System and method for phonetic representation

• US 20090043584A1
• Filed: 08/06/2007
• Published: 02/12/2009
• Est. Priority Date: 08/06/2007
• Status: Abandoned Application

• Current Assignee: Lawrence Brooke Frank Philips
• Original Assignee: Lawrence Brooke Frank Philips
• Inventors: Philips, Lawrence Brooke Frank
• Application Number:
  US11/890,334
• Publication Number:
  US 20090043584A1
• Time in Patent Office: Days
• US Class Current: 704/260
• 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

A method for generating an Approximate Phonetic Representation (APR) of a given word, the word having a sequence of characters, the method comprising: Receiving the word; Generating the APR by applying at least one metaphone3 translation rule to encode one or more of the characters of the given word into a resulting APR; and Returning either the generated APR and/or one or more words matching the APR from a dictionary of words.

Claims

1. A method for generating an Approximate Phonetic Representation (APR) of a given word, the word having a sequence of characters, the method comprising:

Receiving the word;

Generating the APR by applying at least one unique metaphone3 translation rule to encode one or more of the characters of the given word into a resulting APR; and

Returning the generated APR.

2. The Method of claim 1, wherein the Generating further includes applying zero or more public-domain translation rules.

3. The Method of claim 1, wherein the method further comprises:

receiving a Vowel Encoding Setting with a value or either Initial-Only or All-Vowels; and

the Generating further comprises encoding only an initial vowel character when the value is Initial-Only, and encoding each vowel character when the value is All-Vowels.

4. The method of claim 1, wherein the metaphone3 translation rules are selected from one or more metaphone3 Consonant vowel transposition rules.

5. The method of claim 1, wherein the metaphone3 translation rules are selected from one or more metaphone3 Familiar non-English rules.

6. The method of claim 1, wherein the metaphone3 translation rules are selected from one or more metaphone3 Vowel following consonant rules.

7. The method of claim 1, wherein the metaphone3 translation rules are selected from one or more metaphone3 H following consonant rules.

8. The method of claim 1, wherein the metaphone3 translation rules are selected from one or more metaphone3 Unpronounced consonant rules.

9. The method of claim 1, wherein the metaphone3 translation rules are selected from one or more metaphone3 Letter group rules.

10. The method of claim 1, wherein the metaphone3 translation rules are selected from one or more metaphone3 Exception Rules.

11. The method of claim 1, wherein the Generating the APR further comprises applying at least one rule from each of the following metaphone3 rules sets:

Consonant Vowel Transposition Rule set and Vowel Following Consonant Rule set.

12. The method of claim 1, further comprising receiving a dictionary of words, the generating an APR of each word in the dictionary, returning the generated APR for each of the words in the dictionary.

13. The method of claim 1, further comprising receiving a dictionary of words, the generating an APR of each word in the dictionary, returning the generated APR for each of the words in the dictionary.

14. The method of claim 13, wherein the generated APR for each of the words in the dictionary is stored on a computer.

15. A method for generating an Approximate Phonetic Representation (APR) of a given word, the word having a sequence of characters, the method comprising:

Receiving the word;

Generating the APR by applying at least one metaphone3 translation rule to encode one or more of the characters of the given word into a resulting APR; and

Returning at least on Word corresponding to the resulting APR.

16. The Method of claim 15, wherein the Generating further includes applying at least one public-domain translation rules.

17. The Method of claim 15, wherein the method further comprises:

receiving a Vowel Encoding Setting with a value or either Initial-Only or All-Vowels; and

the Generating further comprises encoding only an initial vowel character when the value is Initial-Only, and encoding each vowel character when the value is All-Vowels.

18. The Method of claim 17, wherein the Generating the APR further comprises encoding at least one non-initial vowel character of the given word into the resulting APR.

19. A digital data processing apparatus for identifying one or more dictionary word corresponding with at least one approximate phonetic representation of a given word, said apparatus comprising:

An input means for accepting the given word;

A rule storage means for storing unique Metaphone3 rules;

An APR generation means coupled with said input means and the rule storage means for generating an APR representation of the given word;

A database storage means for storing a approximate phonetic representation corresponding to each of the one or more dictionary word, said database storage means including a storage means for storing digitally encoded information representative of said phonetic expression;

A database access means connected with said APR generation means and said database storage means for accessing the one or more dictionary word corresponding to the APR generated by the APR generation means; and

An output means connected with said database access means for generating a signal representative of the one or more dictionary word corresponding to the APR generated.

20. The digital data processing apparatus of claim 19, wherein the rule storage means further comprises a means for storing at least one public-domain translation rules.