20240031764. EFFICIENT HRTF APPROXIMATION VIA MULTI-LAYER OPTIMIZATION simplified abstract (Microsoft Technology Licensing, LLC)

EFFICIENT HRTF APPROXIMATION VIA MULTI-LAYER OPTIMIZATION

Organization Name

Microsoft Technology Licensing, LLC

Inventor(s)

Mick Kekoa Marchan of Seattle WA (US)

Andrew Stewart Allen of San Diego CA (US)

EFFICIENT HRTF APPROXIMATION VIA MULTI-LAYER OPTIMIZATION - A simplified explanation of the abstract

This abstract first appeared for US patent application 20240031764 titled 'EFFICIENT HRTF APPROXIMATION VIA MULTI-LAYER OPTIMIZATION

Simplified Explanation

The abstract of this patent application describes techniques for using a network to decompose an HRTF (Head-Related Transfer Function) data set and generate approximation data. This approximation data includes mixing channel gains, FIR filter coefficients, and basis filter shapes, which control various components in the network. The network iteratively fine-tunes the approximation data until the output approximated HRTF data set matches the input HRTF data set. Once the approximation data is sufficiently tuned, it is used by the network to render an audio signal.

• The patent application describes a network-based approach to decompose and approximate HRTF data sets.
• The network generates approximation data that includes mixing channel gains, FIR filter coefficients, and basis filter shapes.
• The approximation data is fine-tuned iteratively until it sufficiently matches the input HRTF data set.
• The fine-tuned approximation data is later used by the network to render an audio signal.

Potential applications of this technology:

• Virtual reality (VR) and augmented reality (AR) audio systems.
• Spatial audio rendering for immersive experiences.
• Personalized audio systems for individual listeners.
• Audio processing for gaming and entertainment.

Problems solved by this technology:

• Complex and computationally intensive HRTF data decomposition.
• Difficulty in accurately approximating HRTF data sets.
• Lack of personalized and optimized audio rendering for individual listeners.

Benefits of this technology:

• Improved spatial audio quality and accuracy.
• Personalized and optimized audio experiences for individual listeners.
• Efficient and effective HRTF data decomposition and approximation.
• Enhanced immersion and realism in virtual and augmented reality audio.

Original Abstract Submitted

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.