Speech synthesis system and its method

• US 5,848,390 A
• Filed: 02/02/1995
• Issued: 12/08/1998
• Est. Priority Date: 02/04/1994
• Status: Expired due to Term

• Current Assignee: Fujitsu Limited
• Original Assignee: Fujitsu Limited
• Inventors: Matsumoto, Tatsuro
• Primary Examiner(s): Hudspeth, David R.
• Assistant Examiner(s): Opsasnick, Michael N.
• Application Number:
  US08/382,531
• Time in Patent Office: 1,405 Days
• Field of Search: 381/51, 381/52, 381/53, 395/2.14, 395/2.16, 395/2.38, 395/2.39, 395/2.44, 395/2.67, 395/2.69, 395/2.7, 395/2.77, 395/2.79, 395/2.82, 395/2.84, 704/205, 704/207, 704/229, 704/236, 704/235, 704/258, 704/260, 704/261, 704/268, 704/270, 704/273, 704/275
• US Class Current: 704/260
• CPC Class Codes: G06F 11/34 Recording or statistical evaluation of computer activity, e.g. of down time, of input/output operation ; Recording or statistical evaluation of user activity, e.g. usability assessment
• CPC Class Codes: G06F 11/3409 for performance assessment
• CPC Class Codes: G10L 13/04 Details of speech synthesis systems, e.g. synthesiser structure or memory management

Abstract

A speech synthesis system has a word lexicon stored in a storage device for managing word information, a speech lexicon stored in the storage device for managing speech data. A language processing module carries out language processing of an input text by using the word information, and an acoustic processing module generates a synthesized speech signal by using the speech data in response to a processing result of the language processing module. A D/A converter converts the synthesized speech signal to an analog signal. The system further includes a detection module detecting performance or conditions of a hardware part of a computer in which the system is implemented, and an adjustment module adjusting the word information and/or the speech data which are to be given to the language processing module and the acoustic processing module, based on a detection result of the detection module. Therefore, an optimum speech-synthesis processing suitable for any computer is performed.

Claims

1. A speech synthesis system having a word lexicon stored in a storage device for managing word information, a speech lexicon stored in the storage device for managing speech data, a language processing module carrying out language processing of an input text by using said word information, an acoustic processing module generating a synthesized speech signal by using said speech data in response to a processing result of the language processing module, and a D/A converter converting the synthesized speech signal to an analog signal, the system comprising:

a detection module detecting, before inputting of the input text to be speech-synthesized and without depending thereupon, at least one of performance and conditions of a hardware part of a computer in which said speech synthesis system is implemented; and

an adjustment module adjusting at least one of said word information and said speech data to be given to said language processing module and said acoustic processing module, based on a detection result of said detection module to make speech-synthesis processing suitable for detected computer hardware conditions.

2. The speech synthesis system as claimed in claim 1, wherein said detection module comprises a portion or a whole of:

a CPU-function detection module detecting CPU performance or CPU load;

a memory-function detection module detecting a memory capacity or a memory-using situation;

a storage-device-function detection module detecting access performance of said storage device; and

a D/A-converter-function detection module detecting D/A converting performance of said D/A converter.

3. The speech synthesis system as claimed in claim 1, further comprising a bit adjustment module adjusting bit data of said synthesized speech signal generated in said acoustic processing module according to a number of quantization bits of said D/A converter detected by a D/A-converter-function detection module, when the number of the quantization bits of the D/A converter is not identical to a number of quantization bits of the synthesized speech signal generated in the acoustic processing module.

4. The speech synthesis system as claimed in claim 2, further comprising a bit adjustment module adjusting bit data of said synthesized speech signal generated in said acoustic processing module according to a number of quantization bits of said D/A converter detected by said D/A-converter-function detection module, when the number of the quantization bits of the D/A converter is not identical to a number of quantization bits of the synthesized speech signal generated in the acoustic processing module.

5. The speech synthesis system as claimed in claim 1, further comprising a frequency adjustment module adjusting a sampling frequency of said synthesized speech signal generated in said acoustic processing module according to a sampling frequency of said D/A converter detected by a D/A-converter-function detection module, when the sampling frequency of the D/A converter is not identical to a sampling frequency of the synthesized speech signal generated in the acoustic processing module.

6. The speech synthesis system as claimed in claim 2, further comprising a frequency adjustment module adjusting a sampling frequency of said synthesized speech signal generated in said acoustic processing module according to a sampling frequency of said D/A converter detected by said D/A-converter-function detection module, when the sampling frequency of the D/A converter is not identical to a sampling frequency of the synthesized speech signal generated in the acoustic processing module.

7. The speech synthesis system as claimed in claim 1, wherein said adjustment module adjusts said word information for said language processing module by adjusting an amount of supply of the word information.

