Motor Noise Reduction Circuit Eigenfiltering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Digital cameras with video recording capabilities face the challenge of capturing audio signals while minimizing the noise from the zoom or focus motor, which is wideband and quasi-stationary, sharing the same frequency bands as the audio signal, making traditional DSP filtering techniques ineffective and potentially distorting the original audio.
Innovation Solution
A novel two-microphone noise reduction scheme using eigenfiltering, where linear filters are applied to the microphone signals to separate noise from the desired audio signal, with filter coefficients calculated offline based on noise-only and signal-only data sets, allowing for effective noise reduction without distorting the audio.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If traditional DSP filtering techniques (lowpass filtering, notch filter) are used to remove motor noise, then noise reduction is achieved, but the input audio signal is distorted
Solution Approach 1:
The patent segments the audio signal into two distinct components: motor noise and desired audio signal. By using two microphones positioned at different locations, the system creates separate observation channels that allow independent analysis and processing of each signal component through eigenfiltering operations
Solution Approach 2:
The patent introduces eigenfilters as an intermediary processing mechanism that operates on the microphone signals. These eigenfilters are computed from covariance matrices and serve as mathematical mediators that selectively attenuate motor noise while preserving the audio signal, avoiding direct distortion
2Object-affected harmful factors
If quieter zoom and focus motors are used to reduce noise, then noise level is reduced, but hardware cost increases
Solution Approach 1:
The patent replaces the mechanical solution (using quieter, more expensive motors) with a digital signal processing solution. By using eigenfiltering algorithms on standard microphones and processors, the system achieves noise reduction without requiring specialized low-noise motor hardware, thereby maintaining cost-effectiveness
3Object-affected harmful factors
If beamforming or active noise control methods are used, then noise reduction is achieved, but computational complexity increases
Solution Approach 1:
The patent performs preliminary computation of eigenfilters offline by calculating covariance matrices from recorded data and deriving eigenfilters through eigendecomposition. These pre-computed eigenfilters are then stored and applied during real-time operation, significantly reducing the computational burden during actual noise reduction while maintaining effectiveness
Data Source
AI summary
A method of reducing noise in an environment where the noise source is in a fixed location relative to a pair of microphones, such as in a camera with a zoom motor, involves receiving signals x1(t), x2(t) from the respective microphones, and filtering each of the signals x1(t), x2(t) with respective first and second linear filters having filter coefficients obtained by computing eigenfilters corresponding to data samples from the respective microphones for noise only and signal only conditions.