Signal conditioning apparatus

• US 5,839,099 A
• Filed: 06/11/1996
• Issued: 11/17/1998
• Est. Priority Date: 06/11/1996
• Status: Expired due to Term

• Current Assignee: Callahan Cellular LLC (Intellectual Ventures LLC)
• Original Assignee: Guvolt, Inc.
• Inventors: Munsell, Clyde Stephan, Schaefer, Albert Russell, Cohen, Rodney Guy, Cover, John Richard
• Primary Examiner(s): Dorvil, Richemond
• Application Number:
  US08/661,542
• Time in Patent Office: 889 Days
• Field of Search: 395/2.16, 395/2.55, 395/2.62, 395/2.57, 395/2.33, 704/275, 704/207, 704/243, 704/251, 704/253, 704/233, 704/254, 704/205, 704/267, 704/270, 704/246
• US Class Current: 704/207
• CPC Class Codes: G10L 15/065 Adaptation

Abstract

An signal conditioning apparatus and method to condition a human voice signal for input to a time-domain voice recognition system. The signal conditioner normalizes sampled human voice inputs so that all inputs result in an output with substantially the same pitch and frequency bandwidth characteristics. The signal conditioner includes a pitch altering circuit that alters the pitch of the input human voice signal and a frequency band limiting circuit that limits the range of spectral information contained in a signal. The pitch altering circuit converts the input into a digital signal at a first conversion rate and converts the digital signal back to an analog signal at a second rate which is unequal to the first rate. The frequency band limit circuit is a plurality of filters at different points in the frequency spectrum.

Claims

1. An apparatus for producing a spectrally normalized human voice signal from a human voice input, said spectrally normalized human voice signal being sent to a voice recognition system, said apparatus comprising:

a pitch altering circuit receiving said human voice input and producing an intermediate analog output;

a frequency bandwidth limiting circuit receiving said intermediate analog output and producing said spectrally normalized human voice signal; and

a means for receiving a control signal from the voice recognition system to control a characteristic of said normalized human voice signal.

2. The apparatus of claim 1 wherein said pitch altering circuit further comprises:

an analog to digital converter having a first conversion rate, said analog to digital converter receiving said human voice input and producing a bit stream; and

a digital to analog converter receiving said bit stream and producing said intermediate output, said digital to analog converter having a second conversion rate, said second conversion rate unequal respective to said first conversion rate, thereby altering the pitch of said intermediate output respective to said human voice input.

3. The apparatus of claim 2 further comprising a rate control circuit for determining said first conversion rate and said second conversion rate.

4. The apparatus of claim 3 wherein said second conversion rate is a variable second conversion rate.

5. The apparatus of claim 2 wherein said second conversion rate is a fixed second conversion rate.

6. The apparatus of claim 1 wherein said frequency bandwidth limiting circuit further comprises:

a low frequency bandpass filter having a first frequency range, said low frequency bandpass filter receiving said intermediate output and producing a low frequency signal;

a high frequency bandpass filter having a second frequency range, said high frequency bandpass filter receiving said intermediate output and producing a high frequency signal; and

a mixer circuit for combining said low frequency signal and said high frequency signal to produce said spectrally normalized human voice signal having a narrowed frequency bandwidth.

7. The apparatus of claim 6 wherein said low frequency bandpass filter has a predominant frequency range between 20 Hertz and 60 Hertz.

8. The apparatus of claim 6 wherein said high frequency bandpass filter has a predominant frequency range between 14,000 Hertz and 18,000 Hertz.

9. The apparatus of claim 6 wherein said low frequency bandpass filter has a predominant frequency range of 40 Hertz.

10. The apparatus of claim 6 wherein said high frequency bandpass filter has a predominant frequency range of 16,000 Hertz.

11. The apparatus of claim 6 further comprising:

a first predetermined control signal for setting said first frequency range; and

a second predetermined control signal for setting said second frequency range.

12. The apparatus of claim 6 further comprising:

a first variable control signal for selecting said first frequency range; and

a second variable control signal for selecting said second frequency range.

13. The apparatus of claim 6 further comprising:

an amplitude control signal for controlling amplitude of said conditioned human voice signal.

14. A method for producing a conditioned human voice signal for input to a voice recognition system, said method comprising the steps of:

converting a human voice input to an audio signal, said audio signal having a pitch characteristic and a frequency bandwidth characteristic;

altering said pitch characteristic of said audio signal;

narrowing said frequency bandwidth characteristic of said audio signal to produce said conditioned human voice signal; and

receiving a signal from the voice recognition system to control a characteristic of the audio signal.

15. The method of claim 14 wherein said pitch characteristic is a first pitch characteristic and the step of altering said first pitch further comprises the steps of:

converting said audio signal to a digital signal, said digital signal comprising a plurality of bits, said plurality of bits produced at a first rate; and

converting said plurality of bits to an analog signal having a second pitch, said analog signal produced at a second rate, said second rate unequal respective to said first rate thereby said second pitch characteristic being unequal to said first pitch characteristic.

16. The method of claim 15 wherein the step of narrowing said frequency bandwidth characteristic further comprises the steps of:

filtering said analog signal to produce a low frequency signal;

filtering said analog signal to produce a high frequency signal; and

combining said low frequency signal and said high frequency signal to produce said conditioned human voice signal with limited frequency bandwidth characteristic.

17. The method of claim 16 wherein the step of filtering said analog signal to produce said low frequency signal, said low frequency signal having a predominant frequency range between 20 Hertz and 60 Hertz.

18. The method of claim 16 wherein the step of filtering said analog signal to produce said high frequency signal, said high frequency signal having a predominant frequency range between 14,000 Hertz and 18,000 Hertz.

19. The method of claim 16 wherein the step of filtering said analog signal to produce said low frequency signal, said low frequency having a predominant frequency range of 40 Hertz.

20. The method of claim 16 wherein said step of filtering said analog signal to produce said high frequency signal, said high frequency signal having a predominant range of 16,000 Hertz.

21. The method of claim 15 wherein said second rate is predetermined.

22. The method of claim 15 wherein said second rate is variable.

23. An apparatus for normalizing a human voice signal for input into a voice recognition system, said apparatus comprising:

an input device receiving said human voice input and producing a first intermediate analog signal;

a pitch altering circuit receiving said first intermediate analog signal and producing a second intermediate analog signal;

a frequency bandwidth limiting circuit receiving said second intermediate analog signal and producing said normalized human voice signal for input to said voice recognition system; and

a means for receiving a control signal from said voice recognition system to control a characteristic of said normalized human voice signal.

24. The apparatus of claim 23 wherein said input device is a microphone.

25. The apparatus of claim 23 wherein said input device is a playback device.