Ultra-narrow bandwidth voice coding

• US 7,162,415 B2
• Filed: 11/05/2002
• Issued: 01/09/2007
• Est. Priority Date: 11/06/2001
• Status: Expired due to Fees

• Current Assignee: Lawrence Livermore National Security LLC (Government of the United States of America)
• Original Assignee: Regents of the University of California (University of California)
• Inventors: Ng, Lawrence C., Holzrichter, John F.
• Primary Examiner(s): Hudspeth; David
• Assistant Examiner(s): Jackson; Jakieda R.
• Application Number:
  US10/288,992
• Publication Number:
  US 20030097254A1
• Time in Patent Office: 1,526 Days
• Field of Search: 704/201
• US Class Current: 704/201
• CPC Class Codes: G10L 19/00 Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis in musical instruments G10H

Abstract

A system of removing excess information from a human speech signal and coding the remaining signal information, transmitting the coded signal, and reconstructing the coded signal. The system uses one or more EM wave sensors and one or more acoustic microphones to determine at least one characteristic of the human speech signal.

Claims

1. A system for removing excess information from a human speech signal and coding the remaining signal information, comprising:

at least one EM wave sensor,

at least one acoustic microphone,

processing means for removing said excess information from said human speech signal that produces a remaining signal, and

processing means for coding said remaining signal to provide a coded signal;

wherein said processing means for removing said excess information and said processing means for coding said remaining signal uses said at least one EM wave sensor and said at least one acoustic microphone to determine at least one characteristic of said human speech signal.

2. The system for removing excess information from a human speech signal and coding the remaining signal information of claim 1 wherein said at least one characteristic of said human speech signal comprises an average glottal period time duration value of voiced speech.

3. The system for removing excess information from a human speech signal and coding the remaining signal information of claim 1 wherein said at least one characteristic of said human speech signal comprises an excitation function and its coded description.

4. The system of claim 3 wherein said excitation function comprises at least one of the following:

one numerically parameterized excitation function, one onset of excitation timing function, one directly measured excitation function, and at least one table lookup excitation function.

5. The system of claim 3 wherein said at least one acoustic microphone provides acoustic sensor signal information and said excitation function is time aligned with said acoustic sensor signal information.

6. The system for removing excess information from a human speech signal and coding the remaining, signal information of claim 1 wherein said at least one characteristic of said human speech signal comprises time of onset, time duration, and time of end for each type of speech in a sequences of segments of said speech-types.

7. The system of claim 1, where said at least one characteristic includes at least 3 types of speech.

8. The system for removing excess information from a human speech signal and coding the remaining signal information of claim 1 wherein said at least one characteristic of said human speech signal comprises number of glottal periods and one or more spectral formant values within a continuous segment of voiced speech.

9. The system for removing excess information from a human speech signal and coding the remaining signal information of claim 1 wherein said at least one characteristic of said human speech signal comprises the type of unvoiced speech segment, and its amplitude compared to voiced speech.

10. The system for removing excess information from a human speech signal and coding the remaining signal information of claim 1 wherein said at least one characteristic of said human speech signal comprises header-information that describes speech properties of the user.

11. The system for removing excess information from a human speech signal and coding the remaining signal information of claim 1 wherein said at least one characteristic of said human speech signal comprises one or more of an average glottal-period'"'"'s time-duration value of voiced speech, an excitation function and its coded description, time of onset, time duration, and time of end for each of 3 types of speech in a sequences of segments of said speech-types, the number of glottal periods, variations in glottal period durations, and one or more spectral formant values within a continuous segment of voiced speech, the type of unvoiced speech segment, and its amplitude compared to voiced speech, and header-information that describes recurring speech properties of the user.

12. The system for removing excess information from a human speech signal and coding the remaining signal information of claim 1 wherein said at least one EM wave sensor comprises a coherent wave EM sensor.

13. The system for removing excess information from a human speech signal and coding the remaining signal information of claim 1 wherein said at least one EM wave sensor comprises a coherent wave EM sensor for measuring essential information comprised of air-pressure-induced tissue movement in the human vocal tract for purposes of glottal timing, excitation function description, and voiced speech segment onset times.

14. The system for removing excess information from a human speech signal and coding the remaining signal information of claim 1 wherein said at least one EM wave sensor comprises a coherent optical-frequency EM sensor for obtaining vocal tract wall movement by measuring surface motion of skin tissues connected to said vocal tract wall-tissues.

15. A method of removing excess information from a human speech signal and coding the remaining signal information, comprising the steps of:

producing a human speech signal using one or more EM wave sensors and one or more acoustic microphones,

using processing means for removing excess information from said human speech signal and producing a remaining signal, and

using processing means for coding said remaining signal to provide a coded signal to determine at least one characteristic of said human speech signal.

16. The method of removing excess information from a human speech signal and coding the remaining signal information of claim 15 wherein said at least one characteristic of said human speech signal comprises an average glottal period time duration value of voiced speech.

