Speech synthesis

• US 20050102144A1
• Filed: 11/06/2003
• Published: 05/12/2005
• Est. Priority Date: 11/06/2003
• Status: Abandoned Application

• Current Assignee: Ezra J. Rapoport
• Original Assignee: Ezra J. Rapoport
• Inventors: Rapoport, Ezra J.
• Application Number:
  US10/704,326
• Publication Number:
  US 20050102144A1
• Time in Patent Office: Days
• US Class Current: 704/269
• CPC Class Codes: G10L 13/033 Voice editing, e.g. manipulating the voice of the synthesiser

Abstract

A method for speech synthesis includes combining principal components corresponding to a phoneme with a set of coefficients to produce a signal representing a synthesized expression of the phoneme. The method may also include applying the synthesized expression to a transducer to generate synthesized speech. The method may include generating the phoneme from text.

Claims

1. A method for speech synthesis, comprising:

combining principal components corresponding to a phoneme with a set of coefficients to produce a signal representing a synthesized expression of the phoneme.

2. The method of claim 1, further comprising:

applying the synthesized expression to a transducer to generate synthesized speech.

3. The method of claim 1, further comprising:

generating the phoneme from text.

4. The method of claim 1, further comprising:

receiving speech spoken from a user; and

extracting the set of coefficients.

5. The method of claim 4 wherein extracting the set of coefficients comprises:

changing the set of coefficients to include changes in intonation.

6. The method of claim 1, further comprising:

extracting the principal components from a database.

7. An article comprising a machine-readable medium that stores executable instructions for speech synthesis, the instructions causing a machine to:

combine principal components corresponding to a phoneme with a set of coefficients to produce a signal representing a synthesized expression of the phoneme.

8. The article of claim 7, further comprising instructions causing a machine to:

apply the synthesized expression to a transducer to generate synthesized speech.

9. The article of claim 7, further comprising instructions causing a machine to:

generate the phoneme from text.

10. The method of claim 7, further comprising instructions causing a machine to:

receive speech spoken from a user; and

extract the set of coefficients.

11. The article of claim 10 wherein instructions causing a machine to extract the set of coefficients comprises instructions causing a machine to:

change the set of coefficients to include changes in intonation.

12. The article of claim 7, further comprising instructions causing a machine to:

extract the principal components from a database.

13. An apparatus comprising:

a memory that stores executable instructions for speech synthesis; and

a processor that executes the instructions to;

combine principal components corresponding to a phoneme with a set of coefficients to produce a signal representing a synthesized expression of the phoneme.

14. The apparatus of claim 13, further comprising instructions to:

apply the synthesized expression to a transducer to generate synthesized speech.

15. The apparatus of claim 13, further comprising instructions to:

generate the phoneme from text.

16. The apparatus of claim 13, further comprising instructions to:

receive speech spoken from a user; and

extract the set of coefficients.

17. The apparatus of claim 16 wherein instructions to extract the set of coefficients comprises instructions to:

change the set of coefficients to include changes in intonation.

18. The apparatus of claim 13, further comprising instructions to:

extract the principal components from a database.