Two-Stage Audio Noise Suppression for Real-Time CPU Efficiency

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing communication platforms struggle to efficiently suppress non-stationary noises in real-time using AI-based noise suppression due to high CPU resource consumption, especially on client devices.

Innovation Solution

A two-stage noise suppression method is employed, first using DSP techniques to filter out stationary noise, followed by a Noisy Signal Classifier to determine if further AI-based noise suppression is needed, minimizing CPU usage by deploying AI-based noise suppression only when necessary.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If AI-based noise suppression is applied continuously to handle non-stationary noises, then noise suppression effectiveness is improved, but CPU resource consumption increases significantly

Engineering Contradiction:

Improvenoise suppression effectivenessVSAvoidCPU resource consumption

Core Design Contradiction:

ReliabilityVSUse of energy by moving object

Solution Approach 1:

The noise suppression system is segmented into two distinct stages: a first stage using DSP techniques for stationary noise filtering, and a second stage using AI-based techniques for non-stationary noise suppression. This segmentation allows the system to apply computationally intensive AI processing only when necessary, after initial filtering has removed stationary noise components.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system dynamically adjusts its processing approach by first applying DSP techniques to all audio signals, then selectively applying AI-based noise suppression only to signals that require it. This dynamic adaptation enables the system to maintain high noise suppression effectiveness while minimizing CPU resource consumption by avoiding unnecessary AI processing.

Inventive Principle:

Principle #15Dynamics

2Productivity

If low resource DSP techniques are used for noise filtering, then CPU efficiency is improved, but ability to filter non-stationary noises deteriorates

Engineering Contradiction:

ImproveCPU efficiencyVSAvoidnon-stationary noise filtering capability

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The filtering capability is segmented between two different approaches: DSP techniques handle stationary noise filtering efficiently, while AI-based techniques handle non-stationary noise filtering. By segmenting the functionality this way, the system achieves both high CPU efficiency for common cases and effective non-stationary noise filtering when needed.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The DSP-based noise suppression acts as an intermediary stage that processes audio signals first, removing stationary noise components. This intermediary processing reduces the computational burden on subsequent AI-based processing, allowing the system to maintain CPU efficiency while improving non-stationary noise filtering capability.

Inventive Principle:

Principle #24Intermediary (Mediator)

3Speed

If AI-based noise suppression is deployed on client devices for real-time processing, then real-time noise suppression is achieved, but device resource constraints are violated

Engineering Contradiction:

Improvereal-time processing speedVSAvoiddevice resource consumption

Core Design Contradiction:

SpeedVSUse of energy by moving object

Solution Approach 1:

The real-time processing workload is segmented into two parts: DSP-based processing that runs continuously on client devices for stationary noise removal, and AI-based processing that is selectively applied only when non-stationary noise is detected. This segmentation enables real-time processing capability while respecting device resource constraints by avoiding continuous heavy AI computation.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system applies partial AI-based processing only to the extent necessary - specifically, only when non-stationary noise is detected and only after initial DSP processing has been applied. This partial action approach enables real-time noise suppression for challenging audio scenarios while minimizing overall device resource consumption compared to continuous full AI processing.

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS20260038526A1Intelligent noise suppression for audio signals within a communication platform

Publication Date: 2026.02.05 ZOOM COMMUNICATIONS INC

AI summary

Techniques for intelligent noise suppression for audio signals within a communication platform are disclosed. In an example method, a computing system extracts multiple audio features from an audio signal, in which the audio signal is a raw waveform. The computing system provides the multiple audio features to a first neural network. The computing system classifies, using the first neural network, whether the audio signal contains noise beyond a noise threshold. The computing system, responsive to a classification that the audio signal contains noise beyond the noise threshold, applies artificial intelligence (“AI”)-based denoising to the audio signal to generate a denoised version of the audio signal.