# US 12,770,626 B2 · Encoding method and encoding device using complex signal and decoding method and decoding device using complex signal

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

- Number: US 12,770,626 B2
- Document ID: US12770626B2
- Status: Granted
- Kind: B2
- Application number: 18/108,431
- Filed: 2023-06-12
- Published: 2023-09-21
- Granted: 2026-10-20
- Earliest claimed: 2023-06-12
- Assignees: Electronics and Telecommunications Research Institute
- Applicants: Electronics and Telecommunications Research Institute
- Inventors: Seung Kwon Beack; Jongmo Sung; Tae Jin Lee; Woo-taek Lim; Inseon Jang; Byeongho Cho
- Web page: https://patentgrep.com/patent/US12770626B2

## CPC

- G10L 19/008: Multichannel audio signal coding or decoding using interchannel correlation to reduce redundancy, e.g. joint-stereo, intensity-coding or matrixing
- G10L: SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING

## Abstract

An encoding method and an encoding device using a complex signal and a decoding method and a decoding device using a complex signal are provided. The encoding method includes converting a first channel signal and a second channel signal constituting an audio signal corresponding to a stereo signal from a real domain to a complex domain, determining one of a sum operation, a difference operation, and a bypass operation to be performed on the second channel signal converted to the complex domain, determining a complex spatial cue according to the determined operation, converting a residual signal for the second channel signal to a real domain using the complex spatial cue, converting the first channel signal to a real domain, encoding the first channel signal converted to the real domain, and encoding the residual signal for the second channel signal converted to the real domain.

## Claims

### Claim 1

An encoding method, comprising:
converting a first channel signal x1(n) constituting an audio signal corresponding to a stereo signal from a real domain to a complex domain to a complex domain and outputting a first channel signal in a complex form;
converting a second channel signal x2(n) constituting an audio signal corresponding to a stereo signal from a real domain to a complex domain and outputting a second channel signal in a complex form;
performing one of a sum operation, a difference operation, and a bypass operation on the first channel signal in a complex form and the second channel signal in a complex form and outputting a computed first channel signal xc,1(n) and a computed second channel signal xc,2(n);
deriving a complex spatial cue from the computed first channel signal xc,1(n) and the computed second channel signal xc,2(n);
correcting a phase difference and a gain value of the computed first channel signal xc,1(n) with the complex spatial cue and outputting a differential signal {circumflex over (x)}c,2(n);
converting a residual signal rc,2(n) to a real domain, which is a difference between the computed second channel signal xc,2(n) and the differential signal {circumflex over (x)}c,2(n);
converting the computed first channel signal xc,1(n) to a real domain;
encoding the first channel signal converted to the real domain;
encoding the residual signal converted to the real domain;
performing gain estimation by comparing a power of the residual signal and a power of the computed second channel signal, and
determining which operation to be performed among the sum operation, the difference operation, and the bypass operation on the second channel signal in a complex form using a SNRtype, which is an SNR of an estimated gain value,
wherein the SNRtype is determined according to Equation 5, and in Equation 5, sum, diff, bypass, which are the three types of
x
c
,
2
type
(
n
)
,
is determined according to Equation 6,
SNR
type
=
log
10
⁢
pow
⁡
(
x
c
,
2
type
(
n
)
)
pow
⁡
(
r
c
,
2
type
(
n
)
)
[
Equation
⁢
5
]
x
c
,
2
sum
⁢
(
n
)
=
real_to
⁢
_complex
⁢
{
x
1
⁢
(
n
)
}
+
real_to
⁢
_complex
⁢
{
x
2
⁢
(
n
)
}
x
c
,
2
diff
(
n
)
=
real_to
⁢
_complex
⁢
{
x
1
(
n
)
}
-
real_to
⁢
_complex
⁢
{
x
2
(
n
)
}
x
c
,
2
bypass
(
n
)
=
real_to
⁢
_complex
⁢
{
x
2
(
n
)
}
[
Equation
⁢
6
]
wherein the determining which operation to be performed among the sum operation, the difference operation, and the bypass operation comprises:
determining the operation to be performed on the second channel signal in a complex form as a type having a largest SNR among SNRs for each operation defined in Equation 6;
defining
x
c
,
1
diff
(
n
)
when defined as
x
c
,
2
sum
(
n
)
in Equation 6; and
defining
x
c
,
1
sum
(
n
)
when defined as
x
c
,
2
diff
(
n
)
in Equation 6.

### Claim 2

An encoding device comprising a processor,
wherein the processor is configured to:
convert a first channel signal x1(n) constituting an audio signal corresponding to a stereo signal from a real domain to a complex domain to a complex domain and outputting a first channel signal in a complex form;
convert a second channel signal x2(n) constituting an audio signal corresponding to a stereo signal from a real domain to a complex domain and outputting a second channel signal in a complex form;
perform one of a sum operation, a difference operation, and a bypass operation on the first channel signal in a complex form and the second channel signal in a complex form and outputting a computed first channel signal xc,1(n) and a computed second channel signal xc,2(n);
derive a complex spatial cue from the computed first channel signal xc,1(n) and the computed second channel signal xc,2(n);
correct a phase difference and a gain value of the computed first channel signal xc,1(n) with the complex spatial cue and outputting a differential signal xc,2(n);
convert a residual signal rc,2(n) to a real domain, which is a difference between the computed second channel signal xc,2(n) and the differential signal {circumflex over (x)}c,2(n);
convert the computed first channel signal xc,1(n) to a real domain;
encode the first channel signal converted to the real domain; and
encode the residual signal converted to the real domain;
perform gain estimation by comparing a power of the residual signal and a power of the computed second channel signal; and
determine which operation to be performed among the sum operation, the difference operation, and the bypass operation on the second channel signal in a complex form using a SNRtype which is an SNR of an estimated gain value,
wherein the SNRtype is determined according to Equation 5, and in Equation 5, sum, diff, bypass, which are the three types of
x
c
,
2
type
(
n
)
,
is determined according to Equation 6,
SNR
type
=
log
10
⁢
pow
⁡
(
x
c
,
2
type
(
n
)
)
pow
⁡
(
r
c
,
2
type
(
n
)
)
[
Equation
⁢
5
]
x
c
,
2
sum
⁢
(
n
)
=
real_to
⁢
_complex
⁢
{
x
1
⁢
(
n
)
}
+
real_to
⁢
_complex
⁢
{
x
2
⁢
(
n
)
}
x
c
,
2
diff
(
n
)
=
real_to
⁢
_complex
⁢
{
x
1
(
n
)
}
-
real_to
⁢
_complex
⁢
{
x
2
(
n
)
}
x
c
,
2
bypass
(
n
)
=
real_to
⁢
_complex
⁢
{
x
2
(
n
)
}
[
Equation
⁢
6
]
wherein the determining which operation to be performed among the sum operation, the difference operation, and the bypass operation comprises:
determining the operation to be performed on the second channel signal in a complex form as a type having a largest SNR among SNRs for each operation defined in Equation 6;
defining
x
c
,
1
d
⁢
i
⁢
f
⁢
f
(
n
)
when defined as
x
c
,
2
sum
(
n
)
in Equation 6; and
defining
x
c
,
1
sum
(
n
)
when defined as
x
c
,
2
diff
(
n
)
in Equation 6.