Abstract: Systems, methods, and machine-readable storage devices that receive an input signal representing audio captured using a microphone. The input signal includes portions that represent acoustic output from one or more audio sources, and a portion that represents other acoustic energy in the environment. A frequency domain representation of the input signal is iteratively modified to substantially reduce effects due to all but a selected one of the portions, from which an estimate of the power spectral density, PSD, of the selected portion is determined. Based upon the estimated PSD a noise or echo component is reduced, or a replacement noise is provided. The iterative modification involves a diagonalization of the cross-spectral density matrix to remove content coherent with a first audio input from the auto and cross-spectra of other signals.

Abstract: The technology described in this document can be embodied in a method that includes receiving a plurality of representations of the signal corresponding to samples of the signal within a frame of predetermined time duration, and estimating a power spectral density (PSD) for each of a plurality of frequency bins. The PSD for a particular frequency bin is estimated based on a smoothing parameter calculated from a noise estimate for the particular frequency bin as obtained from samples corresponding to a preceding frame. The method also includes generating, based on the PSD for each of the plurality of frequency bins, an estimate of the steady-state noise floor, and computing a measure of spectral flatness associated with the samples within the frame. The method also includes determining that the measure of spectral flatness satisfies a threshold condition, and in response, computing an updated estimate of the steady-state noise floor.

Abstract: The technology described herein can be embodied in a method for estimating a power spectral density of noise, the method including receiving an input signal representing audio captured using a microphone. The input signal includes a first portion that represents acoustic outputs from two or more audio sources, and a second portion that represents a noise component. The method also includes iteratively modifying a frequency domain representation of the input signal, such that the modified frequency domain representation represents a portion of the input signal in which effects due to the first portion are substantially reduced. The method further includes determining, from the modified frequency domain representation, an estimate of a power spectral density of the noise, and generating a control signal configured to adjust one or more gains of an acoustic transducer. The control signal is generated based on the estimate of the power spectral density of the noise.

Abstract: A system is provided for limiting distortion of an audio speaker. The system includes a first lowpass filter circuitry that is configured to receive a system input signal and generate a first lowpass filtered output signal. The system also includes a first limiter circuitry that is configured to limit the first lowpass filtered output signal. Limiting the first low pass filtered output signal includes reducing an amplitude of the first lowpass filtered output signal below a first predetermined threshold value, thereby to generate a first limited output signal. A second lowpass filter circuitry is configured to receive the first limited output signal and to generate a second lowpass filtered output signal.

Abstract: The technology described herein can be embodied in a method for estimating a power spectral density of noise, the method including receiving an input signal representing audio captured using a microphone. The input signal includes a first portion that represents acoustic outputs from two or more audio sources, and a second portion that represents a noise component. The method also includes iteratively modifying a frequency domain representation of the input signal, such that the modified frequency domain representation represents a portion of the input signal in which effects due to the first portion are substantially reduced. The method further includes determining, from the modified frequency domain representation, an estimate of a power spectral density of the noise, and generating a control signal configured to adjust one or more gains of an acoustic transducer. The control signal is generated based on the estimate of the power spectral density of the noise.

Abstract: The technology described herein can be embodied in a method for estimating a power spectral density of noise, the method including receiving an input signal representing audio captured using a microphone. The input signal includes a first portion that represents acoustic outputs from two or more audio sources, and a second portion that represents a noise component. The method also includes iteratively modifying a frequency domain representation of the input signal, such that the modified frequency domain representation represents a portion of the input signal in which effects due to the first portion are substantially reduced. The method further includes determining, from the modified frequency domain representation, an estimate of a power spectral density of the noise, and generating a control signal configured to adjust one or more gains of an acoustic transducer. The control signal is generated based on the estimate of the power spectral density of the noise.

Abstract: A system that performs noise estimation for an audio adjustment application comprises a coherence calculator that determines at least one coherence value between microphone signals generated by at least two microphones that each independently senses acoustic energy in a listening space. A first microphone of the at least two microphones generates a first microphone signal from the acoustic energy and a second microphone of the at least two microphones generates a second microphone signal from the acoustic energy. The acoustic energy comprises a combination of an audio signal transduced by one or more speakers and environmental noise of the acoustic energy that is local to the listening space. A noise estimate computation processor determines an estimate of a level of the environmental noise based on the at least one coherence value.

Abstract: A method of operating a situationally aware speaker associated with a virtual personal assistant (VPA) service provider that comprises receiving an indication of at least one parameter of an environment proximate the situationally aware speaker, and delivering the response to the vocal query to the user formatted as speech through an audio output of the situationally aware speaker, at least one audio parameter of the response set based on the indication of the at least one parameter.

Abstract: A method of operating a situationally aware speaker associated with a virtual personal assistant (VPA) service provider that comprises receiving an indication of at least one parameter of an environment proximate the situationally aware speaker, and delivering the response to the vocal query to the user formatted as speech through an audio output of the situationally aware speaker, at least one audio parameter of the response set based on the indication of the at least one parameter.

Abstract: The technology described in this document can be embodied in a method that includes receiving a plurality of representations of the signal corresponding to samples of the signal within a frame of predetermined time duration, and estimating a power spectral density (PSD) for each of a plurality of frequency bins. The PSD for a particular frequency bin is estimated based on a smoothing parameter calculated from a noise estimate for the particular frequency bin as obtained from samples corresponding to a preceding frame. The method also includes generating, based on the PSD for each of the plurality of frequency bins, an estimate of the steady-state noise floor, and computing a measure of spectral flatness associated with the samples within the frame. The method also includes determining that the measure of spectral flatness satisfies a threshold condition, and in response, computing an updated estimate of the steady-state noise floor.

