System and apparatus for speech communication and speech recognition

• US 7,346,175 B2
• Filed: 07/02/2002
• Issued: 03/18/2008
• Est. Priority Date: 09/12/2001
• Status: Active Grant

• Current Assignee: Bitwave Pte Ltd.
• Original Assignee: Bitwave Pte Ltd.
• Inventors: Lau, Yean Ming, Loh, Kok Heng, Hui, Siew Kok
• Primary Examiner(s): Chin; Vivian
• Assistant Examiner(s): LAO, LUNSEE
• Application Number:
  US10/487,229
• Publication Number:
  US 20040193411A1
• Time in Patent Office: 2,086 Days
• Field of Search: 381/74, 381 711- 718, 381 7111- 7112, 381 941- 943, 455/41.2, 455/412, 704/233, 704227-228, 704/231, 379/406.8
• US Class Current: 381/74
• CPC Class Codes: G10L 15/20 Speech recognition techniques specially adapted for robustness in adverse environments, e.g. in noise, of stress induced speech G10L21/02 takes precedence
• CPC Class Codes: G10L 2021/02161 Number of inputs available containing the signal or the noise to be suppressed
• CPC Class Codes: G10L 2021/02168 the estimation exclusively taking place during speech pauses
• CPC Class Codes: G10L 21/0208 Noise filtering
• CPC Class Codes: H04M 1/271 controlled by voice recognition
• CPC Class Codes: H04M 1/6033 for providing handsfree use or a loudspeaker mode in telephone sets arrangements for preventing acoustic echo H04M9/08

Abstract

A headset system is proposed including a headset unit to be worn by a user and having two or more microphones, and a base unit in wireless communication with the headset. Signals received from the microphones are processed using a first adaptive filter to enhance a target signal, and then divided and supplied to a second adaptive filter arranged to reduce interference signals and a third filter arranged to reduce noise. The outputs of the second and third filters are combined, and are be subject to further a processing in the frequency domain. The results are transmitted to a speech recognition engine.

Claims

1. A headset system including a base unit and a headset unit for wearing by a user and having a plurality of microphones, the headset unit and base unit being arranged for mutual wireless communication, and at least one of the base unit and the headset unit comprising at least one digital signal processing unit arranged to perform signal processing in the time domain on audio signals received by the plurality of microphones, the at least one digital signal processing unit including a first adaptive filter configured to enhance a wanted signal in the audio signals, a second adaptive filter and a third adaptive filter each configured to receive an output of the first adaptive filter, the second adaptive filter being configured to reduce an unwanted interference signal in the audio signals, and the third adaptive filter being arranged to estimate noise signals.

2. A headset system according to claim 1, in which the base unit comprises a cradle configured to hold the headset unit.

3. A headset system according to claim 1, in which the headset unit is associated with a loudspeaker operable by the headset unit for generating audio signals to the user.

4. A headset system according to claim 1, the at least one digital processing unit being adapted to combine output signals of the second and third adaptive filters, convert the combined signals to the frequency domain and perform further processing in the frequency domain.

5. A headset system according to claim 4, configured to linearly combine an output S_i(t) of the second adaptive filter and an output S_n(t) of the third adaptive filter using weighting factors to derive two interference signals, to subtract a first of the interference signals Ic from the output of the first adaptive filter, and to convert a second of the interference signals Is into the frequency domain.

6. A headset system according to claim 1, configured so that the second and third adaptive filters are not adapted if it is determined that a target signal is present.

7. A headset system according to claim 1, configured so that the second adaptive filter is not adapted if it is determined that an interference signal is not present.

8. A headset system according to claim 6, further, configured to determine signal energy at intervals, and to derive at least one noise threshold from a plurality of values of the signal energy, said determination including a determination of whether a further signal energy is above the at least one noise threshold.

9. A headset system according to claim 8, further configured so that the derivation of said at least one noise threshold includes using the plurality of values of the signal energy to derive a histogram representing the statistical frequencies of signal energy values in each of a number of bands, and deriving the at least one noise threshold from a signal energy value Emax associated with the band having the highest histogram value.

10. A headset system according to claim 1, in which the at least one digital signal processing unit comprises a fourth adaptive filter configured to determine a direction of arrival of the wanted signal.

11. A headset system according to claim 10, configured for the weights of the fourth adaptive filter to be updated, including being configured to repeatedly perform an update process which attenuates each existing weight value by a forgetting factor α

.

