System and method for generating a phrase pronunciation

• US 20050192793A1
• Filed: 02/28/2005
• Published: 09/01/2005
• Est. Priority Date: 02/27/2004
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Dictaphone Corporation (Microsoft Corporation)
• Inventors: Cote, William F., Carrier, Jill
• Granted Patent: US 7,783,474 B2
• Time in Patent Office: Days
• US Class Current: 704/4
• 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
• CPC Class Codes: G10L 15/187 Phonemic context, e.g. pronunciation rules, phonotactical constraints or phoneme n-grams

Abstract

A system and method for a speech recognition technology that allows language models to be customized through the addition of special pronunciations for components of phrases, which are added to the factory language models during customization. It allows components of a phrase to have different pronunciations inside customer-added phrases than are specified for those isolated components in the factory language models.

Claims

1. A method of adding phrase pronunciations to a language model, the method comprising the steps of:

generating a list of pron components, whose pronunciations differ when they occur in a phrase;

assigning at least one pron to each pron component;

determining the pronunciation of a first phrase, by the steps of;

tokenizing the first phrase by generating a list of tokens corresponding to the first phrase;

determining a pron for each of the list of tokens;

assembling the pronunciation of the first phrase based upon a combination of each said pron; and

adding the first phrase and the pronunciation of the first phrase to the language model.

2. The method according to claim 1, wherein tokenizing identifies the initial token.

3. The method according to claim 1, wherein tokenizing identifies the non-initial tokens.

4. The method according to claim 1, wherein the step of assigning at least one pron to each pron component includes generating a phonetic transcription for each pron component.

5. The method according to claim 4, wherein the step of generating a phonetic transcription includes referencing an item in the language model.

6. The method according to claim 4, wherein the step of generating a phonetic transcription includes specifying a literal phonetic transcription.

7. The method according to claim 1, wherein determining a pron for each token includes searching a pron component list to determine said pron.

8. The method according to claim 7, wherein the pron component list includes non-initial components.

9. The method according to claim 7, wherein the pron component list includes initial components.

10. The method according to claim 1, wherein determining a pron for each token includes searching the language model to determine said pron.

11. The method according to claim 1, wherein determining a pron for each token includes searching a background dictionary to determine said pron.

12. The method according to claim 1, wherein determining a pron for each token includes applying a pron guesser to determine said pron.

13. A system for adding phrase pronunciations to a language model, the system comprising:

a computer with a computer code mechanism for processing a list of pron components whose pronunciations differ when they occur in a phrase, assigning at least one pron to each pron component, determining the pronunciation of a first phrase by first tokenizing the first phrase by generating a list of tokens corresponding to the first phrase, then determining a pron for each of the list of tokens, then assembling the pronunciation of the first phrase based on a combination of each said pron, and adding the first phrase and the pronunciation of the first phrase to a language model;

a language model electronically accessible by said computer code mechanism; and

a tokenizer for generating a list of tokens corresponding to the said first phrase, said tokenizer in electronic communication with said computer code mechanism.

14. The system of claim 13, wherein said pron components list includes non-initial components.

15. The system of claim 13, wherein said pron components list includes initial components.

16. The system of claim 13, further comprising a background dictionary electronically accessible by said computer code mechanism, wherein said computer code mechanism searches said background dictionary to determine a pron for each token.

17. The system of claim 13, further comprising a pron guesser in electronic communication with said computer code mechanism, wherein said computer code mechanism applies said pron guesser to determine a pron for each token.