Adaptive reduction of noise signals and background signals in a speech-processing system
First Claim
Patent Images
1. A method for reducing noise signals and background signals in a speech-processing system, comprising:
- adaptively filtering an audio input signal, using a filter, to generate a prediction output signal using a plurality of coefficients to generate a plurality of prediction errors and generating an error from the plurality of prediction errors where the prediction output signal is the sum of the plurality of prediction errors;
where the absolute values of the coefficients are continuously reduced by a plurality of reduction parameters;
where the prediction output signal as a prediction of the audio input signal with reduced noise is used as an input signal for a second filter to generate a second prediction; and
where the second filter comprises a prediction filter having a second filter with a set of second coefficients, wherein a learning rate to adapt the coefficients is selected that is several powers of ten less than a learning rate of the first filter.
9 Assignments
0 Petitions
Accused Products
Abstract
An audio input signal is filtered using an adaptive filter to generate a prediction output signal with reduced noise, wherein the filter is implemented using a plurality of coefficients to generate a plurality of prediction errors and to generate an error from the plurality of prediction errors, wherein the absolute values of the coefficients are continuously reduced by a plurality of reduction parameters.
39 Citations
26 Claims
-
1. A method for reducing noise signals and background signals in a speech-processing system, comprising:
-
adaptively filtering an audio input signal, using a filter, to generate a prediction output signal using a plurality of coefficients to generate a plurality of prediction errors and generating an error from the plurality of prediction errors where the prediction output signal is the sum of the plurality of prediction errors; where the absolute values of the coefficients are continuously reduced by a plurality of reduction parameters; where the prediction output signal as a prediction of the audio input signal with reduced noise is used as an input signal for a second filter to generate a second prediction; and where the second filter comprises a prediction filter having a second filter with a set of second coefficients, wherein a learning rate to adapt the coefficients is selected that is several powers of ten less than a learning rate of the first filter. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13)
where k, with 0<
k<
<
1, is a reduction parameter;μ
, with μ
<
<
1, is a learning rate;e is an error resulting from the difference of all the individual prediction errors (sv1-sv4) from the audio input signal s(t); sv(t) is the prediction output signal resulting from the sum of all the individual prediction errors, where N is the number of coefficients c, (t); and c;
(t) is an individual coefficient having an index i at time t.
-
-
4. The method of claim 3, where the coefficients are computed according to the equation
ci(t+1)=ci(t)+(μ- ·
e·
s(t−
i))−
kci(t)
where
e=s(t)−
sv(t) and
sv(t)=Σ
i=1 . . . Nci(t−
1)·
s(t−
i).
- ·
-
5. The method of claim 1, comprising subtracting the second prediction from the prediction output signal.
-
6. The method of claim 5, where a learning rule is asymmetrically designed to determine the subsequent coefficients such that the absolute values of the subsequent coefficients fall more significantly in absolute value than they rise and can rapidly fall to zero, but rise only with a small gradient.
-
7. The method of claim 1, where the coefficients are limited to prevent drifting of the coefficients-when the audio input signal is normalized.
-
8. The method of claim 1, where an output signal of the first and/or second filter relative to its input signal is used as a measure for the presence of speech in the input signal.
-
9. The method of claim 1, where the step of adaptively filtering comprises least mean squares processing.
-
10. The method of claim 9, where the step of adaptively filtering comprises FIR filtering.
-
11. The method of claim 1, comprising multiplying a sigmoid function by the prediction output signal to prevent an overmodulation of the signal in case of a bad prediction.
-
12. The method of claim 1, comprising mixing the audio input signal with the prediction output signal.
-
13. The method of claim 1, further comprising programming an application-specific integrated circuit.
-
14. A method, for reducing noise signals and background signals in a speech-processing system, comprising:
-
adaptively filtering a sign of an audio input signal to determine individual prediction errors by using a filter, to generate a prediction output signal using a plurality of coefficients to generate a plurality of prediction errors and generating an error from the plurality of prediction errors where the prediction output signal is the sum of the plurality of prediction errors; where the absolute values of the coefficients are continuously reduced by a plurality of reduction parameters. - View Dependent Claims (15, 16)
-
-
17. A method for reducing noise signals and background signals in a speech-processing system, comprising:
-
adaptively filtering an audio input signal, using a filter, to generate a prediction output signal using a plurality of coefficients to generate a plurality of prediction errors and generating an error from the plurality of prediction errors where the prediction output signal is the sum of the plurality of prediction errors; where the absolute values of the coefficients are continuously reduced by a plurality of reduction parameters; and where a maximum of a speech signal component of the audio input signal is detected, and an output signal is renormalized to the maximum. - View Dependent Claims (18)
-
-
19. A device for the reduction of noise signals and background signals in a speech-processing system, comprising:
-
an adaptive filter that filters an audio input signal and provides a prediction output signal with reduced noise; memory that stores a plurality of coefficients for the adaptive filter; a multiplier to weight the optionally time-delayed audio input signal, or to weight the prediction output signal by a weighting factor smaller than one; and an adder to add the weighted signal to the prediction output signal or to the prediction to generate a noise-reduced audio output signal wherein the adaptive filter generates a plurality of prediction errors and an error from the plurality of prediction errors, where a coefficient supply circuit continuously reduces the absolute values of the coefficients using at least one reduction parameter. - View Dependent Claims (20, 21, 22, 23, 24, 25, 26)
-
Specification