SYSTEM AND METHOD FOR GENERATING A PRONUNCIATION DICTIONARY

• US 20070255567A1
• Filed: 04/27/2006
• Published: 11/01/2007
• Est. Priority Date: 04/27/2006
• Status: Abandoned Application

• Current Assignee: AT&T Corporation (AT&T, Inc.)
• Original Assignee: AT&T Corporation (AT&T, Inc.)
• Inventors: Bangalore, Srinivas, Schulz, David
• Application Number:
  US11/380,496
• Publication Number:
  US 20070255567A1
• Time in Patent Office: Days
• US Class Current: 704/260
• CPC Class Codes: G06F 40/242 Dictionaries
• CPC Class Codes: G06F 40/53 Processing of non-Latin text kana-to-kanji conversion G06F40/129; vowelisation G06F40/232
• 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

Disclosed is a system, method and computer-readable medium that stores instructions for controlling a computing device. The method embodiment relates to processing speech data wherein the method comprises generating phoneme transcriptions for words in a first language, generating a pronunciation dictionary comprising three-parts including, for each word in the dictionary, a first part having a first language orthography, a second part having a second language pronunciation and a third part having a second language orthography and applying the pronunciation dictionary in a speech application. The method is especially applicable to languages where the alphabet does not fully represent how a word is pronounced, as in Arabic or Hebrew.

Claims

1. A method of processing speech data, the method comprising:

generating phoneme transcriptions for words in a first language;

generating a pronunciation dictionary comprising three-parts including, for each word in the dictionary, a first part having a first language orthography, a second part having a second language pronunciation and a third part having a second language orthography; and

applying the pronunciation dictionary in a speech application.

2. The method of claim 1, wherein generating the pronunciation dictionary further comprises:

collecting names written in the first language alphabet;

using text-to-speech and name pronunciation software to generate phoneme transcriptions of the names; and

transliterating the phoneme transcriptions into the second language using rules.

3. The method of claim 1, wherein applying the pronunciation dictionary in a speech application comprises at least one of:

using the pronunciation dictionary to constrain possible ways a first language word is pronounced and predicting how a second language spelling of first language name will be pronounced.

4. The method of claim 1, wherein the second language is one of Arabic or Hebrew and the first language is one of Latin or English.

5. The method of claim 1, wherein the three-part dictionary comprises a database containing a second language orthographic string, at least one pronunciation variant, and the first-language spelling.

6. The method of claim 1, further comprising:

using the pronunciation dictionary to train second language letter-to-sound rules.

7. The method of claim 6, further comprising:

analyzing text using the trained letter-to-sound rules to predict a pronunciation for new words not seen before.

8. The method of claim 2, wherein the step of transliterating the phoneme transcriptions into the second language using rules further comprises transliterating the phoneme transcriptions by rule into a number of plausible variants in the second language.

9. A system for processing speech data, the system comprising:

a module configured to generate phoneme transcriptions for words in a first language;

a module configured to generate a pronunciation dictionary comprising three-parts including, for each word in the dictionary, a first part having a first language orthography, a second part having a second language pronunciation and a third part having a second language orthography; and

a module configured to apply the pronunciation dictionary in a speech application.

10. The system of claim 9, wherein the module configured to generate the pronunciation dictionary further:

collects names written in the first language alphabet;

uses text-to-speech and name pronunciation software to generate phoneme transcriptions of the names; and

transliterates the phoneme transcriptions into the second language using rules.

11. The system of claim 9, wherein the module configured to apply the pronunciation dictionary in a speech application further:

uses the pronunciation dictionary to constrain possible ways a first language word is pronounced or predicts how a second language spelling of first language name will be pronounced.

12. The system of claim 9, wherein the second language is one of Arabic or Hebrew and the first language is one of Latin or English.

13. The system of claim 9, wherein the three-part dictionary comprises a database containing a second language orthographic string, at least one pronunciation variant, and the first-language spelling.

14. The system of claim 9, further comprising:

a module configured to use the pronunciation dictionary to train second language letter-to-sound rules.

15. The system of claim 14, further comprising:

a module configured to analyze text using the trained letter-to-sound rules to predict a pronunciation for new words not seen before.

16. The system of claim 10, wherein the module configured to transliterate the phoneme transcriptions into the second language using rules further transliterates the phoneme transcriptions by rule into a number of plausible variants in the second language.

17. A computer-readable medium storing instructions for controlling a computing device to process speech data, the instructions comprising:

generating phoneme transcriptions for words in a first language;

generating a pronunciation dictionary comprising three-parts including, for each word in the dictionary, a first part having a first language orthography, a second part having a second language pronunciation and a third part having a second language orthography; and

applying the pronunciation dictionary in a speech application.

18. The computer-readable medium of claim 17, wherein generating the pronunciation dictionary further comprises:

collecting names written in the first language alphabet;

using text-to-speech and name pronunciation software to generate phoneme transcriptions of the names; and

transliterating the phoneme transcriptions into the second language using rules.

19. The computer-readable medium of claim 17, wherein applying the pronunciation dictionary in a speech application comprises at least one of:

using the pronunciation dictionary to constrain possible ways a first language word is pronounced and predicting how a second language spelling of first language name will be pronounced.

20. The computer-readable medium of claim 17, wherein the three-part dictionary comprises a database containing a second language orthographic string, at least one pronunciation variant, and the first-language spelling.

21. The computer-readable medium of claim 17, the instructions further comprising:

using the pronunciation dictionary to train second language letter-to-sound rules.

22. The computer-readable medium of claim 21, the instructions further comprising:

analyzing text using the trained letter-to-sound rules to predict a pronunciation for new words not seen before.

23. The computer-readable medium of claim 18, wherein the step of transliterating the phoneme transcriptions into the second language using rules further comprises transliterating the phoneme transcriptions by rule into a number of plausible variants in the second language.