Deep Neural Network Active Noise Cancellation for Acoustic Variations
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing active noise cancellation (ANC) systems face challenges in adapting to variations in acoustic properties, user fit, and nonlinear speaker behavior, leading to suboptimal performance and stability issues.
Innovation Solution
Utilizing deep neural networks (DNNs) to estimate and adapt the transfer functions of feedforward and feedback filters, incorporating nonlinear elements and time-varying components into the ANC model, and implementing a unified framework for simultaneous estimation of both filters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional ANC systems use fixed transfer functions for feedforward and feedback filters, then the system structure remains simple, but the system cannot adapt to variations in acoustic properties, user fit, and nonlinear speaker behavior
Solution Approach 1:
The patent transforms the static, fixed transfer functions into dynamic, adaptive transfer functions that can change in real-time. The feedforward and feedback filters use adaptive algorithms to continuously update their coefficients based on incoming signals, allowing the system to adapt to varying acoustic conditions, user fit changes, and speaker nonlinearities without requiring a complete system redesign
Solution Approach 2:
The patent implements a feedback mechanism where the output of the feedback filter is fed back into the system and combined with the original noise signal. This feedback loop allows the system to continuously monitor its own performance and adjust the transfer functions accordingly, enabling adaptation to changing conditions while maintaining system stability through controlled feedback gain
2Object-affected harmful factors
If ANC systems cancel noise effectively, then noise reduction performance improves, but the quality of desirable content in the sound signal deteriorates
Solution Approach 1:
The patent divides the noise cancellation task into two separate, specialized filters: a feedforward filter that primarily handles external noise cancellation and a feedback filter that handles residual noise and speaker nonlinearities. Each filter is optimized for its specific function with dedicated transfer functions, allowing independent optimization of noise cancellation performance while preserving signal quality through specialized processing paths
Solution Approach 2:
The patent applies different processing characteristics to different parts of the signal path. The feedforward path uses one set of transfer function characteristics optimized for external noise, while the feedback path uses different characteristics optimized for residual noise and nonlinearities. This localized optimization allows each segment to excel at its specific task without compromising overall signal quality
Data Source
AI summary
An active noise cancellation system can include an input node for receiving a noise signal, an output node for providing an anti-noise signal, and one or more blocks each configured to provide a transfer function between its input and output, with the block being further configured to be capable of having its transfer function estimated by a deep neural network framework. Such an active noise cancellation system can be supported by a computation engine that can be implemented on a system-on-chip device which in turn can be included in an audio device such as a headset.