Audio Noise Reduction Using Dual Microphone Fourier Subtraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional digital cameras face challenges in effectively reducing noise from the optical lens drive during recording, as the noise pattern inside the lens housing is difficult to accurately obtain, leading to incomplete noise reduction in ambient sounds.
Innovation Solution
An audio processing apparatus with multiple microphones, including a first microphone for ambient sounds and a second microphone for noise sources, performs Fourier transforms, generates noise data using noise parameters, and applies subtraction to reduce noise from the ambient sound signals, utilizing an inverse Fourier transform for effective noise reduction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single microphone is used to collect noise for generating a noise pattern, then the device complexity is reduced, but the measurement precision of the noise pattern deteriorates because the noise inside the lens housing cannot be accurately obtained
Solution Approach 1:
The audio collection system is segmented into two independent microphone channels: a first microphone for collecting ambient sound and a second microphone for collecting noise from the lens housing. This segmentation allows each microphone to specialize in capturing specific sound sources, thereby improving the accuracy of noise pattern generation without requiring a single complex microphone system
Solution Approach 2:
The second microphone acts as an intermediary device specifically positioned to capture noise from the lens housing before it mixes with ambient sound. By introducing this intermediate measurement point, the system can isolate and accurately measure the lens driving noise component, which then serves as the basis for generating precise noise patterns for subtraction
2Device complexity
If spectral subtraction method is applied using noise collected by a single microphone, then the noise reduction process is simplified, but the noise reduction effectiveness deteriorates because an exact noise pattern cannot be obtained
Solution Approach 1:
The noise reduction process is segmented into distinct stages: first, separate noise collection using the second microphone; second, noise pattern generation from the isolated noise signal; third, spectral subtraction applying the generated pattern to the ambient sound from the first microphone. This segmentation ensures that the spectral subtraction method operates on accurate, isolated noise patterns rather than mixed sounds, thereby improving noise reduction effectiveness
Solution Approach 2:
The system performs preliminary noise collection and pattern generation before applying the spectral subtraction method to the ambient sound. By预先 (in advance) capturing the noise characteristics from the lens housing using the second microphone and generating accurate noise patterns beforehand, the subsequent noise reduction process can effectively remove lens driving noise without compromising reliability
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The apparatus effectively reduces noise from the optical lens drive, enhancing the quality of recorded ambient sounds by accurately isolating and subtracting noise patterns, resulting in improved audio data with reduced noise interference.
Implementation Method 1
a first conversion unit configured to perform a Fourier transform on an audio signal from the first microphone to generate a first audio signal
Implementation Method 2
a second conversion unit configured to perform a Fourier transform on an audio signal from the second microphone to generate a second audio signal
Implementation Method 3
a third conversion unit configured to perform an inverse Fourier transform on an audio signal from the subtraction unit
Data Source
AI summary
An audio processing apparatus includes a first microphone configured to obtain ambient sound, a second microphone configured to obtain noise from a noise source, a first conversion unit configured to perform a Fourier transform on an audio signal from the first microphone to generate a first audio signal, a second conversion unit configured to perform a Fourier transform on an audio signal from the second microphone to generate a second audio signal, a generation unit configured to generate noise data by using the second audio signal and a parameter related to the noise from the noise source, a subtraction unit configured to subtract the noise data from the first audio signal, and a third conversion unit configured to perform an inverse Fourier transform on an audio signal from the subtraction unit.