Ambisonic Encoding Matrix for 3D Audio Directional Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for recording 3D sound in VR applications face limitations with spatial aliasing and microphone placement constraints, particularly in consumer devices like phones and tablets, which often use omnidirectional microphones, leading to spectral defects and signal-to-noise ratio issues.
Innovation Solution
A device and method that utilize at least five directive microphones to derive look direction angles, calculate a decoding matrix, invert it to obtain an encoding matrix, and encode microphone signals to generate First Order Ambisonic (FOA) signals, enabling more robust encoding over a larger frequency bandwidth and directional range.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Volume of moving object
If omnidirectional microphones are used in consumer devices, then the device can be made compact and portable, but spectral defects and spatial aliasing occur at higher frequencies
Solution Approach 1:
The patent segments the omnidirectional microphone signals by creating multiple virtual first-order microphones through mathematical processing. Instead of using a single omnidirectional microphone, the system divides the omnidirectional signal into directional components using gradient calculations, effectively segmenting the omnidirectional capture into multiple directional perspectives that can be processed independently to reduce spatial aliasing.
Solution Approach 2:
The patent introduces virtual first-order microphones as an intermediary between the physical omnidirectional microphones and the final Ambisonic decoding. These virtual microphones act as a mathematical mediator that transforms the omnidirectional pressure signals into directional gradient signals, enabling the system to achieve first-order Ambisonic encoding without requiring physical directive microphones that would increase device complexity.
2Measurement precision
If directive microphones are used to capture sound in all directions, then directional accuracy improves, but the microphones must be placed in free-field conditions which is not suitable for consumer devices
Solution Approach 1:
The patent creates virtual copies of directive microphones through mathematical processing of omnidirectional microphone signals. Instead of physically placing directive microphones in free-field conditions, the system calculates virtual first-order microphone signals that replicate the directional response characteristics. This copying approach allows the system to achieve directional accuracy without the physical constraints of free-field placement.
Solution Approach 2:
The patent replaces the mechanical system of physical directive microphones with a mathematical processing system. Instead of using physical microphones with specific directional patterns that require free-field placement, the system uses omnidirectional microphones combined with gradient calculations and matrix operations to achieve the same directional encoding效果. This substitution eliminates the mechanical constraints while preserving the functional benefits.
3Reliability
If a dense distribution of microphones is used to sample the soundfield, then spectral accuracy improves, but the number of microphones required becomes too large for consumer applications
Solution Approach 1:
The patent uses partial action by implementing a simplified version of dense microphone sampling. Instead of using a full spherical microphone array with many capsules, the system uses a limited number of omnidirectional microphones (as few as two) and applies mathematical processing to achieve first-order Ambisonic encoding. This partial approach provides sufficient spectral accuracy for consumer applications without requiring excessive microphones.
Solution Approach 2:
The patent changes the parameter of microphone quantity from many physical microphones to fewer microphones with mathematical processing. By transforming the system from a dense physical array to a sparse array with virtual microphone calculations, the patent maintains spectral accuracy through parameter transformation rather than physical multiplication. The gradient calculations and matrix operations compensate for the reduced number of physical sensors.
4Adaptability or versatility
If virtual first-order microphones are generated from omnidirectional microphones, then device compatibility improves, but signal-to-noise ratio deteriorates at low frequencies
Solution Approach 1:
The patent merges multiple omnidirectional microphone signals to create the virtual first-order microphones. By combining the signals from multiple omnidirectional microphones through gradient calculations, the system improves the signal-to-noise ratio through signal averaging and coherence. The merging of multiple measurements at each frequency provides redundancy that enhances low-frequency signal quality while maintaining device compatibility.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The invention is related to the technical field of audio recording of 3D sound, for instance, for virtual reality (VR) or surround sound. The invention relates to VR compatible audio formats, i.e. First Order Ambisonic (FOA) signals. The invention in particular proposes a device and method, respectively, for obtaining a FOA signal from signals of at least four directive microphones, particularly at least five directive microphones. The device is configured to determine a look direction of each microphone, and to calculate a decoding matrix based on the determined look directions. The decoding matrix is a matrix suitable for decoding a FOA signal into the signals of the microphones. Further, the device is configured to invert the decoding matrix to obtain an encoding matrix, and to encode the signals of the microphones based on the encoding matrix to obtain the FOA signal.