Personalized HRTF Generation via Photogrammetry
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for generating head-related transfer functions (HRTFs) for audio monitoring using headphones are inadequate due to their inability to accurately account for individual user anatomy, leading to poor sound localization and fidelity, as they often rely on generic or anthropometrically-matched data rather than user-specific measurements.
Innovation Solution
A system and method that calculates a user's personalized HRTF using video material and supplemental images of a transparent measurement device superimposed over their biological details, allowing for precise data gathering and transmission of the HRTF in a standardized format, enabling accurate audio reproduction without the need for user calibration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If generic or anthropometrically-matched HRTF data is used, then device complexity is reduced, but measurement precision and sound localization accuracy deteriorate
Solution Approach 1:
The patent uses photogrammetry to create a digital copy (3D model) of the user's head and ear anatomy from photographs. This virtual model is then used to generate personalized HRTF data through acoustic simulation, eliminating the need for physical measurement devices while maintaining high measurement precision.
Solution Approach 2:
The patent replaces traditional mechanical measurement devices (microphones, probes) with a photogrammetry-based optical system. Instead of physically measuring head and ear dimensions with specialized equipment, the system uses standard camera photographs and computational algorithms to extract anatomical data and generate HRTF.
2Measurement precision
If user-specific anatomical measurements are taken, then measurement precision improves, but ease of operation and measurement time worsen
Solution Approach 1:
The system creates accurate anatomical copies from standard photographs taken with a regular camera. Users simply take photos of their head and ears from multiple angles, and the photogrammetry algorithm automatically processes these images to extract precise anatomical measurements without requiring manual measurement procedures.
Solution Approach 2:
The system performs preliminary processing of photographs to automatically extract anatomical features and create 3D models before HRTF generation. This preliminary action of automated feature extraction from images eliminates the need for users to perform complex manual measurements during the actual HRTF measurement process.
3Measurement precision
If traditional measurement devices are used, then measurement precision may be maintained, but device complexity and cost increase
Solution Approach 1:
The patent substitutes complex mechanical measurement devices with a photogrammetry system based on standard digital cameras and computational algorithms. This optical-computational approach replaces physical measurement tools while achieving equal or superior measurement precision through 3D modeling from multiple photograph angles.
Solution Approach 2:
Instead of using physical measurement devices that directly contact or scan the user's anatomy, the system creates digital copies (3D models) from photographs. This virtual copying approach eliminates the need for specialized measurement hardware while maintaining accurate anatomical representation for HRTF generation.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
According to an example aspect of the present invention, there is provided a method comprising: calculating a head-related transfer function for a user, the calculation based at least in part on gathered data comprising at least video material and at least one supplemental image, the at least one supplemental image comprising an image of a transparent measurement device superimposed over at least one biological detail of an user, and transmitting said head-related transfer function.