17. The method of removing excess information from a human speech signal and coding the remaining signal information of claim 15 wherein said at least one characteristic of said human speech signal comprises an excitation function and its coded description.

18. The method of removing excess information from a human speech signal and coding the remaining signal information of claim 15 wherein said at least one characteristic of said human speech signal comprises time of onset, time duration, and time of end for each type of speech in a sequences of segments of said speech-types.

19. The method removing excess information from a human speech signal and coding the remaining signal information of claim 15 wherein said at least one characteristic of said human speech signal comprises time of onset, time duration, and time of end for each of 3 types of speech.

20. The method of removing excess information from a human speech signal and coding the remaining signal information of claim 15 wherein said at least one characteristic of said human speech signal comprises the number of glottal periods and one or more spectral formant values within a continuous segment of voiced speech.

21. The method of removing excess information from a human speech signal and coding the remaining signal information of claim 15 wherein said at least one characteristic of said human speech signal comprises the type of unvoiced speech segment, and its amplitude compared to voiced speech.

22. The method of removing excess information from a human speech signal and coding the remaining signal information of claim 15 wherein said at least one characteristic of said human speech signal comprises header-information that describes speech properties of the user.

23. The method of removing excess information from a human speech signal and coding the remaining signal information of claim 15 wherein said at least one characteristic of said human speech signal comprises one or more of an average glottal-period'"'"'s time-duration-value of voiced speech, an excitation function and its coded description, time of onset, time duration, and time of end for each of 3 types of speech in a sequences of segments of said speech-types, number of glottal periods and one or more spectral formant values within a continuous segment of voiced speech, the type of unvoiced speech segment, and its amplitude compared to voiced speech, and header-information that describes speech properties of the user.

24. The method of removing excess information from a human speech signal and coding the remaining signal information of claim 15 wherein said step of using one or more EM wave sensors comprises using one or more coherent wave EM sensors.

25. The method of removing excess information from a human speech signal and coding the remaining signal information of claim 15 wherein said step of using one or more EM wave sensors comprises using a coherent wave EM sensor to measure air pressure induced tissue movement in the human vocal tract for purposes of glottal timing, excitation function description, and voiced speech segment onset times.

26. The method of removing excess information from a human speech signal and coding the remaining information of claim 15 wherein said step of using one or more EM wave sensors comprises using a coherent optical-frequency EM sensor for obtaining vocal tract wall movement by measuring surface motion.

27. The method of removing excess information from a human speech signal and coding the remaining signal information of claim 15 wherein the remaining signal information is coded and transmitted at a constant bandwidth.

28. The method of removing excess information from a human speech signal and coding the remaining signal information of claim 15 wherein the bandwidth and latency are adjusted to meet user applications.

29. The method of removing excess information from a human speech signal and coding the remaining signal information of claim 15 in which constant bit rate transmission coding uses

coding of speech segment onset times and end times,

coding of speech segments according to their type,

coding of the number and duration of glottal cycles of the user in each voiced speech segment as a function of user defined latency and bandwidth limitations.

30. The method of removing excess information from a human speech signal and coding the remaining signal information of claim 15 wherein the coded and transmitted signal is reconstructed into real time speech segments and then into speech phrases which meet the intelligibility objectives of the listener.

31. A communication apparatus, comprising:

at least one EM wave sensor,

at least one acoustic microphone, and

processing means for removing excess information from a human speech signal that produces a remaining signal, and

processing means for coding said remaining signal to provide a coded signal;

wherein said processing means for removing said excess information and said processing means for coding said remaining signal uses said at least one EM wave sensor and said at least one acoustic microphone to determine at least one of the following;

an average glottal period time duration value and variations of the average value from voiced speech

a voiced speech excitation function and its coded description

time of onset, time duration, and time of end for each type of speech in a sequence of segments of said speech-types

number of glottal periods and one or more spectral formant values within a continuous segment of voiced speech

the type of unvoiced speech segment, and its amplitude compared to voiced speech

header-information that describes speech properties of the user.

32. The apparatus of claim 31 which comprises a hand held wireless telephone transmission and receiving communications device, containing:

a EM wave generator, transmitting structure, and receiver for measuring vocal organ movements,

an acoustic microphone,

a processor and algorithms for removing excess speech information and for coding remaining information, and for formatting said coding into a transmission formant meeting the specifications of the communications channel to which said apparatus is attached.

33. The apparatus of claim 31 including a processor and algorithms for decoding information received from another user of methods herein whereby the received information is formatted into intelligible speech.

34. The apparatus of claim 31 in which a wireless transmitting antenna, transmitter, and receiver also serve as a vocal organ measuring EM sensor.

35. A system for removing excess information characterizing a human speech signal, and coding the remaining signal information, comprising:

at least one EM wave sensor,

at least one acoustic microphone,

processing means for removing said excess information from said acoustic microphone signal and from said EM sensor signal that produces a remaining signal, and

processing means for coding said remaining information to provide a coded signal with at least one characteristic of said human speech signal.