizing said short-term prediction coefficients.

4. The speech encoding method as claimed in claim 1, wherein said step of quantizing includes matrix-quantizing said short-term prediction coefficients.

5. The speech encoding method as claimed in claim 1, wherein

said step of selecting includes selecting one of said first and second codebooks based on a magnitude relation between said pitch value of said input speech signal and a pre-set pitch value.

6. A speech encoding method comprising the steps of:

generating short-term prediction errors based on an input speech signal;

providing first and second codebooks formed of assorted parameters representing said short-term prediction errors, said first and second codebooks relating to at least one of a plurality of characteristic parameters of said input speech signal;

selecting one of said first and second codebooks based on a pitch value of said input speech signal; and

quantizing said short-term prediction errors using said selected codebook.

7. The speech encoding method as claimed in claim 6, wherein said plurality of characteristic parameters includes said pitch value, a pitch intensity, a frame power, a voiced/unvoiced discrimination flag, and a gradient of a signal spectrum.

8. The speech encoding method as claimed in claim 6, wherein said step of quantizing includes vector quantizing said short-term prediction errors.

9. The speech encoding method as claimed in claim 6, wherein said step of quantizing includes matrix-quantizing said short-term prediction errors.

10. A speech encoding apparatus comprising:

short-term prediction means for generating short-term prediction coefficients based on an input speech signal;

first and second codebooks formed of assorted parameters representing said short-term prediction coefficients, said first and second codebooks relating to one or more of a plurality of characteristic parameters of said input speech signal;

selection means for selecting one of said first and second codebooks based on a pitch value of said input speech signal; and

quantization means for quantizing said short-term prediction coefficients using said selected codebook.

11. The speech encoding apparatus as claimed in claim 10, wherein said plurality of characteristic parameters includes said pitch value, a pitch strength, a frame power, a voiced/unvoiced discrimination flag, and a gradient of a signal spectrum.

12. The speech encoding apparatus as claimed in claim 10, wherein said quantizing means vector-quantizes said short-term prediction coefficients.

13. The speech encoding apparatus as claimed in claim 10, wherein said quantizing means matrix-quantizes said short-term prediction coefficients.

14. A speech encoding apparatus comprising:

short-term prediction means for generating short-term prediction coefficients based on an input speech signal;

a first plurality of codebooks formed of assorted parameters representing said short-term prediction coefficients, said first plurality of codebooks relating to reference parameters of said input speech signal, said reference parameters including at least one of a plurality of characteristic parameters of said input speech signal;

selecting means for selecting one of said first plurality of codebooks based on said reference parameters of said input speech signal;

quantization means for quantizing said short-term prediction coefficients based on said codebook selected from said first plurality of codebooks;

a second plurality of codebooks formed on the basis of training data corresponding to said reference parameters; and

synthesis means for synthesizing an excitation signal which relates to an output of a codebook selected from said second plurality of codebooks based on a quantized value from said quantization means.

15. The speech encoding apparatus as claimed in claim 14, wherein said plurality of characteristic parameters includes a pitch value, a pitch strength, a frame power, a voice/unvoiced discrimination flag, and a gradient of a signal spectrum.

16. The speech encoding apparatus as claimed in claim 14, wherein said quantization means vector-quantizes said short-term prediction coefficients.

17. The speech encoding apparatus as claimed in claim 14, wherein said quantization means matrix-quantizes said short-term prediction coefficients.

18. The speech encoding apparatus as claimed in claim 14, wherein

said reference parameters include a pitch value of said input speech signal, and

said selection means selects one of said first plurality of codebooks based on said pitch value of said input speech signal.

19. The speech encoding apparatus as claimed in claim 14, wherein each of said first plurality of codebooks and said second plurality of codebooks includes a codebook for a male voice and a codebook for a female voice.