Speech synthesis apparatus

• US 7,526,430 B2
• Filed: 09/15/2005
• Issued: 04/28/2009
• Est. Priority Date: 06/04/2004
• Status: Active Grant

• Current Assignee: Panasonic Intellectual Property Corporation of America (Panasonic Holdings Corporation)
• Original Assignee: Panasonic Corporation (Panasonic Holdings Corporation)
• Inventors: Kamai, Takahiro, Kato, Yumiko
• Primary Examiner(s): Chawan; Vijay B
• Assistant Examiner(s): Colucci; Michael C
• Application Number:
  US11/226,331
• Publication Number:
  US 20060009977A1
• Time in Patent Office: 1,321 Days
• Field of Search: 704/260, 704/275, 704/268, 704/273, 704/258, 700/83, 706/21, 715/716, 381/13
• US Class Current: 704/267
• CPC Class Codes: G10L 13/10 Prosody rules derived from text; Stress or intonation

Abstract

A speech synthesis apparatus, which can embed unchangeable additional information into synthesized speech without causing a deterioration of speech quality and restriction by bands, includes a language processing unit which generates synthesized speech generation information necessary for generating synthesized speech in accordance with a language string, a prosody generating unit which generates prosody information of speech based on the synthesized speech generation information, and a waveform generating unit which synthesizes speech based on the prosody information, in which the prosody generating unit embed code information as watermark information in the prosody information of a segment having a predetermined time duration within a phoneme length including a phoneme boundary.

Claims

1. A speech synthesis apparatus which synthesizes speech, said apparatus comprising:

a prosody generating unit for generating prosody information of the speech based on synthesized speech generation information; and

a synthesis unit for synthesizing the speech based on the prosody information,

wherein said prosody generation unit is for;

specifying a time position including a phoneme boundary in the speech to be synthesized into which a micro-prosody pattern is to be embedded, based on the synthesized speech generation information;

extracting a micro-prosody pattern from a storage unit, the micro-prosody pattern being a pattern of a fine time structure of prosody including the phoneme boundary; and

embedding the extracted micro-prosody pattern into the specified time position as watermark information, the embedded micro-prosody pattern indicating that the speech is synthesized speech.

2. The speech synthesis apparatus according to claim 1,

wherein a duration for embedding the extracted micro-prosody pattern is a duration in a range from 10 milliseconds to 50 milliseconds.

3. The speech synthesis apparatus according to claim 1, further comprising

an encoding unit for encoding additional information,

wherein said encoding unit is for encoding information for associating the micro-prosody pattern stored in said storage unit with the additional information, and

wherein said prosody generation unit is for selecting from the storage unit, based on the encoded information, the micro-prosody pattern associated with the additional information, and embedding the selected micro-prosody pattern into the specified time position including the phoneme boundary.

4. The speech synthesis apparatus according to claim 3,

wherein said encoding unit is further for generating key information which corresponds to the encoded information for decoding the additional information.

14. The speech synthesis apparatus according to claim 1,

wherein said prosody generating unit is for identifying, as the time position including the phoneme boundary in the speech to be synthesized, a portion of at least one vowel of:

a vowel which follows immediately after silence;

a vowel which follows immediately after a consonant other than a semivowel;

a vowel which immediately precedes silence; and

a vowel which immediately precedes a consonant other than a semivowel.

15. The speech synthesis apparatus according to claim 1,

wherein said prosody generating unit is for identifying, as the time position including the phoneme boundary in the speech to be synthesized, at least one of:

a portion, including a starting point of a phoneme, of a vowel which follows immediately after silence;

a portion, -including the starting point of the phoneme, of a vowel which follows immediately after a consonant other than a semivowel;

a portion, including an ending point of the phoneme, of a vowel which immediately precedes silence; and

a portion, including the ending point of the phoneme, of a vowel which immediately precedes a consonant other than a semivowel.

5. A synthesis speech identifying apparatus which identifies whether or not inputted speech is synthesized speech, said apparatus comprising:

a fundamental frequency calculating unit for calculating a speech fundamental frequency of the inputted speech on a per frame basis, each frame having a predetermined duration;

a storage unit in which a micro-prosody pattern is stored, the micro-prosody pattern being a pattern of a fine time structure of prosody including a phoneme boundary, and being used to identify the inputted speech as synthesized speech; and

an identifying unit for;

extracting, in a segment having a duration including a phoneme boundary within which a micro-prosody pattern of the inputted speech exists as watermark information, the fundamental frequency of the speech calculated by said fundamental frequency calculation unit;

matching a pattern of the extracted fundamental frequency with the micro-prosody pattern stored in said storage unit; and

identifying whether or not the inputted speech is synthesized speech.

6. An additional information reading apparatus which decodes additional information embedded in inputted speech, said apparatus comprising:

a fundamental frequency calculating unit for calculating a speech fundamental frequency of the inputted speech on a per frame basis, each frame having a predetermined duration;

a storage unit in which a micro-prosody pattern associated with the additional information is stored, the micro-prosody pattern being a pattern of a fine time structure of prosody including a phoneme boundary; and

an additional information extracting unit for;

extracting, in a segment having a duration including a phoneme boundary within which a micro-prosody pattern of the inputted speech exists as water mark information, a micro-prosody pattern from the speech fundamental frequency calculated by said fundamental frequency calculating unit;

comparing the extracted micro-prosody pattern with the micro-prosody pattern associated with the additional information; and

extracting predetermined additional information included in the extracted micro-prosody pattern.

