# US 10,482,893 B2 · Sound processing method and sound processing apparatus

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

- Number: US 10,482,893 B2
- Document ID: US10482893B2
- Status: Granted
- Kind: B2
- Application number: 15/800,488
- Filed: 2017-11-01
- Published: 2018-05-03
- Granted: 2019-11-19
- Earliest claimed: 2016-11-02
- Assignees: Yamaha Corporation
- Applicants: Yamaha Corporation
- Inventors: Ryunosuke Daido; Hiraku Kayama
- Web page: https://patentgrep.com/patent/US10482893B2

## CPC

- G10L 21/0364: for improving intelligibility
- G10L 21/007: characterised by the process used
- G10L: SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING

## Abstract

A sound processing method includes a step of applying a nonlinear filter to a temporal sequence of spectral envelope of an acoustic signal, wherein the nonlinear filter smooths a fine temporal perturbation of the spectral envelope without smoothing out a large temporal change. A sound processing apparatus includes a smoothing processor configured to apply a nonlinear filter to a temporal sequence of spectral envelope of an acoustic signal, wherein the nonlinear filter smooths a fine temporal perturbation of the spectral envelope without smoothing out a large temporal change.

## Claims

### Claim 1

A sound processing method comprising:
supplying an acoustic signal;
improving a sound quality of the supplied acoustic signal by:
applying a nonlinear filter to a temporal sequence of original spectral envelope of the supplied acoustic signal to smooth fine temporal perturbation of the original spectral envelope without smoothing out a larger temporal change of the original spectral envelope; and
adjusting the supplied acoustic signal having the original spectral envelope using a temporal sequence of spectral envelope smoothed by the nonlinear filter to generate an acoustic signal having the spectral envelope in which the fine temporal perturbation has been smoothed; and
outputting the acoustic signal having the spectral envelope in which the fine temporal perturbation has been smoothed.

### Claim 2 (depends on 1)

The sound processing method according to claim 1, wherein the nonlinear filter is an epsilon separation type nonlinear filter that generates an output spectral envelope corresponding to a first spectral envelope through a product-sum arithmetic operation of calculating a nonlinear function corresponding to each of two or more second spectral envelopes on periphery of the first spectral envelope among a plurality of spectral envelopes calculated at different time points on the time axis, multiplying each of the nonlinear functions by a coefficient and accumulating the products.

### Claim 3 (depends on 2)

The sound processing method according to claim 2, wherein for each second spectral envelope, among the two or more second envelopes:
in a case where the second spectral envelope is more similar to the first envelope than a predetermined threshold, then a difference vector between the first and second spectral envelopes is used as the nonlinear function, and
in a case where the second spectral envelope is less similar to the first spectral envelope than the threshold, a zero vector or a suppressed vector of the difference is used as the nonlinear function.

### Claim 4 (depends on 3)

The sound processing method according to claim 3, wherein the threshold is set to a small numerical value for a period in which the level of the acoustic signal is small.

### Claim 5 (depends on 1)

The sound processing method according to claim 1, wherein the nonlinear filter performs a product-sum operation on a spectral envelope at a time point and one or more spectral envelopes near the time point and more similar to the spectral envelope at the time point than a threshold to obtain a smoothed spectral envelope at the time point.

### Claim 6

A sound processing apparatus comprising:
a sound supplying device that supplies an acoustic signal;
a smoothing processor configured to improve sound quality of the supplied acoustic signal by:
applying a nonlinear filter to a temporal sequence of original spectral envelope of the supplied acoustic signal to smooth fine temporal perturbation of the original spectral envelope without smoothing out a larger temporal change of the original spectral envelope; and
adjusting the supplied acoustic signal having the original spectral envelope using a temporal sequence of spectral envelope smoothed by the nonlinear filter to generate an acoustic signal having the spectral envelope in which the fine temporal perturbation has been smoothed; and
a sound emitting device that outputs the acoustic signal having the spectral envelope in which the fine temporal perturbation has been smoothed.