SYSTEM AND METHOD FOR PERFORMING SPEECH SYNTHESIS WITH A CACHE OF PHONEME SEQUENCES

• US 20090043585A1
• Filed: 08/09/2007
• Published: 02/12/2009
• Est. Priority Date: 08/09/2007
• Status: Active Grant

• Current Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Original Assignee: AT&T Corporation (AT&T, Inc.)
• Inventors: CONKIE, Alistair D.
• Granted Patent: US 7,983,919 B2
• Time in Patent Office: Days
• US Class Current: 704/267
• CPC Class Codes: G10L 13/04 Details of speech synthesis systems, e.g. synthesiser structure or memory management
• CPC Class Codes: G10L 13/08 Text analysis or generation of parameters for speech synthesis out of text, e.g. grapheme to phoneme translation, prosody generation or stress or intonation determination

Abstract

Disclosed are systems, methods, and computer readable media for performing speech synthesis. The method embodiment comprises applying a first part of a speech synthesizer to a text corpus to obtain a plurality of phoneme sequences, the first part of the speech synthesizer only identifying possible phoneme sequences, for each of the obtained plurality of phoneme sequences, identifying joins that would be calculated to synthesize each of the plurality of respective phoneme sequences, and adding the identified joins to a cache for use in speech synthesis.

Claims

1. A method of performing speech synthesis, the method comprising:

applying a first part of a speech synthesizer to a text corpus to obtain a plurality of phoneme sequences, the first part of the speech synthesizer only identifying possible phoneme sequences;

for each of the obtained plurality of phoneme sequences, identifying joins that would be calculated to synthesize each of the plurality of respective phoneme sequences; and

adding the identified joins to a cache for use in speech synthesis.

2. The method of claim 1, the method further comprising:

recording a frequency of occurrence for each of the plurality of phoneme sequences; and

pruning the cache.

3. The method of claim 1, the method further comprising:

building a plurality of caches of different sizes based on values or parameters.

4. The method of claim 3, wherein the values or parameters comprise computational costs or frequency of occurrence.

5. A method of synthesizing a speech signal, the method comprising:

(1) selecting one or more acoustic units from an acoustic unit database;

(2) determining whether a join cost of an acoustic unit sequential pair resides in a cache created by steps comprising;

(a) applying a first part of a speech synthesizer to a text corpus to obtain a plurality of phoneme sequences, the first part of the speech synthesizer only identifying possible phoneme sequences;

(b) for each of the obtained plurality of phoneme sequences, identifying joins that would be calculated to synthesize each of the plurality of respective phoneme sequences; and

(c) adding the identified joins to a cache for use in speech synthesis

(3) if the cache contains the join, extracting the join from the cache for use in speech synthesis; and

(4) if the cache does not contain the join, calculating a value of the join for use in speech synthesis.

6. The method of claim 5, wherein calculating the value of the join cost is performed to enhance accuracy over speed.

7. A system for performing speech synthesis, the system comprising:

a module configured to apply a first part of a speech synthesizer to a text corpus to obtain a plurality of phoneme sequences, the first part of the speech synthesizer only identifying possible phoneme sequences;

for each of the obtained plurality of phoneme sequences, a module configured to identify joins that would be calculated to synthesize each of the plurality of respective phoneme sequences; and

a module configured to add the identified joins to a cache for use in speech synthesis.

8. The system of claim 7, the system further comprising:

a module configured to record a frequency of occurrence for each of the plurality of phoneme sequences; and

a module configured to prune the cache.

9. The system of claim 7, the system further comprising:

a module configured to build a plurality of caches of different sizes based on values or parameters.

10. The system of claim 9, wherein the values or parameters comprise computational costs or frequency of occurrence.

11. A system for synthesizing a speech signal, the system comprising:

(1) a module configured to select one or more acoustic units from an acoustic unit database;

(2) a module configured to determine whether a join cost of an acoustic unit sequential pair resides in a cache created by steps comprising;

(a) applying a first part of a speech synthesizer to a text corpus to obtain a plurality of phoneme sequences, the first part of the speech synthesizer only identifying possible phoneme sequences;

(b) for each of the obtained plurality of phoneme sequences, identifying joins that would be calculated to synthesize each of the plurality of respective phoneme sequences; and

(c) adding the identified joins to a cache for use in speech synthesis

(3) if the cache contains the join, a module configured to extract the join from the cache for use in speech synthesis; and

(4) if the cache does not contain the join, a module configured to calculate a value of the join for use in speech synthesis.

12. The system of claim 11, wherein calculating the value of the join cost is performed to enhance accuracy over speed.

13. A computer readable medium storing a computer program having instructions for performing speech synthesis, the instructions comprising:

applying a first part of a speech synthesizer to a text corpus to obtain a plurality of phoneme sequences, the first part of the speech synthesizer only identifying possible phoneme sequences;

for each of the obtained plurality of phoneme sequences, identifying joins that would be calculated to synthesize each of the plurality of respective phoneme sequences; and

adding the identified joins to a cache for use in speech synthesis.

14. The computer readable medium of claim 13, the instructions further comprising:

recording a frequency of occurrence for each of the plurality of phoneme sequences; and

pruning the cache.

15. The computer readable medium of claim 13, the instructions further comprising:

building a plurality of caches of different sizes based on values or parameters.

16. The computer readable medium of claim 15, wherein the values or parameters comprise computational costs or frequency of occurrence.

17. A computer readable medium storing a computer program having instructions for synthesizing a speech signal, the instructions comprising:

(1) selecting one or more acoustic units from an acoustic unit database;

(2) determining whether a join cost of an acoustic unit sequential pair resides in a cache created by steps comprising;

(a) applying a first part of a speech synthesizer to a text corpus to obtain a plurality of phoneme sequences, the first part of the speech synthesizer only identifying possible phoneme sequences;

(b) for each of the obtained plurality of phoneme sequences, identifying joins that would be calculated to synthesize each of the plurality of respective phoneme sequences; and

(c) adding the identified joins to a cache for use in speech synthesis

(3) if the cache contains the join, extracting the join from the cache for use in speech synthesis; and

(4) if the cache does not contain the join, calculating a value of the join for use in speech synthesis.

18. The computer readable medium of claim 17, wherein calculating the value of the join cost is performed to enhance accuracy over speed.