Synthetically generated speech responses including prosodic characteristics of speech inputs

• US 7,280,968 B2
• Filed: 03/25/2003
• Issued: 10/09/2007
• Est. Priority Date: 03/25/2003
• Status: Active Grant

• Current Assignee: Cerence Operating Company (Cerence Inc.)
• Original Assignee: International Business Machines Corporation
• Inventors: Blass, Oscar J.
• Primary Examiner(s): Opsasnick; Michael N
• Application Number:
  US10/396,077
• Publication Number:
  US 20040193421A1
• Time in Patent Office: 1,659 Days
• Field of Search: 704/4, 704/260, 704/269
• US Class Current: 704/266
• CPC Class Codes: G06F 40/295 Named entity recognition
• CPC Class Codes: G10L 13/02 Methods for producing synthetic speech; Speech synthesisers
• CPC Class Codes: G10L 13/04 Details of speech synthesis systems, e.g. synthesiser structure or memory management

Abstract

A method for digitally generating speech with improved prosodic characteristics can include receiving a speech input, determining at least one prosodic characteristic contained within the speech input, and generating a speech output including the prosodic characteristic within the speech output.

Claims

1. A method for synthetically generating speech with improved prosodic characteristics comprising the steps of:

storing at least one pre-defined prosodic characteristic;

receiving a speech input;

extracting at least one prosodic characteristic contained within said speech input;

selecting at least one prosodic characteristic for generating a speech output, wherein said at least one prosodic characteristic is selected from said at least one extracted prosodic characteristic and said at least one pre-defined prosodic characteristic, wherein said at least one pre-defined prosodic characteristic is selected if said at least one extracted prosodic characteristic matches at least one condition; and

,

generating a speech output including said at least one selected prosodic characteristic within said speech output.

2. The method of claim 1, wherein said speech output includes a portion of said speech input, wherein said portion of said speech output utilizes said at least one prosodic characteristic.

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

upon completing said determining step, storing said prosodic characteristic into a data store; and

,

before said generating step, retrieving said prosodic characteristic from said data store.

4. The method of claim 1, wherein said receiving step occurs during a first session and wherein said generating step occurs during a second session, and wherein said first session and said second session represent two different interactive periods for a common user.

5. The method of claim 1, wherein at least one prosodic characteristic is selected from the group consisting of the speed before and after a word, the pause before and after a word, the rhyme of words, the relative tones of a word, the relative stresses applied to a word, the relative stresses applied to a syllable, and the relative stresses applied to a syllable combination.

6. The method of claim 1, wherein said receiving step and said generating step are performed by an interactive voice response system.

7. The method of claim 1, further comprising the steps of:

converting said speech input into an input text string; and

,

performing a function responsive to said converting step.

8. The method of claim 7, further comprising the steps of:

generating an output text string responsive to said performing step; and

,

converting said output text string into said speech output.

9. The method of claim 1, further comprising the steps of:

identifying a part of speech associated with at least one word within said speech input; and

,

detecting said at least one prosodic characteristic for said selected part of speech.

10. The method of claim 9, wherein said part of speech is a proper noun.

11. A system for generating synthetic speech comprising:

a speech recognition component capable of extracting at least one prosodic characteristic from speech input;

a prosodic characteristic store configured to store and permit retrieval of said at least one extracted prosodic characteristic and at least one pre-defined prosodic characteristic; and

,

a text-to-speech component capable of modifying at least a portion of synthetically generated speech based upon at least one prosodic characteristic, wherein said at least one prosodic characteristic is selected from said at least one extracted prosodic characteristic and said at least pre-defined prosodic characteristics, wherein said at least one pre-defined prosodic characteristic is selected if said at least one extracted prosodic characteristic matches at least one condition.

12. The system of claim 11, wherein said system is an interactive voice response system.

13. A machine readable storage having stored thereon, a computer program having a plurality of code sections, said code sections executable by a machine for causing the machine to perform the steps of:

storing at least one pre-defined prosodic characteristic;

receiving a speech input;

extracting at least one prosodic characteristic contained within said speech input;

selecting at least one prosodic characteristic for generating a speech output, wherein said at least one prosodic characteristic is selected from said at least one extracted prosodic characteristic and said at least one pre-defined prosodic characteristic, wherein said at least one pre-defined prosodic characteristic is selected if said at least one extracted prosodic characteristic matches at least one condition; and

,

generating a speech output including said at least one selected prosodic characteristic within said speech output.

14. The machine readable storage of claim 13, wherein said speech output includes a portion of said speech input, wherein said portion of said speech output utilizes said at least one prosodic characteristic.

15. The machine readable storage of claim 13, further comprising the steps of:

upon completing said determining step, storing said prosodic characteristic into a data store; and

,

before said generating step, retrieving said prosodic characteristic from said data store.

16. The machine readable storage of claim 13, wherein said receiving step occurs during a first session and wherein said generating step occurs during a second session, and wherein said first session and said second session represent two different interactive periods for a common user.

17. The machine readable storage of claim 13, wherein at least one prosodic characteristic is selected from the group consisting of the speed before and after a word, the pause before and after a word, the rhyme of words, the relative tones of a word, the relative stresses applied to a word, the relative stresses applied to a syllable, and the relative stresses applied to a syllable combination.

18. The machine readable storage of claim 13, wherein said receiving step and said generating step are performed by an interactive voice response system.

19. The machine readable storage of claim 13, further comprising the steps of:

converting said speech input into an input text string; and

,

performing a function responsive to said converting step.

20. The machine readable storage of claim 19, further comprising the steps of:

generating an output text string responsive to said performing step; and

,

converting said output text string into said speech output.

21. The machine readable storage of claim 13, further comprising the steps of:

identifying a part of speech associated with at least one word within said speech input; and

,

detecting said at least one prosodic characteristic for said selected part of speech.

22. The machine readable storage of claim 21, wherein said part of speech is a proper noun.

23. A method for synthetically generating speech comprising the steps of:

receiving a speech input;

analyzing said speech input to generate special handling instructions, said instructions comprising generating speech using at least one prosodic characteristic selected from said at least one extracted prosodic characteristic and said at least one pre-defined prosodic characteristic, wherein said at least one pre-defined prosodic characteristic is selected if said at least one extracted prosodic characteristic matches at least one condition; and

,

altering at least one speech generation characteristic of a text-to-speech application based upon said special handling instructions.

24. The method of claim 23, wherein said at least one condition is based upon at least one of a language proficiency level and an emotional state of the listener.

25. The method of claim 23, wherein said speech generation characteristic alters at least one of clarity and pace of speech output.

26. A machine readable storage having stored thereon, a computer program having a plurality of code sections, said code sections executable by a machine for causing the machine to perform the steps of:

receiving a speech input;

analyzing said speech input to generate special handling instructions, said instructions comprising generating speech using at least one prosodic characteristic selected from said at least one extracted prosodic characteristic and said at least one pre-defined prosodic characteristic, wherein said at least one pre-defined prosodic characteristic is selected if said at least one extracted prosodic characteristic matches at least one condition; and

,

altering at least one speech generation characteristic of a text-to-speech application based upon said special handling instructions.

27. The machine readable storage of claim 26, wherein wherein said at least one condition is based upon at least one of a language proficiency level and an emotional state of the listener.

28. The machine readable storage of claim 26, wherein said speech generation characteristic alters at least one of clarity and pace of speech output.