# US 2018/0293997 A1 · Identification of noise signal for voice denoising device

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

- Number: US 2018/0293997 A1
- Document ID: US20180293997A1
- Status: Published application
- Kind: A1
- Application number: 15/951,928
- Filed: 2018-04-12
- Published: 2018-10-11
- Earliest claimed: 2015-10-13
- Assignees: Alibaba Group Holding Limited
- Applicants: Alibaba Group Holding Limited
- Inventors: Zhijun Du
- Web page: https://patentgrep.com/patent/US20180293997A1

## CPC

- G10L 25/21: the extracted parameters being power information
- G10L 25/84: for discriminating voice from noise
- G10L 21/0216: characterised by the method used for estimating noise
- G10L 21/0232: Processing in the frequency domain
- G10L 21/0324: Details of processing therefor
- G10L 25/51: for comparison or discrimination
- G10L 25/18: the extracted parameters being spectral information of each sub-band
- G10L 2021/02168: {the estimation exclusively taking place during speech pauses}
- G10L 2025/783: {based on threshold decision}
- G10L: SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING

## Abstract

Methods, systems, and computer-readable storage media for voice denoising. Implementations include actions of performing a mathematical transform on each frame signal in an audio signal segment to generate multiple power spectra. Each power spectrum corresponds to a respective frame signal. Power value variances corresponding to frame signals at various frequencies are determined. A noise signal is identified in each frame signal based on the power value variance. The identified noise signal is removed from each frame signal of the plurality of frame signals.

## Claims

### Claim 1

A computer-implemented method for voice denoising, the method being executed by one or more processors and comprising:
performing, by the one or more processors, a mathematical transform on each frame signal in an audio signal segment comprising a plurality of frame signals to generate a plurality of power spectra, each power spectrum of the plurality of power spectra corresponding to a respective frame signal;
determining, by the one or more processors, a plurality of power value variances, each power value variance of the plurality of power value variances corresponding to the respective frame signal at various frequencies based on the plurality of power spectra;
identifying, by the one or more processors, a noise signal in each frame signal of the plurality of frame signals based on the power value variance corresponding to the frame signal; and
removing, by the one or more processors, the identified noise signal from each frame signal of the plurality of frame signals from the audio signal segment.

### Claim 2 (depends on 1)

The computer-implemented method of claim 1, further comprising determining the audio signal segment based on comparing an amplitude variation to a threshold.

### Claim 3 (depends on 1)

The computer-implemented method of claim 1, wherein identifying the noise signal comprises comparing the variance corresponding to each frame signal in the audio signal segment to a noise threshold.

### Claim 4 (depends on 1)

The computer-implemented method of claim 1, wherein determining the variance of power values comprises:
at least classifying power values of the frame signal at various frequencies into a first power value set corresponding to a first frequency interval according to frequency intervals to which frequencies corresponding to the power spectrum belong; and
determining a first variance of power values comprised in the first power value set.

### Claim 5 (depends on 1)

The computer-implemented method of claim 1, wherein determining the variance of power values comprises:
at least classifying power values of each frame signal at various frequencies into a first power value set corresponding to a first frequency interval and a second power value set corresponding to a second frequency interval according to frequency intervals to which frequencies corresponding to the power values of the frame signal belong, wherein the first frequency interval is lower than the second frequency interval; and
determining a first variance of power values comprised in the first power value set.

### Claim 6 (depends on 1)

The computer-implemented method of claim 1, further comprising ranking the frame signals in the audio signal segment according to magnitudes of the variances.

### Claim 7 (depends on 6)

The computer-implemented method of claim 6, further comprising: in response to ranking the frame signals, determining whether each frame signal in the audio signal segment is a noise signal based on the variance of power values of each ranked frame signal at various frequencies.

