Abstract: Techniques for using a network to decompose an HRTF data set to generate approximation data and to render an audio signal using the approximation data are disclosed herein. An input HRTF data set is fed into the network, which then generates approximation data that includes mixing channel gains, FIR filter coefficients, and basis filter shapes. This approximation data controls various components in the network. When the input HRTF data set is fed as input into the network, then the output of the network is an output approximated HRTF data set. The network iteratively fine tunes the approximation data until the output approximated HRTF data set sufficiently matches the input HRTF data set. After the approximation data is sufficiently tuned, the approximation data is later used by the network to render an audio signal.

Abstract: The description relates to representing acoustic characteristics of real or virtual scenes. One method includes generating directional impulse responses for a scene. The directional impulse responses can correspond to sound departing from multiple sound source locations and arriving at multiple listener locations in the scene. The method can include processing the directional impulse responses to obtain coherent sound signals and incoherent sound signals. The method can also include encoding first perceptual acoustic parameters from the coherent sound signals and second perceptual acoustic parameters from the incoherent sound signals, and outputting the encoded first perceptual acoustic parameters and the encoded second perceptual acoustic parameters.

Abstract: Techniques for using a network to decompose an HRTF data set to generate approximation data and to render an audio signal using the approximation data are disclosed herein. An input HRTF data set is fed into the network, which then generates approximation data that includes mixing channel gains, FIR filter coefficients, and basis filter shapes. This approximation data controls various components in the network. When the input HRTF data set is fed as input into the network, then the output of the network is an output approximated HRTF data set. The network iteratively fine tunes the approximation data until the output approximated HRTF data set sufficiently matches the input HRTF data set. After the approximation data is sufficiently tuned, the approximation data is later used by the network to render an audio signal.

Abstract: Techniques for using a network to decompose an HRTF data set to generate approximation data and to render an audio signal using the approximation data are disclosed herein. An input HRTF data set is fed into the network, which then generates approximation data that includes mixing channel gains, FIR filter coefficients, and basis filter shapes. This approximation data controls various components in the network. When the input HRTF data set is fed as input into the network, then the output of the network is an output approximated HRTF data set. The network iteratively fine tunes the approximation data until the output approximated HRTF data set sufficiently matches the input HRTF data set. After the approximation data is sufficiently tuned, the approximation data is later used by the network to render an audio signal.

Abstract: The description relates to representing acoustic characteristics of real or virtual scenes. One method includes generating directional impulse responses for a scene. The directional impulse responses can correspond to sound departing from multiple sound source locations and arriving at multiple listener locations in the scene. The method can include processing the directional impulse responses to obtain coherent sound signals and incoherent sound signals. The method can also include encoding first perceptual acoustic parameters from the coherent sound signals and second perceptual acoustic parameters from the incoherent sound signals, and outputting the encoded first perceptual acoustic parameters and the encoded second perceptual acoustic parameters.