Customized Audio Profiling from Mobile In-Ear Acoustic Capture
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for generating personalized audio profiles for listeners are expensive and not feasible with typical user computing devices, and existing approaches using anechoic chambers or neural networks may not accurately capture the intricate sound diffraction of the pinnae.
Innovation Solution
Generate personalized audio profiles using a mobile device with in-ear microphones, capturing sensor data and audio signals while the user moves the device around their head, determining impulse responses and transfer functions based on this data to tailor audio to the user's unique physical characteristics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If measurements are made in an anechoic chamber using audio equipment, then audio profile accuracy is improved, but cost and feasibility deteriorate
Solution Approach 1:
The patent uses visual data (images/videos) of the pinnae as a copy or proxy for direct acoustic measurements. Instead of performing expensive anechoic chamber measurements, the system captures visual representations of the ear structures and uses these to derive audio profiles through neural network processing, thereby achieving accurate personalization at low cost using consumer devices
Solution Approach 2:
The patent replaces the mechanical/acoustic measurement system (anechoic chamber equipment) with an optical system (camera/imaging device). By substituting visual capture and computational processing for physical acoustic measurements, the system eliminates the need for specialized facilities while maintaining measurement accuracy through digital image analysis and neural network-based audio profile generation
2Extent of automation
If neural networks are used to generate audio profiles, then automation is improved, but accuracy in capturing pinnae sound diffraction deteriorates
Solution Approach 1:
The patent performs preliminary capture and analysis of visual pinnae data before audio profile generation. By first obtaining high-quality visual images of the pinnae structures and preprocessing this data, the system prepares accurate input for the neural network, ensuring that the automated process has precise geometric information to work with, thereby maintaining diffraction accuracy while preserving automation
Solution Approach 2:
The patent transforms the problem from direct acoustic measurement parameters to visual geometric parameters. By changing the input parameters from acoustic signals to visual characteristics of the pinnae (shape, size, contours), the system enables accurate modeling of sound diffraction through image-based geometric analysis combined with neural network processing
3Measurement precision
If professional audio equipment is used, then audio profile quality is improved, but device complexity and accessibility deteriorate
Solution Approach 1:
The patent enables the consumer device itself to perform the measurement and profile generation functions. The smartphone or computing device uses its own camera, processor, and neural network capabilities to capture pinnae visual data and generate the audio profile without requiring external professional equipment, making the system self-sufficient and accessible
Solution Approach 2:
The patent makes the consumer device perform multiple functions: visual capture of pinnae, image processing, neural network inference, and audio profile generation. By designing a universal system that consolidates these functions into a single consumer device, the patent eliminates the need for separate professional equipment while maintaining high audio profile quality
Data Source
AI summary
Disclosed implementations for generating personalized audio. In response to receiving sensor data corresponding with a physical characteristic of a user, a model is scaled to the physical characteristic. A function representing an audio response is modified based on the scaled model to produce a modified function. An audio stream is generated based on the modified function.