### Claim 8

A non-transitory, computer-readable medium storing one or more instructions executable by a computer system to perform operations for performing voice denoising, the operations comprising:
performing a mathematical transform on each frame signal in an audio signal segment comprising a plurality of frame signals to generate a plurality of power spectra, each power spectrum of the plurality of power spectra corresponding to a respective frame signal;
determining a plurality of power value variances, each power value variance of the plurality of power value variances corresponding to the respective frame signal at various frequencies based on the plurality of power spectra;
identifying, by the one or more processors, a noise signal in each frame signal of the plurality of frame signals based on the power value variance corresponding to the frame signal; and
removing, by the one or more processors, the identified noise signal from each frame signal of the plurality of frame signals.

### Claim 9 (depends on 8)

The non-transitory, computer-readable medium of claim 8, the operations further comprising determining the audio signal segment based on comparing an amplitude variation to a threshold.

### Claim 10 (depends on 8)

The non-transitory, computer-readable medium of claim 8, wherein identifying the noise signal comprises comparing the variance corresponding to each frame signal in the audio signal segment to a noise threshold.

### Claim 11 (depends on 9)

The non-transitory, computer-readable medium of claim 9, wherein determining the variance of power values comprises:
at least classifying power values of the frame signal at various frequencies into a first power value set corresponding to a first frequency interval according to frequency intervals to which frequencies corresponding to the power spectrum belong; and
determining a first variance of power values comprised in the first power value set.

### Claim 12 (depends on 8)

The non-transitory, computer-readable medium of claim 8, wherein determining the variance of power values comprises:
at least classifying power values of each frame signal at various frequencies into a first power value set corresponding to a first frequency interval and a second power value set corresponding to a second frequency interval according to frequency intervals to which frequencies corresponding to the power values of the frame signal belong, wherein the first frequency interval is lower than the second frequency interval; and
determining a first variance of power values comprised in the first power value set.

### Claim 13 (depends on 8)

The non-transitory, computer-readable medium of claim 8, the operations further comprising ranking the frame signals in the audio signal segment according to magnitudes of the variances.

### Claim 14 (depends on 13)

The non-transitory, computer-readable medium of claim 13, the operations further comprising in response to ranking the frame signals, determining whether each frame signal in the audio signal segment is a noise signal based on the variance of power values of each ranked frame signal at various frequencies.

### Claim 15

A computer-implemented system for voice denoising, comprising:
one or more computers; and
one or more computer memory devices interoperably coupled with the one or more computers and having tangible, non-transitory, machine-readable media storing instructions that, if executed by the one or more computers, perform operations comprising:
performing a mathematical transform on each frame signal in an audio signal segment comprising a plurality of frame signals to generate a plurality of power spectra, each power spectrum of the plurality of power spectra corresponding to a respective frame signal;
determining a plurality of power value variances, each power value variance of the plurality of power value variances corresponding to the respective frame signal at various frequencies based on the plurality of power spectra;
identifying a noise signal in each frame signal of the plurality of frame signals based on the power value variance corresponding to the frame signal; and
removing the identified noise signal from each frame signal of the plurality of frame signals from the audio signal segment.

### Claim 16 (depends on 15)

The computer-implemented system of claim 15, the operations further comprising determining the audio signal segment based on comparing an amplitude variation to a threshold.

### Claim 17 (depends on 15)

The computer-implemented system of claim 15, wherein identifying the noise signal comprises comparing the variance corresponding to each frame signal in the audio signal segment to a noise threshold.

### Claim 18 (depends on 15)

The computer-implemented system of claim 15, wherein determining the variance of power values comprises:
at least classifying power values of the frame signal at various frequencies into a first power value set corresponding to a first frequency interval according to frequency intervals to which frequencies corresponding to the power spectrum belong; and
determining a first variance of power values comprised in the first power value set.

### Claim 19 (depends on 15)

The computer-implemented system of claim 15, wherein determining the variance of power values comprises:
at least classifying power values of each frame signal at various frequencies into a first power value set corresponding to a first frequency interval and a second power value set corresponding to a second frequency interval according to frequency intervals to which frequencies corresponding to the power values of the frame signal belong, wherein the first frequency interval is lower than the second frequency interval; and
determining a first variance of power values comprised in the first power value set.

### Claim 20 (depends on 15)

The computer-implemented system of claim 15, the operations further comprising ranking the frame signals in the audio signal segment according to magnitudes of the variances.
