Audio Enhancement Parameter Adjustment for Voice Clarity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for enhancing sound quality during voice calls fail to effectively remove noise corresponding to the voice band, leading to voice loss and require large computational resources for machine learning-based solutions that are not optimized for individual users.
Innovation Solution
An electronic device that communicates with a server to transmit audio signals, determines the signal-to-noise ratio, and updates sound quality enhancement parameters to enhance voice signals by filtering out noise, using machine learning models personalized for the user's voice.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If machine learning-based noise suppression is applied, then noise removal capability is improved, but computational resource consumption increases
Solution Approach 1:
The patent applies parameter changes by dynamically adjusting noise suppression parameters based on signal-to-noise ratio thresholds. The system changes operational parameters (suppression intensity, filtering strength) according to measured audio conditions, achieving effective noise removal only when necessary and at optimized intensity levels, thus reducing unnecessary computational resource consumption while maintaining noise removal capability.
2Adaptability or versatility
If generic machine learning models are used for noise suppression, then broad applicability is achieved, but voice fidelity deteriorates due to voice loss
Solution Approach 1:
The patent implements local quality by applying different noise suppression strategies tailored to specific audio conditions and user profiles. Instead of using a uniform generic model, the system adapts suppression parameters and model selection based on local characteristics such as individual user voice patterns, specific noise environments, and signal-to-noise ratio levels, thereby preserving voice fidelity while maintaining broad applicability through conditional adaptation.
Solution Approach 2:
The system employs dynamics by making noise suppression parameters and model selection adaptive rather than static. The patent dynamically adjusts suppression intensity, selects appropriate machine learning models, and modifies processing parameters based on real-time audio analysis and user-specific characteristics, enabling the system to optimize voice fidelity for each user while maintaining versatility across different usage scenarios.
3Object-affected harmful factors
If aggressive noise suppression is applied, then noise removal effectiveness is improved, but voice signal quality deteriorates due to voice loss
Solution Approach 1:
The patent applies partial action by implementing conditional noise suppression that activates only when signal-to-noise ratio falls below specific thresholds. Rather than continuously applying aggressive suppression, the system selectively applies noise removal only when and where needed, using partial suppression intensity matched to actual noise levels, thereby achieving effective noise removal while preserving voice signal quality through avoided over-suppression.
Solution Approach 2:
The system employs feedback mechanisms by continuously monitoring signal-to-noise ratio and adjusting suppression intensity accordingly. The patent uses real-time audio analysis feedback to modulate noise suppression strength, reducing or eliminating suppression when voice signals are dominant and applying stronger suppression only when noise levels exceed thresholds, thus maintaining voice signal quality while achieving effective noise removal through adaptive control.
Data Source
AI summary
In accordance with an aspect of the disclosure, an electronic device comprises a communication circuitry configured to establish a voice call with an external electronic device; a microphone; a memory configured to store a first sound quality enhancement parameter; and a processor, wherein the processor is configured to: obtain an audio signal associated with speech through the microphone, during the voice call; transmit, to a server, voice data based on the audio signal when the ratio is within a first range; transmit, to the server, noise data based on the audio signal, when the ratio is within a second range; receive an updated sound quality enhancement parameter from the server with the communication circuit during the voice call; and adjust the first sound quality enhancement parameter stored in the memory, based on the updated sound quality enhancement parameter received from the server.