System and method for enhancing speech of target speaker from audio signal in an ear-worn device using voice signatures
An ear-worn device is provided that operates to isolate and individually treat the received speech of a target speaker or multiple target speakers from an audio input signal detected in a multi-speaker environment. The ear-worn device uses a machine learning model that receives a voice signature of each of one or more target speakers as input signals, to identify and isolate the component of the audio input signal attributable to the target speaker(s). Once isolated, the target speaker's speech may be enhanced, de-emphasized, or otherwise processed in a manner desired by the wearer of the ear-worn device. The wearer may use an external electronic device, e.g., a phone, to select one or more target speakers in a conversation and/or configure various settings associated with processing the speech on the ear-worn device.
1 . A system, comprising:
an electronic device; and
an ear-worn device in communication with the electronic device and comprising:
one or more microphones; and
a processor configured to run one or more neural networks;
wherein:
the electronic device is configured to:
receive user input for an identifier of a speaker;
add the identifier to a registry;
associate a voice signature with the identifier in the registry; and
transmit the voice signature to the ear-worn device; and
the ear-worn device is configured to:
detect, with the one or more microphones, an audio signal;
provide the audio signal to the processor;
receive the voice signature from the electronic device; and
process, with the processor and using the one or more neural networks, the audio signal using the voice signature.
2 . The system of claim 1 , wherein the electronic device is further configured to:
output a prompt to the speaker; and
collect an audio segment including speech from the speaker.
3 . The system of claim 2 , wherein the electronic device is configured, when outputting the prompt to the speaker, to display a script.
4 . The system of claim 2 , wherein the electronic device is further configured to extract the voice signature from the audio segment.
5 . The system of claim 4 , wherein the electronic device is configured, when extracting the voice signature from the audio segment, to use a voice signature neural network.
6 . The system of claim 2 , further comprising a server, and wherein the server is configured to:
receive the audio segment from the electronic device;
extract the voice signature from the audio segment; and
provide the voice signature to the electronic device.
7 . The system of claim 6 , wherein the server is configured, when extracting the voice signature from the audio segment, to use a voice signature neural network.
8 . The system of claim 1 , wherein the ear-worn device is configured, when processing the audio signal using the voice signature, to apply a larger gain to speech identified as being from the speaker based on the voice signature than to speech of another speaker.
9 . The system of claim 8 , wherein the ear-worn device is configured, when applying the larger gain to speech identified as being from the speaker, to:
use the voice signature to generate a mask; and
apply the mask to the audio signal to isolate the speech of the speaker from the audio signal.
10 . The system of claim 1 , wherein the ear-worn device is configured, when processing the audio signal using the voice signature, to increase a signal-to-noise ratio of speech of the speaker in the audio signal using the voice signature.
11 . The system of claim 1 , wherein:
the voice signature comprises a first voice signature;
the ear-worn device is configured, when processing the audio signal using the first voice signature, to process the audio signal using the first voice signature but not a second voice signature during a first time period; and
the ear-worn device is further configured to:
receive the second voice signature from the electronic device; and
process, with the processor and using the one or more neural networks, the audio signal using the second voice signature but not the first voice signature during a second time period.
12 . The system of claim 1 , wherein the electronic device is further configured to:
receive a user selection of the identifier associated with the voice signature; and
transmit the voice signature to the ear-worn device based on receiving the user selection.
13 . The system of claim 12 , wherein the electronic device is further configured to:
display, in a graphical user interface (GUI), multiple identifiers associated with multiple voice signatures in the registry, wherein the identifier is a first identifier of the multiple identifiers; and
receive the user selection from the GUI.
14 . The system of claim 12 , wherein the electronic device is further configured to:
receive the audio signal;
determine a match between the voice signature and the audio signal; and
display the identifier associated with the voice signature based on determining the match.
15 . The system of claim 14 , wherein the electronic device is configured to receive the audio signal from the ear-worn device.
16 . The system of claim 14 , wherein the electronic device is further configured to receive the audio signal using one or more microphones of the electronic device.
17 . The system of claim 1 , wherein the electronic device is further configured to:
receive the audio signal;
determine that the audio signal contains an unidentified voice signature; and
display an indication that the audio signal contains an unidentified voice signature.
18 . The system of claim 17 , wherein the electronic device is further configured to:
receive a user selection to add the unidentified voice signature to the registry; and
add the unidentified voice signature to the registry based on receiving the user selection.
19 . The system of claim 1 , wherein the electronic device is further configured to:
receive the audio signal;
determine a match between the voice signature and the audio signal; and
transmit the voice signature to the ear-worn device based on determining the match.
20 . The system of claim 1 , wherein the ear-worn device is further configured to process, with the processor and using the one or more neural networks, the audio signal using an own voice signature such that speech from a wearer of the ear-worn device is suppressed.