# US 2021/0174819 A1 · Apparatus and method for generating audio signal with noise attenuated on basis of phase change rate according to change in frequency of audio signal

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2021/0174819 A1
- Document ID: US20210174819A1
- Status: Published application
- Kind: A1
- Application number: 16/769,675
- Filed: 2018-12-18
- Published: 2021-06-10
- Earliest claimed: 2018-02-12
- Assignees: Samsung Electronics Co., Ltd.
- Applicants: Samsung Electronics Co., Ltd.
- Inventors: Soonho BAEK; Hangil MOON; Beakkwon SON; Jaemo YANG; Kiho CHO; Juntai KIM
- Web page: https://patentgrep.com/patent/US20210174819A1

## CPC

- G10L 21/0208: Noise filtering
- G10L 21/0232: Processing in the frequency domain
- G06N 3/04: Architecture, e.g. interconnection topology
- G10L 21/0216: characterised by the method used for estimating noise
- G10L 21/0316: by changing the amplitude
- G10L 25/18: the extracted parameters being spectral information of each sub-band
- G10L 25/30: using neural networks
- G10L 25/51: for comparison or discrimination
- H04R 1/406: {microphones}
- H04R 3/005: {for combining the signals of two or more microphones (specially adapted for hearing aids H04R25/407)}
- G10L 2021/02166: {Microphone arrays; Beamforming}
- G10L: SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- H04R: LOUDSPEAKERS, MICROPHONES, GRAMOPHONE PICK-UPS OR LIKE ACOUSTIC ELECTROMECHANICAL TRANSDUCERS; ELECTRIC HEARING AIDS; PUBLIC ADDRESS SYSTEMS (producing sounds with frequency not determined by supply frequency G10K)

## Abstract

An electronic device is disclosed. In addition, various embodiments recognized through the specification are possible. The electronic device includes a plurality of input devices for receiving a plurality of input signals, each of which includes a speech signal and a noise signal and a processor electrically connected with the input devices. The processor is configured to determine a signal to noise ratio (SNR) value for the plurality of input signals for each frequency band, determine a first parameter indicating a phase shift to frequency in the plurality of input signals in a first frequency band where the SNR value is greater than or equal to a specified threshold, determine a second parameter indicating a phase shift to frequency in the plurality of input signals in a second frequency band where the SNR value is less than the threshold, and perform beamforming for the plurality of input signals based on the first parameter and the second parameter.

## Claims

### Claim 1

An electronic device, comprising:
a plurality of input devices configured to receive a plurality of input signals, each of which includes a speech signal and a noise signal; and
a processor electrically connected with the input devices,
wherein the processor is configured to:
determine a signal to noise ratio (SNR) value for the plurality of input signals for each frequency band;
determine a first parameter indicating a phase shift to frequency in the plurality of input signals in a first frequency band where the SNR value is greater than or equal to a specified threshold;
determine a second parameter indicating a phase shift to frequency in the plurality of input signals in a second frequency band where the SNR value is less than the threshold, based on the first parameter; and
perform beamforming for the plurality of input signals based on the first parameter and the second parameter.

### Claim 2 (depends on 1)

The electronic device of claim 1, wherein the processor is configured to:
determine weights for the plurality of input signals based on the first parameter and the second parameter and generate an output signal where at least some of noise signals included in the plurality of input signals are attenuated by applying the weights to the plurality of input signals, as at least a portion of the performing of the beamforming.

### Claim 3 (depends on 1)

The electronic device of claim 1, wherein the processor is configured to:
determine the first parameter based on a covariance matrix for a speech signal or a covariance matrix for a noise signal in the first frequency band of the plurality of input signals.

### Claim 4 (depends on 3)

The electronic device of claim 3, wherein the processor is configured to:
determine a probability that a speech signal will be present in each of the plurality of input signals in the first frequency band;
determine a spatial matrix for the plurality of input signals in the first frequency band;
determine a covariance matrix for the speech signal or a covariance matrix for the noise signal based on the spatial matrix and the probability; and
determine the first parameter based on the covariance matrix for the speech signal or the covariance matrix for the noise signal.

### Claim 5 (depends on 4)

The electronic device of claim 4, wherein the processor is configured to:
determine the first parameter based on the covariance matrix for the speech signal, a difference between the covariance matrix for the speech signal and the covariance matrix for the noise signal, or an inverse matrix of the covariance matrix for the noise signal and the covariance matrix for the speech signal.

### Claim 6 (depends on 1)

The electronic device of claim 1, wherein the processor is configured to:
determine the second parameter from the first parameter by using a deep neural networks (DNN) model.

### Claim 7 (depends on 4)

The electronic device of claim 4, wherein the processor is configured to:
determine the probability for a single channel in the first frequency band or determine the probability for a plurality of channels in the first frequency band.

### Claim 8 (depends on 1)

The electronic device of claim 1, wherein the processor is configured to:
perform fast Fourier transform of the plurality of input signals.

### Claim 9

An electronic device, comprising:
a first microphone;
a second microphone; and
a processor,
wherein the processor is configured to:
obtain a first audio signal using the first microphone and obtain a second audio signal using the second microphone;
identify a specified frequency band where an SNR belongs to a specified range in conjunction with the first audio signal and the second audio signal;
determine a first change rate at which the first audio signal and the second audio signal vary in phase with frequency in the specified frequency band;
estimate a second change rate, at which the first audio signal and the second audio signal vary in phase with frequency in a frequency band different from the specified frequency band, based at least on the first change rate;
determine a first weight corresponding to the first audio signal and a second weight corresponding to the second audio signal based at least on the first change rate and the second change rate; and
generate a third audio signal, a noise signal of which is attenuated, using the first audio signal and the second audio signal, based at least on the first weight and the second weight.

### Claim 10 (depends on 9)

The electronic device of claim 9, wherein the processor is configured to:
determine the first change rate based at least on a covariance matrix for the first audio signal and the second audio signal in the specified frequency band.

### Claim 11 (depends on 10)

The electronic device of claim 10, wherein the processor is configured to:
determine a presence probability value for a speech signal included in at least one of the first audio signal and the second audio signal in the specified frequency band; and
determine the covariance matrix based on the presence probability value.

### Claim 12 (depends on 9)

The electronic device of claim 9, wherein the processor is configured to:
determine the first change rate based on a difference between a covariance matrix for a speech signal and a covariance matrix for a noise signal in the first audio signal and the second audio signal.

### Claim 13 (depends on 9)

The electronic device of claim 9, wherein the processor is configured to:
estimate the second change rate from the first change rate by using a DNN model.

### Claim 14 (depends on 9)

The electronic device of claim 9, wherein the processor is configured to:
estimate the second change rate based at least on linearity of the first change rate.

### Claim 15 (depends on 11)

The electronic device of claim 11, wherein the processor is configured to:
perform FFT of the first audio signal and the second audio signal.