7. The additional information reading apparatus according to claim 6,

wherein the additional information is encoded, and

said additional information reading apparatus further comprises

a decoding unit for decoding the encoded additional information using key information for decoding.

8. A speech synthesis method of synthesizing speech, comprising

generating prosody information of the speech based on synthesized speech generation information,

wherein said generating includes:

specifying a time position including a phoneme boundary in the speech to be synthesized into which a micro-prosody pattern is to be embedded, based on the synthesized speech generation information;

extracting a micro-prosody pattern from a storage unit, the micro-prosody pattern being a pattern of a fine time structure of prosody including the phoneme boundary; and

embedding the extracted micro-prosody pattern into the specified time position as watermark information, the embedded micro-prosody pattern indicating that the speech is synthesized speech.

9. The speech synthesis method according to claim 8,

wherein a duration for embedding the extracted micro-prosody pattern is a duration in a range from 10 milliseconds to 50 milliseconds.

10. A program embodied on a computer readable recording medium, for making a computer function as a speech synthesis apparatus, said program making the computer function as the following:

a prosody generating unit for generating prosody information of speech based on synthesized speech generation information; and

a synthesis unit for synthesizing the speech based on the prosody information,

wherein the prosody generating unit is for;

specifying a time position including a phoneme boundary in the speech to be synthesized into which a micro-prosody pattern is to be embedded, based on the synthesized speech generation information;

extracting a micro-prosody pattern from a storage unit, the micro-prosody pattern being a pattern of a fine time structure of prosody including the phoneme boundary; and

embedding the extracted micro-prosody pattern into the specified time position as watermark information, the embedded micro-prosody pattern indicating that the speech is synthesized speech.

11. The program embodied on a computer readable recording medium, according to claim 10,

wherein a duration for embedding the extracted micro-prosody pattern is a duration in a range from 10 milliseconds to 50 milliseconds.

12. A computer readable recording medium on which a program for making a computer function as a speech synthesis apparatus is recorded,

wherein said program makes a computer function as the following:

a prosody generating unit for generating prosody information of speech based on synthesized speech generation information; and

a synthesis unit for synthesizing the speech based on the prosody information,

wherein the prosody generating unit is for;

specifying a time position including a phoneme boundary in the speech to be synthesized into which a micro-prosody pattern is to be embedded, based on the synthesized speech generation information;

extracting a micro-prosody pattern from a storage unit, the micro-prosody pattern being a pattern of a fine time structure of prosody including the phoneme boundary; and

embedding the extracted micro-prosody pattern into the specified time position as watermark information, the embedded micro-prosody pattern indicating that the speech is synthesized speech.

13. The computer readable recording medium according to claim 12,

wherein a duration for embedding the extracted micro-prosody pattern is a duration in a range from 10 milliseconds to 50 milliseconds.

16. A speech synthesis apparatus which synthesizes speech, said apparatus comprising:

a prosody generating unit for generating prosody information of the speech based on synthesized speech generation information; and

a synthesis unit for synthesizing the speech based on the prosody information,

wherein said prosody generation unit is for;

specifying a time position in the speech to be synthesized into which a micro-prosody pattern is to be embedded, based on the synthesized speech generation information;

extracting a micro-prosody pattern from a storage unit, the micro-prosody pattern being a pattern of a fine time structure of prosody including a phoneme boundary; and

embedding the extracted micro-prosody pattern into the specified time position as watermark information, the embedded micro-prosody pattern indicating that the speech is synthesized speech, and the embedded micro-prosody pattern being used to identify a manufacturer of said speech synthesis apparatus.

17. A synthesis speech identifying apparatus which identifies whether or not inputted speech is synthesized speech, said apparatus comprising:

a fundamental frequency calculating unit for calculating a speech fundamental frequency of the inputted speech on a per frame basis, each frame having a predetermined duration;

a storage unit in which a micro-prosody pattern is stored, the micro-prosody pattern being a pattern of a fine time structure of prosody including a phoneme boundary, and the micro-prosody pattern being used to identify the inputted speech as synthesized speech and to identify a manufacturer of said speech synthesis apparatus that has generated the synthesized speech; and

an identifying unit for;

extracting, in a segment having a duration within which a micro-prosody pattern of the inputted speech exists as watermark information, the fundamental frequency of the speech calculated by said fundamental frequency calculation unit;

matching a pattern of the extracted fundamental frequency with the micro-prosody pattern stored in said storage unit; and

identifying whether or not the inputted speech is synthesized speech and, in the case where the inputted speech is synthesized speech, identify a manufacturer of said speech synthesis apparatus that has generated the synthesized speech.

18. An additional information reading apparatus which decodes additional information embedded in inputted speech, said apparatus comprising:

a fundamental frequency calculating unit for calculating a speech fundamental frequency of the inputted speech on a per frame basis, each frame having a predetermined duration;

a storage unit in which a micro-prosody pattern associated with the additional information is stored, the micro-prosody pattern being a pattern of a fine time structure of prosody including a phoneme boundary, and the micro-prosody pattern being used to identify a manufacturer of said speech synthesis apparatus; and

an additional information extracting unit for;

extracting, in a segment having a duration including a phoneme boundary within which a micro-prosody pattern of the inputted speech exists as watermark information, a micro-prosody pattern from the speech fundamental frequency calculated by said fundamental frequency calculating unit;

comparing the extracted micro-prosody pattern with the micro-prosody pattern associated with the additional information;

extracting predetermined additional information included in the extracted micro-prosody pattern; and

identifying a manufacturer of said speech synthesis apparatus that has generated the synthesized speech.