# US 2023/0276170 A1 · Microphone system and beamforming method

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

- Number: US 2023/0276170 A1
- Document ID: US20230276170A1
- Status: Published application
- Kind: A1
- Application number: 17/895,980
- Filed: 2022-08-25
- Published: 2023-08-31
- Earliest claimed: 2022-02-25
- Assignees: British Cayman Islands Intelligo Technology Inc.
- Applicants: British Cayman Islands Intelligo Technology Inc.
- Inventors: Hua-Jun HONG; CHIH-SHENG CHEN; HSUEH-YING LAI; Yu-Pao TSAI; TSUNG-LIANG CHEN
- Web page: https://patentgrep.com/patent/US20230276170A1

## CPC

- G06N 3/084: Backpropagation, e.g. using gradient descent
- H04R 1/406: {microphones}
- G06F 1/1616: {with folding flat displays, e.g. laptop computers or notebooks having a clamshell configuration, with body parts pivoting to an open position around an axis parallel to the plane they define in closed position}
- G06F 1/1677: {for detecting open or closed state or particular intermediate positions assumed by movable parts of the enclosure, e.g. detection of display lid position with respect to main body in a laptop, detection of opening of the cover of battery compartment}
- G06F 1/1684: {Constructional details or arrangements related to integrated I/O peripherals not covered by groups G06F1/1635 - G06F1/1675}
- G06N 3/04: Architecture, e.g. interconnection topology
- G06N 3/08: Learning methods
- G10L 21/0224: Processing in the time domain
- H04R 3/005: {for combining the signals of two or more microphones (specially adapted for hearing aids H04R25/407)}
- G10L 21/0208: Noise filtering
- G10L 2021/02166: {Microphone arrays; Beamforming}
- H04R 2201/401: 2D or 3D arrays of transducers
- H04R 2410/01: Noise reduction using microphones having different directional characteristics
- H04R 2430/20: Processing of the output signals of the acoustic transducers of an array for obtaining a desired directivity characteristic (H04R2203/12 takes precedence)
- H04R 2499/11: Transducers incorporated or for use in hand-held devices, e.g. mobile phones, PDA's, camera's
- 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)
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G10L: SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING

## Abstract

A microphone system of the invention is applicable to an electronic device comprising an adjustable mechanism that causes a change in geometry of a microphone array. The microphone system comprises the microphone array, a sensor and a beamformer. The microphone array comprises multiple microphones that detect sound and generate multiple audio signals. The sensor detects a mechanism variation of the electronic device to generate a sensing output. The beamformer is configured to perform a set of operations comprising: performing a spatial filtering operation over the multiple audio signals using a trained model based on the sensing output, one or more first sound sources in one or more desired directions and one or more second sound sources in one or more undesired directions to generate a beamformed output signal originated from the one or more first sound sources.

## Claims

### Claim 1

A microphone system, applied to an electronic device comprising an adjustable mechanism that causes a change in geometry of a microphone array, comprising:
the microphone array comprising multiple microphones that detect sound and generate multiple audio signals;
a sensor for detecting a mechanism variation of the electronic device to generate a sensing output; and
a beamformer configured to perform a set of operations comprising:
performing spatial filtering over the multiple audio signals using a trained model based on the sensing output, one or more first sound sources in one or more desired directions and one or more second sound sources in one or more undesired directions to generate a beamformed output signal originated from the one or more first sound sources.

### Claim 2 (depends on 1)

The beamformer according to claim 1, wherein the operation of performing the spatial filtering further comprises:
determining a boundary between the one or more first sound sources and the one or more second sound sources using the trained model according to the sensing output; and
performing the spatial filtering over the multiple audio signals using the trained model based on the boundary to generate the beamformed output signal.

### Claim 3 (depends on 2)

The system according to claim 2, wherein the operation of determining the boundary comprises:
calculating coordinates of the multiple microphones according to the sensing output and locations of the multiple microphones disposed on the electronic device; and
determining the boundary between the one or more first sound sources and the one or more second sound sources using the trained model according to the coordinates of the multiple microphones.

### Claim 4 (depends on 1)

The system according to claim 1, wherein the set of operations further comprises:
performing a denoising operation over the multiple audio signals using the trained model based on the sensing output, the one or more first sound sources and the one or more second sound sources to generate the beamformed output signal without noise originated from the one or more first sound sources.

### Claim 5 (depends on 1)

The system according to claim 1, wherein the operation of performing the spatial filtering further comprises:
performing the spatial filtering over a feature vector for the multiple audio signals using the trained model based on the sensing output, the one or more first sound sources and the one or more second sound sources to generate the beamformed output signal originated from the one or more first sound sources;
wherein the set of operations further comprises:
extracting the feature vector from multiple spectral representations of the multiple audio signals;
wherein the feature vector comprises multiple magnitude spectrums, multiple phase spectrums and multiple phase-difference spectrums.

### Claim 6 (depends on 1)

The system according to claim 1, wherein the electronic device is a laptop, and the multiple microphones are respectively disposed on an outer rim of a laptop screen and C side, and wherein the adjustable mechanism is a hinge.

