Primary-Ambient Audio Decomposition With Reduced Leakage

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current primary-ambient decomposition techniques in audio signal processing often result in artifacts such as 'leakage' of primary components into the ambience, which affects the quality of audio reproduction.

Innovation Solution

The proposed method employs vector-space primary-ambient decomposition, where audio signals are converted into subband representations using STFT, and then decomposed into primary and ambient components using principal component analysis (PCA) and orthogonality constraints, ensuring that the sum of components equals the original signal and satisfying desired orthogonality conditions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of manufacture

If real-valued multiplier methods are used to determine ambient components, then the decomposition process is simple, but artifacts such as leakage of primary components into ambience occur

Engineering Contradiction:
Improvesimplicity of decomposition processVSAvoidleakage artifacts in audio reproduction
Core Design Contradiction:
Ease of manufactureVSObject-generated harmful factors

Solution Approach 1:

The patent segments the audio signal into distinct primary and ambient components using independent vector spaces. By representing the audio signal as a sum of orthogonal vectors (primary component vector and ambient component vector), the method achieves complete separation without contamination, resolving the leakage artifact problem while maintaining mathematical simplicity through vector decomposition.

Inventive Principle:
Principle #1Segmentation

2Device complexity

If traditional primary-ambient decomposition is used, then computational complexity is low, but audio reproduction quality deteriorates due to artifacts

Engineering Contradiction:
Improvecomputational complexity of decompositionVSAvoidaudio reproduction quality
Core Design Contradiction:
Device complexityVSManufacturing precision

Solution Approach 1:

The patent introduces an intermediary orthogonal vector space as a mediator between the primary and ambient components. This orthogonal vector space acts as a mathematical intermediary that ensures complete separation of components through orthogonality constraints, eliminating leakage artifacts while adding only minimal computational overhead for vector projection calculations.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Object-generated harmful factors

If vector-space decomposition with orthogonality constraints is applied, then leakage artifacts are reduced, but algorithm complexity increases

Engineering Contradiction:
Improveleakage of primary components into ambienceVSAvoidcomplexity of decomposition algorithm
Core Design Contradiction:
Object-generated harmful factorsVSDevice complexity

Solution Approach 1:

The patent changes the mathematical parameters of the decomposition by imposing orthogonality constraints on the component vectors. By requiring that the primary component vector and ambient component vector be orthogonal (their dot product equals zero), the method fundamentally changes the decomposition parameters to achieve complete separation, eliminating leakage while the computational complexity remains manageable through efficient vector operations.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP2272169B1Adaptive primary-ambient decomposition of audio signals
Publication Date: 2017.09.06 CREATIVE TECHNOLOGY LTD
  • EP2272169B1 patent drawingFigure 1
  • EP2272169B1 patent drawingFigure 2(a)~2(c)
  • EP2272169B1 patent drawingFigure 3

AI summary

A stereo audio signal is processed to determine primary and ambient components by transforming the signal into vectors corresponding to subband signals, and decomposing the left and right channel vectors into ambient and primary components by matrix and vector operations. Principal component analysis is used to determine a primary component unit vector, and ambience components are determined according to a correlation-based cross-fade or an orthogonal basis derivation.