Voice recording tool for creating database used in text to speech synthesis system

• US 7,890,330 B2
• Filed: 12/30/2006
• Issued: 02/15/2011
• Est. Priority Date: 12/30/2005
• Status: Active Grant

• Current Assignee: Alpine Electronics Incorporated (Alps Alpine Co., Ltd.)
• Original Assignee: Alpine Electronics Incorporated (Alps Alpine Co., Ltd.)
• Inventors: Arthur, William, Ozkaragoz, Inci, Ao, Benjamin
• Primary Examiner(s): Chawan; Vijay B
• Application Number:
  US11/647,825
• Publication Number:
  US 20070203704A1
• Time in Patent Office: 1,508 Days
• Field of Search: 704/260, 704/257, 704/9, 704/251, 704/255, 704/270, 704/275, 704/258, 704/2, 704/8, 704/266, 704/273, 434/169, 434/156, 434/178, 455/168.1, 348/468, 715/236, 715/250, 715/258, 715/210, 715/234, 707/999.006, 707/999.004
• US Class Current: 704/266
• CPC Class Codes: G01C 21/3629 Guidance using speech or audio output, e.g. text-to-speech text to speech systems per se G10L13/00
• CPC Class Codes: G10L 13/06 Elementary speech units used in speech synthesisers; Concatenation rules

Abstract

A method records verbal expressions of a person for use in a vehicle navigation system. The vehicle navigation system has a database including a map and text describing street names and points of interest of the map. The method includes the steps of obtaining from the database text of a word having at least one syllable, analyzing the syllable with a greedy algorithm to construct at least one text phrase comprising each syllable, such that the number of phrases is substantially minimized, converting the text phrase to at least one corresponding phonetic symbol phrase, displaying to the person the phonetic symbol phrase, the person verbally expressing each phrase of the phonetic symbol phrase, and recording the verbal expression of each phrase of the phonetic symbol phrase.

Claims

1. A method for recording verbal expressions of a person implemented in a concatenative speech synthesis system, the method comprising the following steps of:

designing a system with syllable-in-context constructs where all syllables of a language are represented, and if a syllable is bounded by vowel on either side, a phonetic context of the vowel is encoded;

creating a database of text phrases in which each text phrase is represented as a sequence of syllable-in-context constructs;

using a proprietary method to sort all of the text phrases based on numbers of new syllable-in-context constructs that each of the text phrases contains;

displaying selected phrases one at a time through a proprietary speech recording software program for a voice talent in order to produce speech data of the selected phrases; and

recording the verbal expression one text phrase at a time while checking a speech signal to be within a predetermined normal range so that it is not clipped or distorted at either side;

wherein the syllable includes a regular syllable having exactly one vowel preceded and/or followed by any number of consonants, and a super-syllable having two or more syllables with a vowel hiatus across a syllable boundary.

2. The method for recording verbal expressions of claim 1, wherein the verbal expression describes at least one street name and one point of interest.

3. The method for recording verbal expressions of claim 1, further comprises a step of comparing an amplitude of the recorded verbal expression to a predefined amplitude range to determine if the amplitude is within the predefined amplitude range.

4. The method for recording verbal expressions of claim 3, wherein when the amplitude of the recorded verbal expression is within the predefined amplitude range, then the method further comprises a step of storing in a verbal expression database the recorded verbal expression.

5. The method for recording verbal expressions of claim 3, wherein when the amplitude of the recorded verbal expression is outside of the predefined amplitude range, then the method further comprises the following steps of:

displaying to the person the at least one phonetic symbol phrase;

the person verbally expressing each phrase of the at least one phonetic symbol phrase; and

recording the verbal expression of each phrase of the at least one phonetic symbol phrase.

6. The method for recording verbal expressions of claim 1, further comprises a step of recording a first blank period, wherein the first blank period is substantially silent.

7. The method for recording verbal expressions of claim 6, further comprises a step of recording a second blank period, wherein the second blank period is substantially silent.

8. The method for recording verbal expressions of claim 7, wherein a length of each of the first blank period and the second blank period is at least one millisecond.

9. The method for recording verbal expressions of claim 7, wherein the step of recording the first blank period occurs prior to the step of recording the verbal expression of each phrase of the at least one phonetic symbol phrase and the step of recording the second blank period occurs after the step of recording the verbal expression of each phrase of the at least one phonetic symbol phrase.

10. The method for recording verbal expressions of claim 1, further comprises a step of indexing the recording of the verbal expression to a description of the recording and storing the indexed recording of the verbal expression in a verbal expression database.

11. The method for recording verbal expressions of claim 10, wherein the description of the recording comprises the text of each phrase of the verbal expression.

12. The method for recording verbal expressions of claim 11, wherein the description of the recording comprises phonetic symbols of each phrase of the verbal expression.

13. A method for recording verbal expressions of a person implemented in a concatenative speech synthesis system for use in a vehicle navigation system having a database including a map and text describing street names and points of interest of the map, the method comprising the following steps of:

designing a system with syllable-in-context constructs where all syllables of a language are represented, and if a syllable is bounded by vowel on either side, a phonetic context of the vowel is encoded;

creating a database of text phrases in which each text phrase is represented as a sequence of syllable-in-context constructs;

using a proprietary method to sort all of the text phrases based on numbers of new syllable-in-context constructs that each of the text phrases contains;

displaying selected phrases one at a time through a proprietary speech recording software program for a voice talent in order to produce speech data of the selected phrases; and

recording the verbal expression one text phrase at a time while checking a speech signal to be within a predetermined normal range so that it is not clipped or distorted at either side;

wherein the syllable includes a regular syllable having exactly one vowel preceded and/or followed by any number of consonants, and a super-syllable having two or more syllables with a vowel hiatus across a syllable boundary.

14. The method for recording verbal expressions of claim 13, further comprises a step of comparing an amplitude of the recorded verbal expression of each phrase to a predefined amplitude range to determine if the amplitude is within the predefined amplitude range.

15. The method for recording verbal expressions of claim 14, wherein when the amplitude of the recorded verbal expression of each phrase is within the predefined amplitude range, then the method further comprises a step of storing in a verbal expression database the recorded verbal expression of each phrase.

16. The method for recording verbal expressions of claim 13, further comprises a step of indexing the recording of the verbal expression of each phrase to a description of the recording and storing the indexed recording of the verbal expression of each phrase in a verbal expression database.