Personalized Binaural Rendering for Accurate Sound Localization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing binaural audio rendering technologies rely on generic head-related transfer functions (HRTFs) that fail to accurately localize sound sources, leading to incorrect perception and distortions, especially when used with headphones.
Innovation Solution
A personalized binaural audio rendering apparatus and method that adjusts panning trajectories using a personalized adjustment function to correct localization errors, employing a processing circuitry to determine an adjusted target direction and generate driving signals for left and right ear transducers based on user-specific HRTFs and interaural time difference corrections.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If generic HRTFs are used for binaural rendering, then the device complexity is reduced and ease of manufacture is improved, but the localization accuracy and audio perception quality deteriorate
Solution Approach 1:
The system performs preliminary measurement of the user's actual perceived sound directions for multiple reference targets, then pre-computes correction values stored in a lookup table. During actual binaural rendering, the system simply retrieves and applies these pre-computed corrections rather than performing complex real-time measurements, thus achieving personalized accuracy without real-time measurement complexity
Solution Approach 2:
The system creates a simplified copy or approximation of the user's actual HRTF characteristics by measuring perceived directions at discrete reference points and interpolating between them. This creates a personalized correction model that captures essential individual differences without requiring complete personalized HRTF measurement across all possible directions
2Measurement precision
If personalized HRTFs are obtained through traditional measurement methods, then the localization accuracy is improved, but the measurement process complexity and time requirement increase significantly
Solution Approach 1:
The system performs the complex measurement process in advance by determining the user's perceived sound directions for multiple reference targets before actual use. These measurements are stored as correction values that can be quickly applied during binaural rendering without requiring repeated measurements
Solution Approach 2:
The system measures perceived directions at a limited set of discrete reference targets (e.g., 9 directions) rather than continuously across all possible angles. This partial sampling is sufficient to capture the user's individual characteristics and enables accurate interpolation for intermediate directions, reducing measurement complexity while maintaining effectiveness
3Measurement precision
If personalized HRTFs are obtained through traditional measurement methods, then the localization accuracy is improved, but the time required for measurement and setup increases
Solution Approach 1:
The system performs personalized measurement and correction value computation in advance, storing results for rapid retrieval during actual use. This separates the time-consuming measurement phase from the usage phase, making the initial setup time a one-time investment that enables fast subsequent operations
Solution Approach 2:
The system uses a limited set of reference targets (e.g., 9 discrete directions) for measurement rather than exhaustive sampling. This partial measurement approach captures sufficient individual characteristics to enable accurate interpolation, significantly reducing measurement time while maintaining personalization benefits
Data Source
AI summary
An apparatus provides personalized binaural audio rendering of an input signal. The apparatus has a left ear transducer configured to generate a left ear audio signal and a right ear transducer configured to generate a right ear audio signal. Moreover, the apparatus has processing circuitry configured to determine, based on a current target direction of the input signal, an adjusted current target direction using a personalized adjustment function that describes a functional relationship between a plurality of reference target directions of a reference sound signal and a plurality of perceived reference target directions of the reference sound signal as perceived by a user. The processing circuitry is further configured to implement a target direction renderer configured to generate based on the input signal and the adjusted target direction a first driving signal for driving the left ear transducer and a second driving signal for driving the right ear transducer.