Voice signal enhancement

• US 9,437,213 B2
• Filed: 08/20/2012
• Issued: 09/06/2016
• Est. Priority Date: 03/05/2012
• Status: Expired due to Fees

• Current Assignee: Malaspina Labs (Barbados), Inc.
• Original Assignee: Malaspina Labs (Barbados), Inc.
• Inventors: Zakarauskas, Pierre, Escott, Alexander, Chu, Clarence S. H., Stevenson, Shawn E.
• Primary Examiner(s): Jackson, Jakieda
• Application Number:
  US13/589,954
• Publication Number:
  US 20130231923A1
• Time in Patent Office: 1,478 Days
• Field of Search: 704/205, 704/226, 704/225
• US Class Current: 1/1
• CPC Class Codes: G10L 2021/02082 the noise being echo, reverberation of the speech
• CPC Class Codes: G10L 21/0208 Noise filtering
• CPC Class Codes: G10L 21/0308 characterised by the type of parameter measurement, e.g. correlation techniques, zero crossing techniques or predictive techniques
• CPC Class Codes: G10L 21/0324 Details of processing therefor
• CPC Class Codes: G10L 21/0364 for improving intelligibility

Abstract

Implementations include systems, methods and/or devices operable to enhance the intelligibility of a target speech signal by targeted voice model based processing of a noisy audible signal. In some implementations, an amplitude-independent voice proximity function voice model is used to attenuate signal components of a noisy audible signal that are unlikely to be associated with the target speech signal and/or accentuate the target speech signal. In some implementations, the target speech signal is identified as a near-field signal, which is detected by identifying a prominent train of glottal pulses in the noisy audible signal. Subsequently, in some implementations systems, methods and/or devices perform a form of computational auditory scene analysis by converting the noisy audible signal into a set of narrowband time-frequency units, and selectively accentuating the time-frequency units associated with the target speech signal and deemphasizing others using information derived from the identification of the glottal pulse train.

Claims

1. A method of discriminating relative to a voice signal, the method comprising:

receiving, via one or more audible sensors, an audible signal including a target voice signal;

converting the audible signal into a corresponding plurality of wideband time-frequency units, wherein the time dimension of each time-frequency unit includes at least one of a plurality of sequential intervals, and wherein the frequency dimension of each time-frequency unit includes at least one of a plurality of wide sub-bands;

calculating one or more characterizing metrics from the plurality of wideband time-frequency units;

calculating a gain function from one or more characterizing metrics calculated from the plurality of wideband time-frequency units;

converting the audible signal into a corresponding plurality of narrowband time-frequency units;

applying the gain function, calculated from the plurality of wideband time-frequency units, to the plurality of narrowband time-frequency units to produce a corresponding plurality of narrowband gain-corrected time-frequency units;

converting the plurality of narrowband gain-corrected time-frequency units into a corrected audible signal, wherein the corrected audible signal includes an improved target voice signal relative to the received audible signal; and

outputting the corrected audible signal through an output device.

2. The method of claim 1, further comprising receiving the audible signal from a single audio sensor device.

3. The method of claim 1, further comprising receiving the audible signal from a plurality of audio sensors.

4. The method of claim 1, wherein the plurality of wide sub-bands is contiguously distributed throughout the frequency spectrum associated with human speech.

5. The method of claim 1, wherein converting the audible signal into the corresponding plurality of wideband time-frequency units includes applying a Fast Fourier Transform to the audible signal.

6. The method of claim 1, wherein the one or more characterizing metrics comprises:

a strength metric associated the number of glottal pulses identified in the plurality of wideband time-frequency units;

a relative period value indicative of how far an identified period in a respective wide sub-band is from an identified dominant period; and

an autocorrelation coefficient associated with an identified glottal pulse in a respective sub-band.

7. The method of claim 6, wherein one or more of the strength metric, the relative period value and the autocorrelation coefficient are determined from one or more outputs of a voice activity detector.

8. The method of claim 1, further comprising calculating a respective signal-to-noise ratio for each narrow sub-band, and wherein the respective signal-to-noise ratios are included in the calculation of the gain function.

9. The method of claim 1, wherein converting the plurality of narrowband gain-corrected time-frequency units into the corrected audible signal comprises re-synthesizing the audible signal from the plurality of narrowband gain-corrected time-frequency units using an inverse Fast Fourier Transform.

10. The method of claim 1, wherein calculating the gain function includes utilizing a sigmoid function to covert one or more of the characterizing metrics into a respective gain.

