Audio synchronization and delay estimation

Abstract

Systems and techniques are provided for audio synchronization and delay estimation. Audio metadata including a first discrete Fourier transform representation may be received. An audio signal may be pre-processed. A second discrete Fourier transform representation may be generated from the pre-processed audio signal. A correlation result in a discrete Fourier transform representation may be generated based on an element-wise multiplication of the first and second discrete Fourier transform representations. An inverse Fourier transform may be performed on the correlation result in a discrete Fourier transform representation to generate a correlated signal including samples that may have a position and a value. A relative delay value may be determined based on the position of a sample having a value with the greatest magnitude. Playback of a second audio signal may be adjusted based on a current delay value adjusted based on the relative delay value.

1 . A method comprising:
 pre-processing an audio signal at a transmitter to generate a transmitter pre-processed audio signal comprising samples, each sample comprising a value and having a position in the transmitter pre-processed audio signal;   reversing, by the transmitter, the positions of the samples of the transmitter pre-processed audio signal to generate a reversed audio signal;   generating, by the transmitter, a transmitter discrete Fourier transform representation from the reversed audio signal;   transmitting, by the transmitter, at least a section of the transmitter discrete Fourier transform representation to a receiver as audio metadata;   receiving, by the receiver, the audio metadata comprising the at least a section of the transmitter discrete Fourier transform representation;   pre-processing, by the receiver, a second audio signal;   generating, by the receiver, a receiver discrete Fourier transform representation from the pre-processed second audio signal;   generating, by the receiver, a correlation result in a discrete Fourier transform representation based on an element-wise multiplication of the at least a section of the transmitter discrete Fourier transform representation and the receiver discrete Fourier transform representation;   performing, by the receiver, an inverse Fourier transform on the correlation result in a discrete Fourier transform representation to generate a correlated signal comprising one or more samples, each sample of the correlated signal having a position in the correlated signal and comprising a value;   determining, by the receiver, a relative delay value based on the position in the correlated signal of a sample comprising a value with the greatest magnitude of the values of the samples of the correlated signal; and   adjusting, by the receiver, playback of a third audio signal based on a current delay value adjusted based on the relative delay value.