Dual-Path Voice Enhancement for Noisy Bone Conduction Audio
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing voice enhancement methods fail to effectively enhance sound quality and user experience in noisy environments due to the limitations of bone conduction sensors, which collect only low-frequency signals and are susceptible to noise interference.
Innovation Solution
A method involving the acquisition of time-domain microphone and bone conduction signals, followed by noise cancellation using deep neural network (DNN) models, high-pass and low-pass filtering, and signal combination to enhance sound quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If bone conduction sensors are used to collect voice signals, then noise resistance is improved, but sound quality deteriorates due to loss of high-frequency components
Solution Approach 1:
The patent combines bone conduction signals and air conduction signals through signal processing. The bone conduction signal provides noise resistance while the air conduction signal provides high-frequency components. The signals are merged in the frequency domain after separate processing to achieve both noise resistance and complete frequency spectrum.
Solution Approach 2:
The patent segments the voice signal processing into two separate paths: one for bone conduction signals and one for air conduction signals. Each path is processed independently with appropriate filtering and noise cancellation, then the results are combined to achieve the desired performance.
2Manufacturing precision
If air conduction is used to transmit voice signals, then sound quality is improved with rich voice spectrum, but noise interference increases from environmental noise
Solution Approach 1:
The patent merges the advantages of both air conduction and bone conduction by combining their respective signals. The air conduction path captures high-frequency components and rich voice spectrum, while the bone conduction path provides noise resistance. The combined output achieves both goals simultaneously.
Solution Approach 2:
The patent uses an intermediary signal processing system that separately processes air conduction and bone conduction signals before combining them. This intermediary processing allows independent optimization of each path and effective noise cancellation while preserving sound quality.
3Object-affected harmful factors
If only low-frequency bone conduction signals are collected, then noise resistance is improved, but listening experience deteriorates due to dull sound
Solution Approach 1:
The patent combines bone conduction signals and air conduction signals to achieve both noise resistance and pleasant listening experience. The air conduction signal restores the high-frequency components that make sound pleasant, while the bone conduction signal provides the noise resistance.
Solution Approach 2:
The patent changes the frequency parameter by adding high-frequency components from air conduction to the low-frequency bone conduction signal. This parameter change transforms the dull sound into a pleasant sound while maintaining the noise resistance property.
Data Source
AI summary
Disclosed are a voice enhancement method, apparatus and system and a computer-readable storage medium. The method includes acquiring a time-domain microphone signal and a time-domain bone conduction signal at the current moment; determining whether the signals are voice signals, if yes, performing a noise cancellation processing to the time-domain microphone signal by a pre-established DNN noise cancellation model, performing a frequency-domain noise cancellation processing to the time-domain bone conduction signal, if not, setting an output signal at the current moment as zero; performing a high-pass filtering processing to the time-domain microphone signal from which noise has been cancelled, to obtain a first output time-domain signal, performing a low-pass filtering processing to the time-domain bone conduction signal from which noise has been cancelled, to obtain a second output time-domain signal; obtaining an output time-domain signal at the current moment according to the first and second output time-domain signals.