Methods for generating the voiced portion of speech signals

• US 5,581,656 A
• Filed: 04/06/1993
• Issued: 12/03/1996
• Est. Priority Date: 09/20/1990
• Status: Expired due to Term

• Current Assignee: Digital Voice Systems, Inc.
• Original Assignee: Digital Voice Systems, Inc.
• Inventors: Hardwick, John C., Lim, Jae S.
• Primary Examiner(s): MacDonald, Allen R.
• Assistant Examiner(s): EDOUARD, PATRICK NESTOR
• Application Number:
  US08/043,286
• Time in Patent Office: 1,337 Days
• Field of Search: 381/29-40, 381/50-53, 395/2.1-2.39, 395/2.67-2.78
• US Class Current: 704/258
• CPC Class Codes: G10L 19/087 using mixed excitation models, e.g. MELP, MBE, split band LPC or HVXC
• CPC Class Codes: G10L 25/90 Pitch determination of speech signals
• CPC Class Codes: G10L 25/93 Discriminating between voiced and unvoiced parts of speech signals G10L25/90 takes precedence

Abstract

The pitch estimation method is improved. Sub-integer resolution pitch values are estimated in making the initial pitch estimate; the sub-integer pitch values are preferably estimated by interpolating intermediate variables between integer values. Pitch regions are used to reduce the amount of computation required in making the initial pitch estimate. Pitch-dependent resolution is used in making the initial pitch estimate, with higher resolution being used for smaller values of pitch. The accuracy of the voiced/unvoiced decision is improved by making the decision dependent on the energy of the current segment relative to the energy of recent prior segments; if the relative energy is low, the current segment favors an unvoiced decision; if high, it favors a voiced decision. Voiced harmonics are generated using a hybrid approach; some voiced harmonics are generated in the time domain, whereas the remaining harmonics are generated in the frequency domain; this preserves much of the computational savings of the frequency domain approach, while at the same time improving speech quality. Voiced harmonics generated in the frequency domain are generated with higher frequency accuracy; the harmonics are frequency scaled, transformed into the time domain with a Discrete Fourier Transform, interpolated and then time scaled.

Claims

1. A method for synthesizing a portion or all of a speech signal from harmonics, the method comprising the steps of:

processing a signal containing information on a plurality of harmonics that consist of at least two groups of said harmonics;

generating said first group of harmonics from said signal using a time domain synthesis method;

generating a second signal from said information on said second group of harmonics using a frequency domain synthesis method; and

combining said generated first group of harmonics and said generated second signal to generate said portion or all of a speech signal.

2. The method of claim 1 wherein said first group comprises low-frequency harmonics.

3. The method of claim 1 wherein said second group comprises high-frequency harmonics.

4. The method of claim 1 wherein said time domain synthesis is performed by generating a low-order piecewise phase polynomial.

5. The method of claim 1 wherein said frequency domain synthesis is performed using the method comprising the steps of:

linearly frequency scaling said information on said harmonics according to the mapping ##EQU24## where L is some small integer, to generate frequency-scaled harmonics;

performing an L-point Inverse Discrete Fourier Transform (DFT) to simultaneously transform said frequency scaled harmonics into the time domain; and

performing interpolation and time scaling to generate said second signal.

6. The method of claim 1 or 5 wherein said synthesized speech is the voiced portion of a multi-band excitation (MBE) speech coder.

7. The method of claim 1 or 2 wherein said information on said plurality of harmonics represents parameters of a Sinusoidal Transform Coder.

8. A method for synthesizing a portion or all of a speech signal from harmonics, the method comprising the steps of:

processing a signal containing information on a plurality of harmonics that consist of at least two groups of said harmonics;

generating a first signal from said first group of harmonics using a time domain synthesis method;

generating a second signal from said second group of harmonics using a frequency domain synthesis method; and

combining said first signal and said second signal to generate said portion or all of a speech signal.

9. The method of claim 8 wherein said first group comprises low-frequency harmonics.

10. The method of claim 8 wherein said second group comprises high-frequency harmonics.

11. The method of claim 8 wherein said time domain synthesis is performed by generating a low-order piecewise phase polynomial.

12. The method of claim 8 wherein said frequency domain synthesis is performed using the method comprising the steps of:

linearly frequency scaling said information on said harmonics according to the mapping ##EQU25## where L is some small integer, to generate frequency-scaled harmonics;

performing an L-point Inverse Discrete Fourier Transform (DFT) to simultaneously transform said frequency scaled harmonics into the time domain; and

performing interpolation and time scaling to generate said second signal.

13. The method of claim 8 or 12 wherein said synthesized speech is the voiced portion of a multi-band excitation (MBE) speech coder.

14. The method of claim 8 or 12 wherein said information on said plurality of harmonics represents parameters of a Sinusoidal Transform Coder.

18. The method of claim 5, 12 or 15 wherein said DFT is computed with a Fast Fourier Transform.

19. The method of claim 5, 12 or 15 wherein said interpolation is performed with linear interpolation.

15. A method for synthesizing a portion or all of a speech signal from harmonics, the method comprising the steps of:

processing a signal containing information on a plurality of harmonics;

linearly frequency scaling said information on said harmonics according to the mapping ##EQU26## where L is some small integer, to generate frequency-scaled harmonics;

performing an L-point Inverse Discrete Fourier Transform (DFT) to simultaneously transform said frequency scaled harmonics into the time domain; and

performing interpolation and time scaling on said frequency scaled harmonics that have been transformed into the time domain to generate said portion or all of a speech signal.

16. The method of claim 15 wherein said synthesized speech is the voiced portion of a multi-band excitation (MBE) speech coder.

17. The method of claim 15 wherein said information on said plurality of harmonics represents parameters of a Sinusoidal Transform Coder.