Multi-Microphone Signal Processing for Ambient Noise Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing noise reduction techniques for mobile devices in noisy environments are not effective enough and require complex computations, failing to adequately separate speech from ambient noise.
Innovation Solution
A system utilizing multiple microphones, including a standard microphone, an 'in ear' speaker/microphone, and a bone conduction microphone, processes signals to estimate and subtract ambient noise, using equations and MMSE techniques to generate a corrected signal that enhances speech clarity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If traditional noise reduction techniques are used, then some noise filtering is achieved, but the speech clarity is degraded and computational complexity increases
Solution Approach 1:
The patent segments the speech signal processing into multiple independent stages: echo cancellation, noise reduction, and voice activity detection. Each stage processes specific aspects of the signal separately, allowing optimized processing for each function while maintaining overall speech quality.
Solution Approach 2:
The patent introduces an intermediary processing chain between the microphone input and the final output, including multiple filtering stages (Hanning window, pre-emphasis filter, autocorrelation-based pitch detection) that progressively refine the signal while preserving speech characteristics.
2Object-affected harmful factors
If multiple microphones are used to reduce ambient noise, then noise reduction effectiveness improves, but device complexity increases
Solution Approach 1:
The patent makes the mobile device's existing components (microphone, speaker, processor) perform multiple functions. The microphone serves both for capturing speech and for noise analysis, the speaker serves for both output and acoustic echo cancellation, reducing the need for additional dedicated noise reduction hardware.
Solution Approach 2:
The system uses the device's own acoustic path (speaker to microphone feedback) to generate and cancel acoustic echoes, and uses the ambient noise captured by the same microphone to perform noise reduction, making the system self-sufficient without requiring external or additional specialized components.
3Measurement precision
If complex signal processing algorithms are applied, then speech and noise separation improves, but computational cost increases
Solution Approach 1:
The patent applies partial processing to the signal by using voice activity detection to identify speech segments and applying noise reduction only during these segments. The processing is selective rather than continuous, reducing overall computational energy while maintaining speech quality during active speech periods.
Solution Approach 2:
The patent dynamically adjusts processing parameters including Hanning window size, pre-emphasis filter coefficients, and noise reduction strength based on detected speech characteristics and ambient noise levels, optimizing the balance between speech clarity and computational energy consumption for different scenarios.
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 solution effectively reduces ambient noise for both the near and far ends of a phone conversation, improving speech clarity with reduced computational complexity and cost, using a combination of microphones to isolate and cancel noise.
Implementation Method 1
a third input signal that is detected by a bone-conduction microphone
Data Source
AI summary
A system for processing sound, the system including: (a) a processor, configured to process a first input signal that is detected by a first microphone at a detection moment, a second input signal that is detected by a second microphone at the detection moment, and a third input signal that is detected by a bone-conduction microphone at the detection moment, to generate a corrected signal that is responsive to the first, second, and third input signals; and (b) a communication interface, configured to provide the corrected signal to an external system.