Abstract: The technology described in this document can be embodied in a method that includes receiving a plurality of representations of the signal corresponding to samples of the signal within a frame of predetermined time duration, and estimating a power spectral density (PSD) for each of a plurality of frequency bins. The PSD for a particular frequency bin is estimated based on a smoothing parameter calculated from a noise estimate for the particular frequency bin as obtained from samples corresponding to a preceding frame. The method also includes generating, based on the PSD for each of the plurality of frequency bins, an estimate of the steady-state noise floor, and computing a measure of spectral flatness associated with the samples within the frame. The method also includes determining that the measure of spectral flatness satisfies a threshold condition, and in response, computing an updated estimate of the steady-state noise floor.

Abstract: The technology described in this document can be embodied in a method that includes receiving a plurality of representations of the signal corresponding to samples of the signal within a frame of predetermined time duration, and estimating a power spectral density (PSD) for each of a plurality of frequency bins. The PSD for a particular frequency bin is estimated based on a smoothing parameter calculated from a noise estimate for the particular frequency bin as obtained from samples corresponding to a preceding frame. The method also includes generating, based on the PSD for each of the plurality of frequency bins, an estimate of the steady-state noise floor, and computing a measure of spectral flatness associated with the samples within the frame. The method also includes determining that the measure of spectral flatness satisfies a threshold condition, and in response, computing an updated estimate of the steady-state noise floor.

Abstract: The technology described herein can be embodied in a method for estimating a power spectral density of noise, the method including receiving an input signal representing audio captured using a microphone. The input signal includes a first portion that represents acoustic outputs from two or more audio sources, and a second portion that represents a noise component. The method also includes iteratively modifying a frequency domain representation of the input signal, such that the modified frequency domain representation represents a portion of the input signal in which effects due to the first portion are substantially reduced. The method further includes determining, from the modified frequency domain representation, an estimate of a power spectral density of the noise, and generating a control signal configured to adjust one or more gains of an acoustic transducer. The control signal is generated based on the estimate of the power spectral density of the noise.

Abstract: A system that performs noise estimation for an audio adjustment application comprises a coherence calculator that determines at least one coherence value between microphone signals generated by at least two microphones that each independently senses acoustic energy in a listening space. A first microphone of the at least two microphones generates a first microphone signal from the acoustic energy and a second microphone of the at least two microphones generates a second microphone signal from the acoustic energy. The acoustic energy comprises a combination of an audio signal transduced by one or more speakers and environmental noise of the acoustic energy that is local to the listening space. A noise estimate computation processor determines an estimate of a level of the environmental noise based on the at least one coherence value.

Abstract: A signal input module receives at least one of an entertainment audio signal and a telephony audio signal from vehicle sound circuitry. A level control module executes gain level control logic to balance the entertainment audio signal and a telephony audio signal according to a ratio. A gain control signal is applied to at least one of the entertainment audio signal and the telephony audio signal. A routing module mixes the entertainment audio signal and the telephony audio signal and routes the mixed signal to an output channel associated with a speaker.

Abstract: Among other things, one or more non-transitory machine-readable media storing instructions are described. The stored instructions are executable by one or more processing devices to perform operations comprising analyzing an effect of noise in a spatial unit on sound in the spatial unit, at least part of the sound being produced by audio signals; selecting an adjustment curve among a group of adjustment curves based on one or more characteristics of the noise; and determining an amount of adjustment to be made to the audio signals based on the analyzed effect and the selected adjustment curve.

Abstract: A system that performs noise estimation for an audio adjustment application comprises a coherence calculator that determines at least one coherence value between microphone signals generated by at least two microphones that each independently senses acoustic energy in a listening space. A first microphone of the at least two microphones generates a first microphone signal from the acoustic energy and a second microphone of the at least two microphones generates a second microphone signal from the acoustic energy. The acoustic energy comprises a combination of an audio signal transduced by one or more speakers and environmental noise of the acoustic energy that is local to the listening space. A noise estimate computation processor determines an estimate of a level of the environmental noise based on the at least one coherence value.

Abstract: A method of operating a situationally aware speaker associated with a virtual personal assistant (VPA) service provider that comprises receiving an indication of at least one parameter of an environment proximate the situationally aware speaker, and delivering the response to the vocal query to the user formatted as speech through an audio output of the situationally aware speaker, at least one audio parameter of the response set based on the indication of the at least one parameter.

Abstract: Among other things, one or more non-transitory machine-readable media storing instructions are described. The stored instructions are executable by one or more processing devices to perform operations comprising analyzing an effect of noise in a spatial unit on sound in the spatial unit, at least part of the sound being produced by audio signals; selecting an adjustment curve among a group of adjustment curves based on one or more characteristics of the noise; and determining an amount of adjustment to be made to the audio signals based on the analyzed effect and the selected adjustment curve.

Abstract: A system is provided for limiting distortion of an audio speaker. The system includes a first lowpass filter circuitry that is configured to receive a system input signal and generate a first lowpass filtered output signal. The system also includes a first limiter circuitry that is configured to limit the first lowpass filtered output signal. Limiting the first low pass filtered output signal includes reducing an amplitude of the first lowpass filtered output signal below a first predetermined threshold value, thereby to generate a first limited output signal. A second lowpass filter circuitry is configured to receive the first limited output signal and to generate a second lowpass filtered output signal.