11. A method of discriminating against far field audible components, the method comprising:

receiving, via one or more audible sensors, an audible signal including a target voice signal;

converting the audible signal into a corresponding plurality of time-frequency units, wherein the time dimension of each time-frequency unit includes at least one of a plurality of sequential intervals, and wherein the frequency dimension of each time-frequency unit includes at least one of a plurality of sub-bands;

calculating one or more characterizing metrics from the plurality of time-frequency units associated with near field audible components;

calculating a discriminating function from one or more characterizing metrics calculated from the plurality of wideband time-frequency units;

applying the discriminating function, calculated from the plurality of wideband time-frequency units, to the plurality of time-frequency units to produce a corresponding plurality of corrected time-frequency units;

converting the plurality of corrected time-frequency units into a corrected audible signal, wherein the corrected audible signal includes an improved target voice signal relative to the received audible signal; and

outputting the corrected audible signal through an output device.

12. A voice signal enhancement device to discriminate relative to a voice signal, the device comprising:

one or more audio sensors configured to receive and audible signal including a target voice signal;

a first conversion module configured to convert the audible signal into a corresponding plurality of wideband time-frequency units, wherein the time dimension of each time-frequency unit includes at least one of a plurality of sequential intervals, and wherein the frequency dimension of each time-frequency unit includes at least one of a plurality of wide sub-bands;

a second conversion module configured to convert the audible signal into a corresponding plurality of narrowband time-frequency units;

a metric calculator configured to calculate one or more characterizing metrics from the plurality of wideband time-frequency units;

a gain calculator configured to calculate a gain function from one or more characterizing metrics calculated from the plurality of wideband time-frequency units;

a filtering module configured to apply the gain function, calculated from the plurality of wideband time-frequency units, to the plurality of narrowband time-frequency units to produce a corresponding plurality of narrowband gain-corrected time-frequency units;

a third conversion module configured to convert the plurality of narrowband gain-corrected time-frequency units into a corrected audible signal, wherein the corrected audible signal includes an improved target voice signal relative to the received audible signal; and

an output device configured to output the corrected audible signal.

13. The device of claim 12, further comprising an audio sensor to receive the audible signal.

14. The device of claim 12, wherein at least one of the first conversion module and the second conversion module utilizes a Fast Fourier Transform.

15. The device of claim 12, wherein the third conversion module utilizes an Inverse Fast Fourier Transform.

16. The device of claim 12, wherein the metric calculator is operable to determine at least one of:

a strength metric associated the number of glottal pulses identified in the plurality of wideband time-frequency units;

a relative period value indicative of how far an identified period in a respective wide sub-band is from an identified dominant period; and

an autocorrelation coefficient associated with an identified glottal pulse in a respective sub-band.

17. The device of claim 16, further comprising a voice activity detector, and wherein one or more of the strength metric, the relative period value and the autocorrelation coefficient are determined from one or more outputs of the voice activity detector.

18. The device of claim 12, further comprising a narrowband signal-to-noise estimator to determine a respective signal-to-noise ratio for each narrow sub-band, and wherein the respective signal-to-noise ratios are included in the calculation of the gain function.

19. A voice signal enhancement device to discriminate relative to a voice signal, the device comprising:

means for receiving an audible signal including a target voice signal;

means for converting the audible signal into a corresponding plurality of wideband time-frequency units, wherein the time dimension of each time-frequency unit includes at least one of a plurality of sequential intervals, and wherein the frequency dimension of each time-frequency unit includes at least one of a plurality of wide sub-bands;

means for converting the audible signal into a corresponding plurality of narrowband time-frequency units;

means for calculating one or more characterizing metrics from the plurality of wideband time-frequency units;

means for calculating gain function from one or more characterizing metrics calculated from the plurality of wideband time-frequency units;

means for applying the gain function, calculated from the plurality of wideband time-frequency units, to the plurality of narrowband time-frequency units to produce a corresponding plurality of narrowband gain-corrected time-frequency units;

means for converting the plurality of narrowband gain-corrected time-frequency units into a corrected audible signal, wherein the corrected audible signal includes an improved target voice signal relative to the received audible signal; and

means for outputting the corrected audible signal.

20. A voice signal enhancement device to discriminate relative to a voice signal, the device comprising:

one or more audio sensors configured to receive and audible signal including a target voice signal;

a processor;

a memory including instructions, that when executed by the processor cause the device to;

convert an audible signal into a corresponding plurality of wideband time-frequency units, wherein the time dimension of each time-frequency unit includes at least one of a plurality of sequential intervals, and wherein the frequency dimension of each time-frequency unit includes at least one of a plurality of wide sub-bands;

convert the audible signal into a corresponding plurality of narrowband time-frequency units;

calculate one or more characterizing metrics from the plurality of wideband time-frequency units;

calculate gain function from one or more characterizing metrics calculated from the plurality of wideband time-frequency units;

apply the gain function, calculated from the plurality of wideband time-frequency units, to the plurality of narrowband time-frequency units to produce a corresponding plurality of narrowband gain-corrected time-frequency units;

convert the plurality of narrowband gain-corrected time-frequency units into a corrected audible signal, wherein the corrected audible signal includes an improved target voice signal relative to the received audible signal; and

output the corrected audible signal through an output device.