# US 12,190,899 B2 · Apparatus and method for acquiring a plurality of audio signals associated with different sound sources

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

- Number: US 12,190,899 B2
- Document ID: US12190899B2
- Status: Granted
- Kind: B2
- Application number: 17/657,600
- Filed: 2022-03-31
- Published: 2022-07-21
- Granted: 2025-01-07
- Earliest claimed: 2019-10-04
- Assignees: Fraunhofer-Gesellschaft zur Förderung der angewandten Forschung e. V.; Technische Universitat Ilmenau
- Applicants: Fraunhofer-Gesellschaft zur Förderung der angewandten Forschung e.V.; Technische Universität Ilmenau
- Inventors: Gerald Schuller
- Web page: https://patentgrep.com/patent/US12190899B2

## CPC

- G10L 21/0272: Voice signal separating
- G06N 5/01: Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound
- G06N 7/01: Probabilistic graphical models, e.g. probabilistic networks
- 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

## Abstract

Methods, apparatus and techniques for acquiring output signals associated with different sources (such as audio sources) are presented. A first input signal is combined with a delayed and scaled version of a second input signal, to acquire a first output signal. A second input signal is combined with a delayed and scaled version of the first input signal, to acquire a second output signal. Using a random direction optimization, scaling values (forming a candidates' vector) are determined by iteratively modifying the candidates' vector. A Kullback-Leibler divergence is measured. The first output signal and the second output signal are selected to be those measurements associated with the candidate parameters associated with Kullback-Leibler divergence which indicates lowest similarity.

## Claims

### Claim 1

An apparatus for acquiring a plurality of output signals, associated with different sound sources, on the basis of a plurality of input signals, in which signals from the sound sources are combined,
wherein the apparatus is configured to combine a first input signal, or a processed version thereof, with a delayed and scaled version of a second input signal, to acquire a first output signal;
wherein the apparatus is configured to combine a second input signal, or a processed version thereof, with a delayed and scaled version of the first input signal, to acquire a second output signal;
wherein the apparatus is configured to determine, using a random direction optimization:
a first scaling value, which is used to acquire the delayed and scaled version of the first input signal;
a first delay value, which is used to acquire the delayed and scaled version of the first input signal;
a second scaling value, which is used to acquire the delayed and scaled version of the second input signal; and
a second delay value, which is used to acquire the delayed and scaled version of the second input signal,
wherein the random direction optimization is such that candidate parameters form a candidates' vector, the candidates' vector being iteratively refined by modifying the candidates' vector in random directions,
wherein the random direction optimization is such that a metrics indicating the similarity, or dissimilarity, between the first and second output signals is measured, and the first and second output signals are selected to be those measurements associated with the candidate parameters associated with metrics indicating lowest similarity, or highest dissimilarity,
wherein the metrics is processed as a Kullback-Leibler divergence.

### Claim 2 (depends on 1)

The apparatus of claim 1, wherein the delayed and scaled version of the second input signal, to be combined with the first input signal, is acquired by applying a fractional delay to the second input signal.

### Claim 3 (depends on 1)

The apparatus of claim 1, wherein the delayed and scaled version of the first input signal, to be combined with the second input signal, is acquired by applying a fractional delay to the first input signal.

### Claim 4 (depends on 1)

The apparatus of claim 1, wherein the first and second scaling values and first and second delay values are acquired by minimizing an objective function.

### Claim 5 (depends on 1)

The apparatus of claim 1, configured to:
combine the first input signal, or a processed version thereof, with the delayed and scaled version of the second input signal in the time domain and/or in the z transform or frequency domain;
combine the second input signal, or a processed version thereof, with the delayed and scaled version of the first input signal in the time domain and/or in the z transform or frequency domain.

### Claim 6 (depends on 1)

The apparatus of claim 1,
wherein the optimization is performed in the z transform domain.

### Claim 7 (depends on 1)

The apparatus of claim 1,
wherein the optimization is performed in the time domain.

### Claim 8 (depends on 1)

The apparatus of claim 1,
wherein the optimization is performed in the frequency domain.

### Claim 9 (depends on 1)

The apparatus of claim 1, wherein the delay or fractional delay applied to the second input signal is indicative of the relationship and/or difference or arrival between:
the signal from the first source received by the first microphone; and
the signal from the first source received by the second microphone.

### Claim 10 (depends on 1)

The apparatus of claim 1, wherein the delay or fractional delay applied to the first input signal is indicative of the relationship and/or difference or arrival between:
the signal from the second source received by the second microphone; and
the signal from the second source received by the first microphone.

### Claim 11 (depends on 1)

The apparatus of claim 1, wherein the metrics is acquired in form of:
D
KL
⁡
(
P
⁢


⁢
Q
)
=
∑
n
⁢
P
⁡
(
n
)
⁢
log
⁡
(
P
⁡
(
n
)
Q
⁡
(
n
)
)
wherein P (n) is an element associated with the first input signal and Q(n) is an element associated with the second input signal.