8. The speech synthesis system as claimed in claim 7, wherein said adjustment module determines said word information in the word lexicon by selecting between one of a plurality of word lexicons and a word appearance frequency, and supplies selected word information to the language processing module.

9. The speech synthesis system as claimed in claim 1, wherein said adjustment module adjusts the speech data for said acoustic processing module, by using means of adjusting a number of quantization bits of the speech data, adjusting a sampling frequency of the speech data, adjusting an amount of supply of the speech data, or combinations thereof.

10. The speech synthesis system as claimed in claim 9, wherein said adjustment module, when the number of the quantization bits of the speech data is adjusted, selects, in the case that a plurality of speech lexicons are prepared with relation to different numbers of the quantization bits, one of the plurality of speech lexicons, speech data in a selected speech lexicon being directed to said acoustic processing module, or compresses or extends, in the case that only one speech lexicon is prepared, the speech data in the speech lexicon, compressed or extended speech data being directed to the acoustic processing module.

11. The speech synthesis system as claimed in claim 9, wherein said adjustment module, when the sampling frequency of the speech data is adjusted, selects, in the case that a plurality of speech lexicons are prepared with relation to different sampling frequencies, one of the plurality of speech lexicons, speech data in a selected speech lexicon being directed to said acoustic processing module, or samples, in the case that only one speech lexicon is prepared, the speech data in the speech lexicon in a high or a low frequency, the sampled speech data being directed to the acoustic processing module.

12. The speech synthesis system as claimed in claim 9, wherein said adjustment module, when the amount of supply of the speech data is adjusted, determines the speech data in the speech lexicon by selecting between one of a plurality of speech lexicons and a word appearance frequency, the selected speech data being directed to the acoustic processing module.

13. The speech synthesis system as claimed in claim 9, wherein said adjustment module, when the amount of supply of the speech data is adjusted, selects, in the case that a plurality of speech lexicons are prepared with relation to different tones, one or a plurality of the speech lexicons, the speech data in the selected speech lexicon being directed to the acoustic processing module.

24. The speech synthesis system as claimed in claim 1, wherein if the detection module detects that said D/A converter is not installed in this system or is not available, data of the synthesized speech signal generated in the acoustic processing module is stored in a storage device.

33. The speech synthesis system of claim 1, wherein said hardware conditions are detected and said adjustment module adjusted before start of the speech synthesis processing.

14. A speech synthesis system having a word lexicon and speech lexicon stored in a storage device, the word lexicon managing word information and the speech lexicon managing speech data, a language processing module carrying out language processing of an input text by using said word information, an acoustic processing module generating a synthesized speech signal by using said speech data in response to a processing result of the language processing module, and a D/A converter converting the synthesized speech signal to an analog signal, the system comprising:

a setting module by which a user can set at least one of performance and conditions of a hardware part of a computer in which said system is implemented, by a dialogic processing between the user and the computer, before inputting of the input text to be speech-synthesized and without depending thereupon; and

an adjustment module adjusting at least one of said word information and said speech data to be given to said language processing module and said acoustic processing module, based on a setting result of said setting module used in speech-synthesis processing that is suitable for detected computer hardware conditions.

15. The speech synthesis system as claimed in claim 14, further comprising a bit adjustment module adjusting bit data of said synthesized speech signal generated in said acoustic processing module according to a number of quantization bits of said D/A converter set by said setting module, when the number of the quantization bits of the D/A converter is not identical to a number of quantization bits of the synthesized speech signal generated in the acoustic processing module.

16. The speech synthesis system as claimed in claim 14, further comprising a frequency adjustment module adjusting a sampling frequency of said synthesized speech signal generated in said acoustic processing module according to a sampling frequency of said D/A converter set by said setting module, when the sampling frequency of the D/A converter is not identical to a sampling frequency of the synthesized speech signal generated in the acoustic processing module.

17. The speech synthesis system as claimed in claim 14, wherein said adjustment module adjusts said word information for said language processing module by adjusting an amount of supply of the word information.

18. The speech synthesis system as claimed in claim 17, wherein said adjustment module determines the word information in the word lexicon by selecting between one of a plurality of word lexicons and a word appearance frequency, and gives selected word information to said language processing module.

19. The speech synthesis system as claimed in claim 14, wherein said adjustment module adjusts the speech data for said acoustic processing module, by using means of adjusting a number of quantization bits of the speech data, adjusting a sampling frequency of the speech data, adjusting an amount of supply of the speech data, or combinations thereof.

20. The speech synthesis system as claimed in claim 19, wherein said adjustment module, when the number of the quantization bits of the speech data is adjusted, selects, in a case that a plurality of speech lexicons are prepared with relation to different numbers of the quantization bits, one of the plurality of speech lexicons, speech data in a selected speech lexicon being directed to said acoustic processing module, or compresses or extends, in the case that only one speech lexicon is prepared, the speech data in the speech lexicon, compressed or extended speech data being directed to the acoustic processing module.

