Formant detecting device and speech processing apparatus

• US 5,479,560 A
• Filed: 10/27/1993
• Issued: 12/26/1995
• Est. Priority Date: 10/30/1992
• Status: Expired due to Fees

• Current Assignee: New Energy and Industrial Technology Development Organizatio
• Original Assignee: Technology Research Association Of Medical And Welfare Apparatus
• Inventors: Mekata, Tsuyoshi
• Primary Examiner(s): MacDonald, Allen R.
• Assistant Examiner(s): ONKA, THOMAS
• Application Number:
  US08/143,932
• Time in Patent Office: 790 Days
• Field of Search: 395/2.18, 395/2.35, 395/2.36, 395/2.37, 395/2.8, 381/39
• US Class Current: 704/209
• CPC Class Codes: G10L 21/0364 for improving intelligibility
• CPC Class Codes: G10L 25/15 the extracted parameters being formant information
• CPC Class Codes: G10L 25/27 characterised by the analysis technique
• CPC Class Codes: H04R 2225/43 Signal processing in hearing aids to enhance the speech intelligibility

Abstract

A speech processing apparatus for obtaining a processed speech which is natural and comfortable for a listener, by refining a gain value assigned for each frequency band in enhancing formants in a power spectrum. The power spectrum, calculated in a frequency analyzing unit, is subject to contrast enhancement in a contrast enhancing unit, and judged as to whether it is a format or not in each frequency band. In a gain value assigning unit, a gain value of 1 is assigned to a formant, and a gain value smaller than 1 is to a frequency other than formant. A threshold value for each frequency band is determined by a threshold value determining unit in accordance with power spectrum of input speech signal, to eliminate the effect of variation in speech level.

Claims

1. A formant detecting device comprising:

frequency analyzing means for calculating the power spectrum of an input speech signal;

contrast enhancing means for enhancing the contrast between a local maximum portion and a local minimum portion in said power spectrum of said input speech signal; and

single threshold value judging means for comparing the power in said power spectrum enhanced by said contrast enhancing means with a threshold value in each frequency band and for judging a frequency band corresponding to said power to be a formant if said power in said enhanced power spectrum exceeds said threshold value.

2. A formant detecting device according to claim 1, wherein said threshold value is predetermined so that a predefined first and a predefined second formant of each of a predetermined number of vocalized vowels are detected by said formant detecting device with probability of 50% or more.

3. A formant detecting device according to claim 1, further comprising threshold determining means for determining said threshold value in accordance with said power spectrum of said input speech signal.

4. A formant detecting device according to claim 3, wherein said threshold value determining means determines said threshold value in each frequency band so that said threshold value is equal to a product of a constant and the power at the corresponding frequency band of said power spectrum of said input speech signal.

5. A formant detecting device according to claim 4, further comprising constant changing means for changing said constant manually.

6. A formant detecting device according to claim 4, further comprising constant changing means for receiving a background noise level and for changing said constant as a function of said background noise level.

7. A formant detecting device according to claim 3, wherein said threshold value determining means determines said threshold value so that said threshold value is equal to the average power over all the frequency bands in said power spectrum of said input speech signal.

8. A formant detecting device comprising:

frequency analyzing means for calculating the power spectrum of an input speech signal;

contrast enhancing means for enhancing the contrast between a local maximum portion and a local minimum portion in said power spectrum of said input speech signal;

dividing means for dividing the power at each frequency band of said power spectrum enhanced by said contrast enhancing means by the power of said input speech signal in the corresponding frequency band;

threshold value judging means for comparing a divisional result obtained by said dividing means with a single threshold value in each frequency band and for judging a frequency band corresponding to said divisional result to be a formant if said divisional result exceeds said threshold value.

9. A speech processing apparatus comprising:

frequency analyzing means for calculating the power spectrum of an input speech signal;

contrast enhancing means for enhancing the contrast between a local maximum portion and a local minimum portion in said power spectrum of said input speech signal;

threshold value judging means for comparing the power in the power spectrum enhanced by the contrast enhancing means with a single threshold value in each frequency band and for judging a frequency band corresponding to said power to be a formant if said power in the enhanced power spectrum exceeds said threshold value;

gain value assigning means for assigning a first gain value to said frequency band judged to be a formant by said threshold judging means and for assigning a second gain value to other frequency bands; and

speech signal generating means for generating a speech signal having a power spectrum obtained by multiplying the power at each frequency band of said power spectrum of said input speech signal by the gain value assigned to that frequency band by said gain value assigning means.

10. A speech processing apparatus according to claim 9, wherein said frequency analyzing means further calculates the phase of said input speech signal, and said speech signal generating means further comprises:

multiplying means for multiplying the power at each frequency band of said power spectrum of said input speech signal by the gain value assigned to that frequency band by said gain value assigning means; and

inverse transformation means for transforming inversely a multiplicative result obtained by said multiplying means, and said phase of said input speech signal obtained by the frequency analyzing means into the speech signal.

11. A speech processing apparatus according to claim 9, wherein said speech signal generating means comprises frequency characteristics variable filter means for varying frequency characteristics of said input speech signal in accordance with one of said first gain value and said second gain value assigned by said gain value assigning means.

12. A speech processing apparatus according to claim 9, wherein said gain value assigning means has a plurality of candidate values for at least one of said first and second gain values, and said speech processing apparatus further comprises gain value switching means for switching at least one of said first and second gain values to one of said plurality of candidate values.

13. A speech processing apparatus according to claim 9, wherein said gain value assigning means has a plurality of candidate values for at least one of said first and second gain values, and said speech processing apparatus further comprises:

background noise level detecting means for detecting a background noise level from said input speech signal; and

gain value switching means for switching at least one of said first and second gain values to one of said plurality of candidate values.

14. A speech processing apparatus comprising:

frequency analyzing means for calculating the power spectrum of an input speech signal;

contrast enhancing means for enhancing the contrast between a local maximum portion and a local minimum portion in said power spectrum of said input speech signal;

dividing means for dividing the power at each frequency band of said power spectrum enhanced by said contrast enhancing means by the power of said input speech signal in the corresponding frequency band;

threshold value judging means for comparing a divisional result obtained by said dividing means with a single threshold value in each frequency band and for judging a frequency band corresponding to said divisional result to be a formant if said divisional result exceeds said threshold value;

gain value assigning means for assigning a first gain value to said frequency band judged to be a formant by said threshold judging means and for assigning a second gain value to other frequency bands; and

speech signal generating means for generating a speech signal having a power spectrum obtained by multiplying the power at each frequency band of said power spectrum of said input speech signal by the gain value assigned to that frequency band by said gain value assigning means.