### Claim 12 (depends on 1)

The apparatus of claim 1, wherein the metrics is acquired in form of:
D
⁡
(
P
0
,
P
1
)
=
-
∑
n
⁢
[
P
0
⁡
(
n
)
⁢
log
⁡
(
P
0
⁡
(
n
)
P
1
⁡
(
n
)
)
+
P
1
⁡
(
n
)
⁢
log
⁡
(
P
1
⁡
(
n
)
P
0
⁡
(
n
)
)
]
wherein P1(n) is an element associated with the first input signal and P2(n) is an element associated with the second input signal.

### Claim 13 (depends on 1)

The apparatus of claim 1, further configured to transform, into a frequency domain, information associated with the acquired first and second output signals.

### Claim 14 (depends on 1)

The apparatus of claim 1, further configured to encode information associated with the acquired first and second output signals.

### Claim 15 (depends on 1)

The apparatus of claim 1, further configured to store information associated with the acquired first and second output signals.

### Claim 16 (depends on 1)

The apparatus of claim 1, further configured to transmit information associated with the acquired first and second output signals.

### Claim 17 (depends on 1)

The apparatus of claim 1 and at least one of a first microphone for acquiring the first input signal and a second microphone for acquiring the second input signal.

### Claim 18 (depends on 1)

The apparatus of claim 1, configured to sum a plurality of products between:
a respective element of a first set of normalized magnitude values, and
a logarithm of a quotient formed on the basis of:
the respective element of the first set of normalized magnitude values; and
a respective element of a second set of normalized magnitude values,
in order to obtain a value describing a similarity, or dissimilarity, between a signal portion described by the first set of normalized magnitude values and a signal portion described by the second set of normalized magnitude values.

### Claim 19 (depends on 18)

The apparatus of claim 18, wherein, for at least one of the first and second input signals, the respective element is based on the candidate first or second outputs signal as acquired from the candidate parameters.

### Claim 20 (depends on 19)

The apparatus of claim 19, wherein for at least one of the first and second input signals, the respective element is acquired as a fraction between:
a value associated with a candidate first or second output signal; and
a norm associated with the previously acquired values of the first or second output signal.

### Claim 21 (depends on 19)

The apparatus of claim 19, wherein for at least one of the first and second input signals, the respective element is acquired by
P
i
⁡
(
n
)
=

s
i
′
⁡
(
n
)



s
i
′


1
(
7
)

### Claim 22

A method for acquiring a plurality of output signals associated with different sound sources on the basis of a plurality of input signals, in which signals from the sound sources are combined,
the method comprising:
combining a first input signal, or a processed version thereof, with a delayed and scaled version of a second input signal, to acquire a first output signal;
combining a second input signal, or a processed version thereof, with a delayed and scaled version of the first input signal, to acquire a second output signal;
determining, using a random direction optimization, at least one of:
a first scaling value, which is used to acquire the delayed and scaled version of the first input signal;
a first delay value, which is used to acquire the delayed and scaled version of the first input signal;
a second scaling value, which is used to acquire the delayed and scaled version of the second input signal; and
a second delay value, which is used to acquire the delayed and scaled version of the second input signal,
wherein the random direction optimization is such that candidate parameters form a candidates' vector, the candidates' vector begin iteratively refined by modifying the candidates' vector in random directions,
wherein the random direction optimization is such that a metrics indicating the similarity, or dissimilarity, between the first and second output signals is measured, and the first and second output signals are selected to be those measurements associated with the candidate parameters associated with the metrics indicating lowest similarity, or highest dissimilarity,
wherein the metrics is processed as a Kullback-Leibler divergence.

### Claim 23 (depends on 1, 22)

The method of claim 22, configured to use an apparatus according to claim 1.

### Claim 24 (depends on 22)

The method of claim 22, wherein the fractional delay is indicative of the relationship and/or difference between the delay of the signal arriving at the first microphone from the second source and the delay of the signal arriving at the second microphone from the second.

### Claim 25

A non-transitory digital storage medium having a computer program stored thereon to perform the method for acquiring a plurality of output signals associated with different sound sources on the basis of a plurality of input signals, in which signals from the sound sources are combined,
the method comprising:
combining a first input signal, or a processed version thereof, with a delayed and scaled version of a second input signal, to acquire a first output signal;
combining a second input signal, or a processed version thereof, with a delayed and scaled version of the first input signal, to acquire a second output signal;
determining, using a random direction optimization, at least one of:
a first scaling value, which is used to acquire the delayed and scaled version of the first input signal;
a first delay value, which is used to acquire the delayed and scaled version of the first input signal;
a second scaling value, which is used to acquire the delayed and scaled version of the second input signal; and
a second delay value, which is used to acquire the delayed and scaled version of the second input signal,
wherein the random direction optimization is such that candidate parameters form a candidates' vector, the candidates' vector begin iteratively refined by modifying the candidates' vector in random directions,
wherein the random direction optimization is such that a metrics indicating the similarity, or dissimilarity, between the first and second output signals is measured, and the first and second output signals are selected to be those measurements associated with the candidate parameters associated with the metrics indicating lowest similarity, or highest dissimilarity,
wherein the metrics is processed as a Kullback-Leibler divergence,
when said computer program is run by a computer.