Reference signal generation for acoustic echo cancellation
First Claim
Patent Images
1. A speech interface system, comprising:
- a housing;
a speaker positioned at least partly within the housing and configured to generate output audio based at least in part on an output audio signal;
one or more input microphones positioned to produce an input audio signal, the input audio signal representing user speech and one or more echoed components of the output audio from the speaker;
a reference microphone positioned within a compartment of the housing, the reference microphone to produce a reference audio signal, wherein a relative magnitude of the output audio to the user speech is greater in the reference audio signal than in the input audio signal, wherein the compartment is disposed at least partly between the speaker and the one or more input microphones;
an adaptive filter configured to produce an estimated echo signal representing the one or more echoed components of the output audio from the speaker as represented by the input audio signal, based at least in part on the reference audio signal;
a subtraction component configured to subtract the estimated echo signal from the input audio signal to produce an echo-suppressed audio signal; and
one or more speech processing components configured to perform speech recognition on the echo-suppressed audio signal.
2 Assignments
0 Petitions
Accused Products
Abstract
An audio device may have an output speaker that produces audio within the environment of a user and one or more input microphones that capture speech and other sounds from the user environment. The audio device may use acoustic echo cancellation (AEC) to suppress echoed components of the speaker output that may be present in audio captured by the input microphones. The AEC may be implemented using an adaptive filter that estimates echoing based on an output reference signal. The output reference signal may be generated by a reference microphone placed near the speaker of the audio device.
-
Citations
24 Claims
-
1. A speech interface system, comprising:
-
a housing; a speaker positioned at least partly within the housing and configured to generate output audio based at least in part on an output audio signal; one or more input microphones positioned to produce an input audio signal, the input audio signal representing user speech and one or more echoed components of the output audio from the speaker; a reference microphone positioned within a compartment of the housing, the reference microphone to produce a reference audio signal, wherein a relative magnitude of the output audio to the user speech is greater in the reference audio signal than in the input audio signal, wherein the compartment is disposed at least partly between the speaker and the one or more input microphones; an adaptive filter configured to produce an estimated echo signal representing the one or more echoed components of the output audio from the speaker as represented by the input audio signal, based at least in part on the reference audio signal; a subtraction component configured to subtract the estimated echo signal from the input audio signal to produce an echo-suppressed audio signal; and one or more speech processing components configured to perform speech recognition on the echo-suppressed audio signal. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11)
-
-
12. An audio device, comprising:
-
a housing; a speaker positioned at least partly within the housing and configured to generate output audio; one or more input microphones configured to produce an input audio signal, the input audio signal representing user speech and one or more echoed components of the output audio from the speaker; a reference microphone positioned within a compartment of the housing such that the reference microphone is isolated from the user speech, the reference microphone configured to produce a reference audio signal, wherein a relative magnitude of the output audio to the user speech is greater in the reference audio signal than in the input audio signal, wherein the compartment is disposed at least partly between the speaker and the one or more input microphones; an adaptive filter configured to produce, based at least in part on the reference audio signal, an estimated echo signal representing the one or more echoed components of the output audio from the speaker as represented by the input audio signal; and an echo suppression element configured to suppress the one or more echoed components of the output audio as represented by the input audio signal using the estimated echo signal. - View Dependent Claims (13, 14, 15, 16, 17, 18, 19, 20)
-
-
21. A method, comprising:
-
producing, using a speaker located at least partly within a housing, output audio based at least in part on an output audio signal; receiving an input audio signal from one or more input microphones, wherein the input audio signal represents user speech and one or more components of the output audio from the speaker; receiving a reference audio signal from a reference microphone that is positioned within a compartment of the housing, wherein a relative magnitude of the output audio to the user speech is greater in the reference audio signal than in the input audio signal, and wherein the compartment is disposed at least partly between the speaker and the one or more input microphones; generating, based at least in part on the reference audio signal, an estimated echo signal representing the one or more components of the output audio from the speaker as represented by the input audio signal; and suppressing the one or more components of the output audio as represented by the input audio signal based at least in part on the estimated echo signal. - View Dependent Claims (22, 23, 24)
-
Specification