12. A headset system according to claim 10, the at least one digital signal processing unit being configured to determine a ratio P_k indicating the ratio of the highest central weight value A of the fourth adaptive filter to the sum of A and the highest peripheral weight value B, the at least one digital signal processing unit being configured only to adapt the first filter if the ratio P_k is above a given value T_pkl.

13. A headset system according to claim 12, in which, following an adaptation of the first filter, the at least one digital signal processing means unit is configured to calculate a new value P_k2 of the ratio, to determine whether the new value of P_k2 is below previous maximum value of P_k2 and below a threshold T_pk, and if so to restore at least one of the first, second and third adaptive filters to its previous state.

14. A headset system according to claim 12, configured so that the determination that an interference signal is not present includes a determination that the value of said ratio P_k is below a threshold T_pk2.

15. A headset system according to claim 1, configured to adapt the weights of the second or third filters by weighting update factors which are inverse functions of respective error outputs e_c1 of the second filter and e_c2 of the third filter.

16. A headset system according to claim 4, configured to transform the combined signals into two frequency domain signals which are a signal S_f and an interference signal I_f, to transform S_f and I_f into respective modified spectra P_s and P_i, and to warp the modified spectra into respective Bark spectra Bs and Bi.

17. A headset system according to claim 16, further configured, prior to said warping, to apply frequency scanning to the modified spectra P_s and P_i, and to attenuate in P_ipeaks which are found to be common to both.

18. A headset system according to claim 16, further configured to derive a ratio of the sum of the values of B_s over the Bark critical bands up to the voice band upper cutoff, and the sum of the values of B_s over the Bark critical bands at and above the unvoiced bank lower cutoff.

19. A headset system according to claim 18, further configured to derive the ratio above a given threshold, and to amplify the values of B_s above the unvoiced band lower threshold.

20. A headset system according to claim 1, further comprising a speech recognition engine configured to receive the output of the at least one digital signal processing unit.

21. A headset system according to claim 17, further configured for the speech recognition engine to receive from the at least one digital signal processing unit information indicating at least one of:

a) a direction of a target signal T_d,

b) a signal Energy E_r1,

c) a noise threshold used by the at least one digital signal processing unit,

d) an estimated SINR (target signal to interference ratio) and SNR (target signal to noise ratio),

e) a signal A′

indicating the presence of target speech,

f) a spectrum of processed speech signal S_out,

g) potential speech start and end points, and

h) an interference signal spectrum, I_f.

22. A headset system according to claim 1, in which the headset unit comprises two arms configured to be positioned proximate the mouth of the user and to be positioned on either side of the user'"'"'s head.

23. A headset system according to claim 22, in which the headset unit is further configured to be supportable by the user'"'"'s shoulders with the two arms embracing the user'"'"'s neck.

24. A headset system according to claim 22, in which at least one microphone is provided on a free end of each of the two arms.

25. A method of processing signals received from an array of sensors, including sampling and digitizing the received signals into digital signals, the method comprising:

filtering the digital signals using a first adaptive filter arranged to enhance a target signal in the digital signals;

transmitting an output of the first adaptive filter to each of a second adaptive filter and a third adaptive filter, the second adaptive filter being arranged to suppress unwanted interference signals, and the third adaptive filter being arranged to estimate noise signals; and

combining the outputs of the second and third adaptive filters.

26. A headset system including a base unit and a headset unit for wearing by a user and having a plurality of microphones, the headset unit and the base unit being arranged for mutual wireless communication, and at least one of the base unit and the headset unit comprising at least one digital signal processing unit arranged to perform signal processing in a time domain on audio signals generated by the plurality of microphones, the at least one digital signal processing unit including a first adaptive filter configured to enhance a wanted signal in the audio signals, a second adaptive filter and a third adaptive filter each configured to receive an output of the first adaptive filter, the second adaptive filter being configured to reduce an unwanted interference signal in the audio signals, the third adaptive filter being arranged to estimate noise signals, the headset system being configured:

so that the second and third adaptive filters are not adapted if it is determined that a target signal is present;

to determine signal energy at intervals, and to derive at least one noise threshold from a plurality of values of the signal energy, including to determine whether a further signal energy is above the at least one noise threshold; and

so that the derivation of said at least one noise threshold includes using the plurality of values of the signal energy to derive a histogram representing the statistical frequencies of signal energy values in each of a number of bands, and deriving the at least one noise threshold from a signal energy value E_max associated with a band having a highest histogram value.