21. The speech synthesis system as claimed in claim 19, wherein said adjustment module, when the sampling frequency of the speech data is adjusted, selects, in the case that a plurality of speech lexicons are prepared with relation to different sampling frequencies, one of the plurality of speech lexicons, speech data in a selected speech lexicon being directed to said acoustic processing module, or samples, in the case that only one speech lexicon is prepared, the speech data in the speech lexicon in a high or a low frequency, sampled speech data being directed to the acoustic processing module.

22. The speech synthesis system as claimed in claim 19, wherein said adjustment module, when the amount of supply of the speech data is adjusted, determines the speech data in the speech lexicon according to appearance frequency, selected speech data being directed to the acoustic processing module.

23. The speech synthesis system as claimed in claim 19, wherein said adjustment module, when the amount of supply of the speech data is adjusted, selects, in the case that a plurality of speech lexicons are prepared with relation to different tones, one or a plurality of the speech lexicons, the speech data in a selected speech lexicon being directed to the acoustic processing module.

34. The speech synthesis system of claim 14, wherein said hardware conditions are set and said adjustment module adjusted before start of the speech synthesis processing.

25. A speech synthesizing method, comprising the steps of:

(a) detecting, before inputting of an input text to be speech-synthesized and without depending thereupon, at least one of performance and conditions of a hardware part of a computer in which a speech synthesis system is constructed;

(b) adjusting, based on a detection result of said performance or said conditions, word information which is derived from a word lexicon and is given to a language processing module and speech data which is derived from a speech lexicon and is given to an acoustic processing module;

(c) performing language processing for speech synthesis of the input text by using adjusted word information; and

(d) generating a synthesized speech signal of the input text by using adjusted speech data, in response to a result of the language processing.

26. A speech synthesizing method, comprising the steps of:

(a) setting at least one of performance and conditions of a hardware part of a computer in which a speech synthesis system is constructed, by a dialogic processing between a user and the computer, before inputting of the input text to be speech-synthesized and without depending thereupon;

(b) adjusting, based on a setting result of said performance or said conditions, word information which is derived from a word lexicon and is given to a language processing module and speech data which is derived from a speech lexicon and is given to an acoustic processing module;

(c) performing language processing for speech synthesis of an input text by using adjusted word information; and

(d) generating a synthesized speech signal of the input text by using adjusted speech data, in response to a result of the language processing.

27. A storage medium storing information and data describing a sequence for controlling a speech synthesis system, the storage medium comprising:

a word lexicon managing word information;

a speech lexicon managing speech data;

a language processing module performing language processing for speech synthesis of an input text by using word information of the word lexicon;

an acoustic processing module generating a synthesized speech signal of the input text by using speech data of the speech lexicon, based on a processing result of the language processing module;

a detection module detecting, before inputting of the input text to be speech-synthesized and without depending thereupon, at least one of performance and conditions of a hardware part of a computer in which said system is implemented; and

an adjustment module adjusting lexicon information which is to be given to said language processing module and said acoustic processing module, in response to a detection result of said detection module.

28. A storage medium storing information and data describing a sequence for controlling a speech synthesis system, the storage medium comprising:

a word lexicon managing word information;

a speech lexicon managing speech data;

a language processing module performing language processing for speech synthesis of an input text by using word information of the word lexicon;

an acoustic processing module generating a synthesized speech signal of the input text by using speech data of the speech lexicon, based on a processing result of the language processing module;

a setting module by which a user can set at least one of performance and conditions of a hardware part of a computer in which said system is implemented, with a dialogic processing between the user and the computer, before inputting of the input text to be speech-synthesized and without depending thereupon; and

an adjustment module adjusting lexicon information which is to be given to said language processing module and said acoustic processing module, in response to a setting result of said setting module.

29. A speech synthesis system transforming text into an analog speech signal, comprising:

a language processing module coupled to an acoustic processing module; and

an adjustment module varying parameters of the language processing module and the acoustic processing module based on a measurement of hardware conditions, before inputting of the input text to be speech-synthesized and without depending thereupon.

30. The speech synthesis system of claim 29, further comprising a bit adjustment module adjusting bit quantization levels of a D/A converter based on the hardware conditions.

31. The speech synthesis system of claim 29, further comprising a frequency adjustment module adjusting a sampling frequency of a synthesized speech signal generated in the acoustic processing module according to hardware conditions.

32. The speech synthesis system of claim 29, wherein the measurement of hardware conditions is selected from the group consisting essentially of CPU Performance, CPU load, memory capacity, memory-using situation, access performance of a storage device, and D/A converting performance, all independent of data input rate.