### Claim 7 (depends on 1)

The system according to claim 1, wherein the electronic device is a laptop, and the sensor is a gyroscope mounted on one of A side and B side.

### Claim 8 (depends on 1)

The system according to claim 1, wherein the electronic device is a laptop, and the sensor is a Lidar sensor mounted on one of B side and C side.

### Claim 9 (depends on 1)

The system according to claim 1, wherein the electronic device is a headset, and wherein the adjustable mechanism is a clamping headband, and the sensor is a rotary potentiometer disposed in a middle point of the clamping headband.

### Claim 10 (depends on 1)

The system according to claim 1, wherein the electronic device is a headset, and the multiple microphones are respectively disposed around two earcups.

### Claim 11 (depends on 1)

The system according to claim 1, wherein the electronic device is a wearable speaker, and wherein the adjustable mechanism is an adjustable neckband, and the sensor is a rotary potentiometer disposed in a middle point of the adjustable neckband.

### Claim 12 (depends on 1)

The system according to claim 1, wherein the electronic device is a wearable speaker and the multiple microphones are respectively embedded in two front ends of the wearable speaker.

### Claim 13 (depends on 1)

The system according to claim 1, wherein the trained model is a neural network that is trained with the one or more first sound sources, the one or more second sound sources, multiple sensing outputs and a training dataset, and wherein the training dataset is associated with a combination of clean single-microphone audio data and noisy single-microphone audio data originated from the one or more first sound sources and the one or more second sound sources.

### Claim 14

A beamforming method for a microphone system in an electronic device, the electronic device comprising an adjustable mechanism, the method comprising:
detecting sound by multiple microphones in a microphone array to generate multiple audio signals;
detecting a mechanism variation of the electronic device by the sensor to generate a sensing output, wherein the adjustable mechanism causes a change in geometry of the microphone array; and
performing a spatial filtering operation over the multiple audio signals using a trained model based on the sensing output, one or more first sound sources in one or more desired directions and one or more second sound sources in one or more undesired directions to generate a beamformed output signal originated from the one or more first sound sources.

### Claim 15 (depends on 14)

The method according to claim 14, wherein the step of performing the spatial filtering operation further comprises:
determining a boundary between the one or more first sound sources and the one or more second sound sources using the trained model according to the sensing output; and
performing the spatial filtering operation over the multiple audio signals using the trained model based on the boundary to generate the beamformed output signal.

### Claim 16 (depends on 15)

The method according to claim 15, wherein the step of determining the boundary comprises:
calculating coordinates of the multiple microphones according to the sensing output and locations of the multiple microphones disposed on the electronic device; and
determining the boundary between the one or more first sound sources and the one or more second sound sources using the trained model according to the coordinates of the multiple microphones.

### Claim 17 (depends on 14)

The method according to claim 14, further comprising:
performing a denoising operation over the multiple audio signals using the trained model based on the sensing output, the one or more first sound sources and the one or more second sound sources to generate the beamformed output signal without noise originated from the one or more first sound sources.

### Claim 18 (depends on 14)

The method according to claim 14, further comprising:
prior to the step of the spatial filtering operation, extracting a feature vector from multiple spectral representations of the multiple audio signals;
wherein the step of performing the spatial filtering operation further comprises:
performing the spatial filtering operation over the feature vector using the trained model based on the sensing output, the one or more first sound sources and the one or more second sound sources to generate the beamformed output signal originated from the one or more first sound sources;
wherein the feature vector comprises multiple magnitude spectrums, multiple phase spectrums and multiple phase-difference spectrums.

### Claim 19 (depends on 14)

The method according to claim 14, further comprising:
respectively disposing the multiple microphones on an outer rim of a laptop screen and C side;
wherein the electronic device is a laptop.

### Claim 20 (depends on 14)

The method according to claim 14, further comprising:
disposing the sensor on one of A side and B side;
wherein the sensor is a gyroscope and the electronic device is a laptop.

### Claim 21 (depends on 14)

The method according to claim 14, further comprising:
disposing the sensor on one of B side and C side;
wherein the sensor is a Lidar sensor and the electronic device is a laptop.

### Claim 22 (depends on 14)

The method according to claim 14, wherein the electronic device is a headset, and the adjustable mechanism is a clamping headband.

### Claim 23 (depends on 22)

The method according to claim 22, further comprising:
respectively disposing the multiple microphones around two earcups; and
disposing the sensor in a middle point of the clamping headband;
wherein the sensor is a rotary potentiometer.

### Claim 24 (depends on 14)

The method according to claim 14, wherein the electronic device is a wearable speaker, and the adjustable mechanism is an adjustable neckband.

### Claim 25 (depends on 24)

The method according to claim 24, further comprising:
respectively embedding the multiple microphones in two front ends; and
disposing the sensor in a middle point of the adjustable neckband;
wherein the sensor is a rotary potentiometer.

### Claim 26 (depends on 14)

The method according to claim 14, wherein the trained model is a neural network that is trained with the one or more first sound sources, the one or more second sound sources, multiple sensing outputs and a training dataset, and wherein the training dataset is associated with a combination of clean single-microphone audio data and noisy single-microphone audio data originated from the one or more first sound sources and the one or more second sound sources.
