METHOD, APPARATUS AND COMPUTER PROGRAM PRODUCT FOR PROVIDING IMPROVED SPEECH SYNTHESIS

• US 20090299747A1
• Filed: 05/29/2009
• Published: 12/03/2009
• Est. Priority Date: 05/30/2008
• Status: Active Grant

• Current Assignee: Nokia Technologies Oy (Nokia Corporation)
• Original Assignee: Nokia Corporation
• Inventors: Suni, Antti Santeri, Vainio, Martti Tapani, Raitio, Tuomo Johannes, Alku, Paavo Ilmari, Nurminen, Jani Kristian
• Granted Patent: US 8,386,256 B2
• Time in Patent Office: Days
• US Class Current: 704/264
• CPC Class Codes: G10L 13/04 Details of speech synthesis systems, e.g. synthesiser structure or memory management

Abstract

An apparatus for providing improved speech synthesis may include a processor and a memory storing executable instructions. In response to execution of the instructions by the processor, the apparatus may perform at least selecting a real glottal pulse from among one or more stored real glottal pulses based at least in part on a property associated with the real glottal pulse, utilizing the real glottal pulse selected as a basis for generation of an excitation signal, and modifying the excitation signal based on spectral parameters generated by a model to provide synthetic speech.

Claims

1. An apparatus comprising a processor and a memory storing executable instructions that in response to execution by the processor cause the apparatus to at least perform the following:

selecting a real glottal pulse from among one or more stored real glottal pulses based at least in part on a property associated with the real glottal pulse;

utilizing the real glottal pulse selected as a basis for generation of an excitation signal; and

modifying the excitation signal based on spectral parameters generated by a model to provide synthetic speech.

2. An apparatus according to claim 1, wherein the instructions further cause the apparatus to modify the excitation signal based on spectral parameters generated by the model by filtering the excitation signal based on spectral parameters generated by a hidden Markov model framework.

3. An apparatus according to claim 2, wherein the instructions further cause the apparatus to train the hidden Markov model framework using parameters generated at least in part based on glottal inverse filtering.

4. An apparatus according to claim 2, wherein the instructions further cause the apparatus to select the real glottal pulse by selecting the real glottal pulse at least in part based on parameters associated with the hidden Markov model framework.

5. An apparatus according to claim 1, wherein the instructions further cause the apparatus to select the real glottal pulse by selecting a current pulse based at least in part on a previously selected pulse.

6. An apparatus according to claim 1, wherein the instructions further cause the apparatus to select the real glottal pulse by selecting the real glottal pulse based on a fundamental frequency associated with the real glottal pulse.

7. An apparatus according to claim 6, the instructions further cause the apparatus to modify the fundamental frequency.

8. An apparatus according to claim 7, wherein the instructions further cause the apparatus to modify the fundamental frequency by utilizing time domain or frequency techniques for modifying the fundamental frequency.

9. An apparatus according to claim 6, wherein the instructions further cause the apparatus to select the real glottal pulse by selecting at least two pulses and wherein modifying the fundamental frequency comprises combining the at least two pulses into a single pulse.

10. An apparatus according to claim 1, wherein the instructions further cause the apparatus to perform an initial operation of estimating the plurality of stored real glottal pulses from corresponding natural speech signals using glottal inverse filtering.

11. A method comprising:

selecting a real glottal pulse from among one or more stored real glottal pulses based at least in part on a property associated with the real glottal pulse;

utilizing the real glottal pulse selected as a basis for generation of an excitation signal; and

modifying, via a processor, the excitation signal based on spectral parameters generated by a model to provide synthetic speech.

12. A method according to claim 11, wherein modifying the excitation signal based on spectral parameters generated by the model comprises modifying the excitation signal based on spectral parameters generated by a hidden Markov model framework.

13. A method according to claim 11, wherein selecting the real glottal pulse further comprises selecting a current pulse based at least in part on a previously selected pulse.

14. A method according to claim 11, wherein selecting the real glottal pulse further comprises selecting the real glottal pulse based on a fundamental frequency associated with the real glottal pulse.

15. A method according to claim 11, further comprising an initial operation of estimating the plurality of stored real glottal pulses from corresponding natural speech signals using glottal inverse filtering.

16. A computer program product comprising at least one computer-readable storage medium having computer-executable program code portions stored therein, the computer-executable program code portions comprising:

program code instructions for selecting a real glottal pulse from among one or more stored real glottal pulses based at least in part on a property associated with the real glottal pulse;

program code instructions for utilizing the real glottal pulse selected as a basis for generation of an excitation signal; and

program code instructions for modifying the excitation signal based on spectral parameters generated by a model to provide synthetic speech.

17. A computer program product according to claim 16, wherein the program code instructions for modifying the excitation signal include instructions for modifying the excitation signal based on spectral parameters generated by a hidden Markov model framework.

18. A computer program product according to claim 16, wherein the program code instructions for selecting the real glottal pulse include instructions for selecting a current pulse based at least in part on a previously selected pulse.

19. A computer program product according to claim 16, wherein the program code instructions for selecting the real glottal pulse include instructions for selecting the real glottal pulse based on a fundamental frequency associated with the real glottal pulse.

20. A computer program product according to claim 16, further comprising program code instructions for an initial operation of estimating the plurality of stored real glottal pulses from corresponding natural speech signals using glottal inverse filtering.