Directional Microphone Equalization Using Omnidirectional Reference
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Directional microphones in conferencing systems suffer from spectral peaks due to acoustic reflections, leading to an unpleasant 'boxy' artifact, while omnidirectional microphones pick up reverberation and noise, making them less desirable despite being less sensitive to reflections.
Innovation Solution
The spectral response of an omnidirectional microphone is used as a reference to develop scale factors for spectral equalization of directional microphones, decomposing signals into sub-bands and applying these scale factors to correct frequency imbalances caused by acoustic reflections.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If directional microphones are used, then audio directionality and noise rejection are improved, but spectral peaks and 'boxy' artifacts occur due to acoustic reflections
Solution Approach 1:
An omnidirectional microphone is introduced as an intermediary reference device. Its spectral response serves as a baseline that is compared against the directional microphone's spectral response to identify and correct unwanted peaks caused by acoustic reflections.
Solution Approach 2:
The spectral response parameters of the directional microphone are dynamically adjusted by applying scale factors derived from comparing its spectral response to that of the omnidirectional microphone. This equalization process modifies frequency response parameters to eliminate spectral peaks.
2Object-generated harmful factors
If omnidirectional microphones are used, then sensitivity to acoustic reflections is reduced, but reverberation and noise from all directions are picked up
Solution Approach 1:
The omnidirectional microphone serves as a reference intermediary that captures the overall acoustic environment including reflections. This reference signal is then used to compute scale factors that correct the directional microphone's spectral response, allowing the system to benefit from the omnidirectional mic's reflection insensitivity while maintaining directional noise rejection.
Solution Approach 2:
The spectral response characteristics of the omnidirectional microphone are copied and used as a reference template. By comparing and scaling the directional microphone's spectral response against this template, the system creates a corrected version that eliminates unwanted spectral peaks while preserving the directional microphone's advantageous noise rejection properties.
3Object-generated harmful factors
If spectral equalization is applied to directional microphones, then 'boxy' artifacts are eliminated, but system complexity increases
Solution Approach 1:
The system performs self-calibration by automatically comparing the spectral responses of the omnidirectional and directional microphones. The scale factors are computed and applied automatically without requiring manual intervention or complex external calibration equipment, making the equalization process self-service and relatively simple to implement.
Solution Approach 2:
The equalization is achieved by applying simple multiplicative scale factors to the spectral components of the directional microphone signal. This parameter-based approach is computationally efficient and avoids the need for complex filter designs or iterative optimization algorithms, thereby minimizing the increase in system complexity.
Data Source
Figure 1A
Figure 1B
Figure 2A~2D
AI summary
The spectral response of an omnidirectional microphone is used as a reference. This reference is compared to the spectral response of each directional microphone to develop scale factors that are applied to the directional microphone spectral response to perform spectral equalization. The outputs of the omnidirectional microphone and the directional microphones are decomposed into a series of sub-bands and the comparison and equalization is done for each sub-band. The equalized sub-bands are then converted into a time domain signal for further processing by the conference phone or video conference system.