Adaptive nullforming for selective audio pick-up

Abstract

Audio pickup systems and methods are provided to enhance an audio signal by removing noise components related to an acoustic environment. The systems and methods receive a primary signal and a reference signal. The reference signal is adaptively filtered and subtracted from the primary signal to minimize an energy content of a resulting output signal.

What is claimed is: 
     
         1 . A method of enhancing an audio signal, the method comprising:
 receiving a primary signal;   receiving one or more reference signals derived from one or more microphones;   adaptively filtering each of the one or more reference signals to provide one or more filtered signals;   providing a noise estimate signal as a combination of one or more of the one or more filtered signals;   subtracting the noise estimate signal from the primary signal to provide an output signal; and   adapting the adaptive filtering of each of the one or more reference signals to minimize an energy content of the output signal, such that a signal component produced by a new acoustic source is adapted by the adaptive filters, over time, to be included in the noise estimate signal.   
     
     
         2 . The method of  claim 1  further comprising monitoring the output signal for a desired signal, and stopping adapting the filtering of each of the plurality of reference signals upon detecting the desired signal. 
     
     
         3 . The method of  claim 2  wherein monitoring the output signal for a desired signal includes monitoring the output signal for a wake-up word. 
     
     
         4 . The method of  claim 1  further comprising providing the primary signal from a primary microphone. 
     
     
         5 . The method of  claim 1  further comprising combining a plurality of signals from a plurality of microphones to provide the primary signal. 
     
     
         6 . The method of  claim 1  further comprising separating each of the primary signal and each of the one or more reference signals into sub-bands. 
     
     
         7 . The method of  claim 1  further comprising operating a background adaptive filter upon each of the one or more reference signals, and wherein adapting the filtering of each of the one or more reference signals includes copying a set of filter weights from the background adaptive filter. 
     
     
         8 . An audio system, comprising:
 a primary input to receive a primary signal;   one or more reference inputs, each of the reference inputs to receive a reference signal;   one or more adaptive filters, each of the adaptive filters configured to adaptively filter one of the one or more reference signals to provide a filtered signal; and   one or more combiners configured to receive one or more of the filtered signals and subtract the one or more of the filtered signals from the primary signal to provide an output signal,   each of the one or more adaptive filters configured to adapt to a signal component produced by a new acoustic source, over time, such that the signal component is represented in the one or more filtered signals, causing the output signal to be adapted toward a substantially null response in the presence of the new acoustic source.   
     
     
         9 . The audio system of  claim 8  further comprising a detector configured to detect a desired signal content in the output signal and to pause adaptation of the plurality of adaptive filters in response to detecting the desired signal content. 
     
     
         10 . The audio system of  claim 9  wherein the detector is a wake-up word detector. 
     
     
         11 . The audio system of  claim 8  further comprising a plurality of microphones, at least one of the plurality of microphones configured to provide at least one of the reference signals to at least one of the one or more reference inputs. 
     
     
         12 . The audio system of  claim 11  further comprising a beamformer configured to receive signals from one or more of the plurality of microphones and to provide the primary signal to the primary input. 
     
     
         13 . A method of enhancing an audio signal, the method comprising:
 receiving a plurality of audio signals derived from a plurality of microphones;   adaptively filtering at least one of the plurality of audio signals to provide a noise estimate signal;   combining the noise estimate signal with at least one of the plurality of audio signals to provide a resulting signal, wherein the adaptive filtering adapts the noise estimate signal to minimize an energy content of the resulting signal;   monitoring the resulting signal for a predetermined signal content;   responsive to detecting the predetermined signal content in the resulting signal, stopping the adaptive filtering; and   responsive to not detecting the predetermined signal content in the resulting signal, continuing the adaptive filtering such that, over time, energy of the resulting signal is minimized.   
     
     
         14 . The method of  claim 13  further comprising providing the resulting signal to an audio device. 
     
     
         15 . The method of  claim 13  wherein the predetermined signal content is a wake-up word. 
     
     
         16 . The method of  claim 13  further comprising identifying one of the plurality of audio signals as a primary signal, wherein combining includes subtracting the noise estimate signal from the primary signal. 
     
     
         17 . The method of  claim 16  wherein the primary signal is provided by an array forming technique. 
     
     
         18 . The method of  claim 13  wherein stopping the adaptive filter comprises reverting filter coefficients to prior values. 
     
     
         19 . The method of  claim 13  wherein adaptively filtering includes operating a background filter and copying a set of filter coefficients from the background filter at intervals. 
     
     
         20 . The method of  claim 19  wherein stopping the adaptive filtering comprises not copying the set of filter coefficients from the background filter at the next interval.