Adaptive correction of loudspeaker using recurrent neural network
View Patent ↗An audio system is described that corrects for linear and nonlinear distortions. The system can include a physical loudspeaker system responsive to an audio input signal, an adaptive circuit, e.g., with a recurrent neural network, to correct for non-linear distortions from the loudspeaker.
1. An audio system, comprising:
a physical system including a loudspeaker configured to produce audio output in response to an audio input signal;
an audio processor to output a processed signal to the loudspeaker, the audio processor including a recurrent neural network to correct for non-linear distortions from the loudspeaker, the recurrent neural network including a first recurrent neural network to correct for non-linear distortions from the loudspeaker and a second recurrent neural network to predict performance of the loudspeaker receiving an output from the first recurrent neural network; and
an adaptive feedback system receiving an audio output from the loudspeaker and comparing the received audio output to a target to provide correction parameters to the recurrent neural network, wherein the recurrent neural network receives the audio input signal and outputs a desired output signal, and
a summing circuit to sum the audio output and the desired output signal to produce an error signal that is received as a control signal by the recurrent neural network,
the adaptive feedback system further configured to predict performance of the loudspeaker receiving an output from the first recurrent neural network and to provide corrective parameters to the second recurrent neural network.
2. The system of claim 1 , wherein the recurrent neural network receives the audio input signal and outputs a corrected audio signal to the loudspeaker.
3. The system of claim 2 , wherein the recurrent neural network outputs a drive signal loudspeaker.
4. The system of claim 3 , wherein the audio processor applies a target linear transfer function to the input signal to produce the processed signal for the loudspeaker.
5. The system of claim 1 wherein the recurrent neural network is a precorrector.
6. The system of claim 5 , wherein the recurrent neural network is trained using an error signal between an output from the loudspeaker and an output from a forward model.
7. The system of claim 1 , wherein the audio input signal is a multitone, sweep, overlapped log sweeps, and/or music signal.
8. An audio system, comprising:
a loudspeaker that includes non-linear distortion and linear distortion based on an audio signal input to the loudspeaker;
non-linear distortion removal parameters developed from a first recurrent neural network to correct for non-linear distortions from the loudspeaker and a second recurrent neural network to predict performance of the loudspeaker receiving an output from the first recurrent neural network and correct parameters of the first recurrent neural network; and
a summing circuit to sum the system output and the desired output signal to produce an error signal that is received as a control signal by both the first recurrent neural network and the second recurrent neural network;
circuitry to apply the non-linear distortion removal parameters to the audio signal in the loudspeaker.
9. The audio system of claim 8 , wherein the circuitry is in an amplifier that sends an audio signal corrected by the non-linear distortion removal parameters to the loudspeaker to reduce non-linear distortions at the loudspeaker in response to the audio signal.
10. The audio system of claim 9 , wherein the non-linear distortion removal parameters are in an audio signal correction matrix that are mathematically applied to an audio signal input to the amplifier that outputs a corrected audio output signal to the loudspeaker.
11. The audio system of claim 9 , wherein the matrix includes linear distortion correction parameters that are mathematically applied to the audio signal input to the amplifier that outputs the corrected audio output signal to the loudspeaker.
12. The audio system of claim 8 , wherein the first recurrent neural network receives the audio input signal and outputs a corrected audio signal to the second recurrent neural network and the second recurrent neural network outputs a cascade output signal.
13. The audio system of claim 12 , wherein the first recurrent neural network outputs the corrected audio signal to a loudspeaker system model that outputs a system output.
14. The audio system of claim 8 , wherein the first recurrent neural network is a precorrector and the second recurrent neural network is a forward model RNN.