Adaptive Audio Decimation Using Analysis-Guided Branch Switching

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Decimation processes for microphone signals require significant computational resources, particularly in devices with multiple microphones and limited power supply, such as mobile devices.

Innovation Solution

A multi-stage decimation process with a first branch providing lower quality decimation and a second branch with full decimation, where the second branch's usage is dependent on the outcome of audio signal analysis, such as noise detection, allowing for efficient processing by performing analysis on partially decimated signals.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If a full decimation process is applied to all audio input signals, then decimation quality is improved, but computational resources and power consumption increase significantly

Engineering Contradiction:
Improvedecimation qualityVSAvoidpower consumption
Core Design Contradiction:
Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The decimation process is segmented into two separate branches: a first branch providing lower quality decimation and a second branch providing full quality decimation. This segmentation allows the system to process signals through different pathways based on their characteristics, applying full decimation only when necessary.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system dynamically selects which branch to use for each audio signal based on real-time analysis. The audio signal analysis determines whether a signal should undergo full decimation (second branch) or lower quality decimation (first branch), making the processing path adaptive rather than static.

Inventive Principle:
Principle #15Dynamics

2Measurement precision

If a full decimation process is applied to all audio input signals, then decimation quality is improved, but computational resources increase

Engineering Contradiction:
Improvedecimation qualityVSAvoidcomputational efficiency
Core Design Contradiction:
Measurement precisionVSProductivity

Solution Approach 1:

The decimation process is segmented into two separate branches: a first branch providing lower quality decimation and a second branch providing full quality decimation. This segmentation allows the system to process signals through different pathways based on their characteristics, applying full decimation only when necessary.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The first branch performs partial decimation (lower quality) sufficient for noise detection purposes, while the second branch performs full decimation only when needed. This partial action approach avoids the excessive computational cost of applying full decimation to all signals.

Inventive Principle:
Principle #16Partial or excessive action

3Measurement precision

If audio signal analysis is performed on fully decimated signals, then analysis accuracy is improved, but processing time increases

Engineering Contradiction:
Improveanalysis accuracyVSAvoidprocessing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

Noise detection is performed as a preliminary action on partially decimated signals from the first branch before committing to full decimation. This preliminary analysis allows the system to identify signals that do not require full processing, saving time overall.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The first branch acts as an intermediary stage that provides sufficiently processed signals for preliminary noise detection. This intermediate processing level is adequate for the analysis task, avoiding the need to perform analysis on fully decimated signals.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentEP3916726A1Decimator for audio signals
Publication Date: 2021.12.01 NOKIA TECHNOLOGIES OY
  • EP3916726A1 patent drawingFigure 1
  • EP3916726A1 patent drawingFigure 2
  • EP3916726A1 patent drawingFigure 3~4

AI summary

Examples of the disclosure relate to an apparatus that provides a decimator for audio signals. The apparatus comprises means for: receiving one or more audio input signals and applying a multi-stage decimation process to the one or more audio input signals wherein the multi-stage decimation process comprises at least a first branch and a second branch. The apparatus also comprises means for applying audio signal analysis to the one or more audio input signals. The audio signal analysis extracts at least one audio signal parameter from the one or more audio input signals and is performed on the output of the first branch of the multi-stage decimation process. Use of the second branch of the multi-stage decimation process is dependent upon the outcome of the audio signal analysis.