System for developing word-pronunciation pairs

• US 6,363,342 B2
• Filed: 12/18/1998
• Issued: 03/26/2002
• Est. Priority Date: 12/18/1998
• Status: Expired due to Fees

• Current Assignee: Matsushita Electric Industrial Company Limited (Panasonic Holdings Corporation)
• Original Assignee: Matsushita Electric Industrial Company Limited (Panasonic Holdings Corporation)
• Inventors: Shaw, Rhonda, Pearson, Steve, Kuhn, Roland
• Primary Examiner(s): Korzuch, William
• Assistant Examiner(s): Azad, Abul K.
• Application Number:
  US09/216,111
• Publication Number:
  US 20020013707A1
• Time in Patent Office: 1,194 Days
• Field of Search: 704/244, 704/245, 704/243, 704/251, 704/252, 704/253, 704/254, 704/258, 704/260, 704/257, 704/220
• US Class Current: 704/220
• CPC Class Codes: G10L 15/063 Training
• CPC Class Codes: G10L 2015/025 Phonemes, fenemes or fenones being the recognition units
• CPC Class Codes: G10L 2015/0638 Interactive procedures

Abstract

An editing tool is provided for developing word-pronunciation pairs based on a spelled word input. The editing tool includes a transcription generator that receives the spelled word input from the user and generates a list of suggested phonetic transcriptions. The editor displays the list of suggested phonetic transcriptions to the user and provides a mechanism for selecting the desired pronunciation from the list of suggested phonetic transcriptions. The editing tool further includes a speech recognizer to aid the user in selecting the desired pronunciation from the list of suggested phonetic transcriptions based on speech data input that corresponds to the spelled word input, and a syllable editor that enables the user to manipulate a syllabic part of a selected pronunciation. Lastly, the desired pronunciation can be tested at any point through the use of a text-to-speech synthesizer that generates audible speech data for the selected phonetic transcription.

Claims

1. A system for developing word-pronunciation pairs based on a spelled word input, comprising:

a transcription generator receptive of the spelled word input for generating a phonetic transcription that corresponds to the spelled word input, said phonetic transcription being segmented into syllabic portions;

a phoneticizer receptive of the spelled word input for producing a plurality of scored phonetic transcriptions, where the phoneticizer employs letter-only decision trees and phoneme-mixed decision trees to produce said plurality of scored phonetic transcriptions, the letter-only decision trees having nodes representing questions about a given letter and neighboring letters in the spelled word input and the phoneme-mixed decision trees having nodes representing questions about a phoneme and neighboring phonemes in the spelled word input;

a phonemic editor connected to said transcription generator for displaying and editing syllabic portions of said phonetic transcription and connected to said phoneticizer for displaying said plurality of scored phonetic transcriptions; and

a storage mechanism for updating a lexicon with the spelled word input and said phonetic transcription, thereby developing the desired word-pronunciation pair.

2. The system of claim 1 wherein said transcription generator accesses a dictionary to generate said phonetic transcription, the dictionary storing phonetic transcription data corresponding to a plurality of spelled words.

3. The system of claim 1 wherein said transcription generator using letter-to-sound rules to produce said phonetic transcription.

4. The system of claim 1 wherein said phonetic transcription further includes accentuation data for the spelled word input.

5. The system of claim 4 wherein said dictionary storing accentuation data corresponding to each of the plurality of spelled words and said phonemic editor being operative to display and edit the accentuation data associated with said phonetic transcription.

6. The system of claim 1 wherein said phonemic editor provides a language selection mechanism and said transcription generator being connected to a plurality of dictionaries each of which stores phonetic transcription data in a different language, whereby said transcription generator invokes one of said plurality of dictionaries to produce a phonetic transcription that corresponds to the language from said language selection mechanism.

7. The system of claim 1 further includes a pronunciation selection mechanism connected to said phonemic editor for selecting at least one of said plurality of scored phonetic transcriptions.

8. The system of claim 7 wherein said pronunciation selection mechanism provides at least one of said plurality of scored phonetic transcriptions for updating said decision trees.

9. The system of claim 1 wherein the spelled word input and said phonetic transcription stored in said lexicon being used to retrain said transcription generator.

10. The system of claim 1 further includes a speech recognizer connected to said phonemic editor and receptive of speech data corresponding to the spelled word input for rescoring said plurality of scored phonetic transcriptions based on said speech data.

11. The system of claim 1 further includes a speech recognizer receptive of speech data corresponding to the spelled word input and being operative to produce the spelled word input, whereby said transcription generator receptive of the spelled word input from said speech recognizer.

12. The system of claim 1 further includes a speech recognizer receptive of speech data for producing a sounds-like word corresponding to the speech data, such that said phonemic editor being operative to provide a sounds-like phonetic transcription that corresponds to the sounds-like word and replace at least one syllabic portion of said phonetic transcription with said sounds-like phonetic transcription.

13. The system of claim 1 further includes a text-to-speech synthesizer connected to said phonemic editor and receptive of said phonetic transcription for generating speech data.

14. A system for developing word-pronunciation pairs based on a spelled word input, comprising:

a dictionary for storing phonetic transcription data corresponding to a plurality of spelled words;

a transcription generator connected to said dictionary and receptive of the spelled word input for producing a phonetic transcription that corresponds to the spelled word input, said phonetic transcription being segmented into syllabic portions;

a phoneticizer receptive of the spelled word input for producing a plurality of scored phonetic transcriptions, where the phoneticizer employs letter-only decision trees and phoneme-mixed decision trees to produce said plurality of scored phonetic transcriptions, the letter-only decision trees having nodes representing questions about a given letter and neighboring letters in the spelled word input and the phoneme-mixed decision trees having nodes representing questions about a phoneme and neighboring phonemes in the spelled word input; and

a phonemic editor connected to said transcription generator for displaying and editing syllabic portions of said phonetic transcription and connected to said phoneticizer for displaying said plurality of scored phonetic transcriptions, thereby developing the desired word-pronunciation pair.

15. The system of claim 14 wherein said transcription generator being operative to produce said phonetic transcription using letter-to-sound rules.

16. The system of claim 14 further includes a storage mechanism for updating a lexicon with the spelled word and said phonetic transcription.

17. The system of claim 14 wherein said phonetic transcription further includes accentuation data for the spelled word input.

18. The system of claim 17 wherein said dictionary storing accentuation data corresponding to each of the plurality of spelled words and said phonemic editor being operative to display and edit the accentuation data associated with said phonetic transcription.

19. The system of claim 14 wherein the spelled word and said phonetic transcription being used to retrain said transcription generator.

20. The system of claim 14 wherein said phonemic editor provides a language selection mechanism and said transcription generator being connected to a plurality of dictionaries each of which stores phonetic transcription data in a different language, whereby said transcription generator invokes one of said plurality of dictionaries to produce a phonetic transcription that corresponds to the language from said language selection mechanism.

21. The system of claim 14 further includes a pronunciation selection mechanism connected to said phonemic editor for selecting at least one of said plurality of scored phonetic transcriptions.

22. The system of claim 21 wherein said pronunciation selection mechanism provides at least one of said plurality of scored phonetic transcriptions for updating said decision trees.

23. The system of claim 14 further includes a speech recognizer connected to said phonemic editor and receptive of speech data corresponding to the spelled word input for rescoring said plurality of scored phonetic transcriptions based on said speech data.

24. The system of claim 14 further includes a speech recognizer receptive of speech data corresponding to the spelled word input and being operative to produce the spelled word input, whereby said transcription generator receptive of the spelled word input of said speech recognizer.

25. The system of claim 14 further includes a speech recognizer receptive of speech data for producing a sounds-like word corresponding to the speech data, such that said phonemic editor being operative to provide a sounds-like phonetic transcription that corresponds to the sounds-like word and replace at least one syllabic portion of said phonetic transcription with said sounds-like phonetic transcription.

26. The system of claim 14 further includes a text-to-speech synthesizer connected to said phonemic editor and receptive of said phonetic transcription for generating speech data.