Audio Signal Quantization Using Psychoacoustic Masking Thresholds
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current audio coding techniques for 3D sound scenes fail to guarantee minimal distortion in the listening domain, leading to suboptimal audio quality due to inadequate quantization noise management in multichannel systems.
Innovation Solution
A method for determining a quantization function based on psychoacoustic masking thresholds and inverse linear transformation errors, which minimizes noise in the listening domain by iteratively adjusting parameters to ensure compliance with perceptual criteria and reduce bit rate.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If conventional quantization methods are used on transformed components, then device complexity is reduced, but audio quality deteriorates due to quantization noise not being properly masked in the listening domain
Solution Approach 1:
The patent introduces an intermediary transformation process: quantization is performed in the transformed domain (e.g., ambisonic or surround domain), and then an inverse transformation is applied to map the quantized components back to the listening domain. This intermediary approach allows quantization noise to be properly masked by the audio signals in the listening domain, improving audio quality while maintaining relatively simple device complexity.
Solution Approach 2:
The patent changes the domain parameters by applying a linear transformation to convert audio signals from the listening domain to a different domain (e.g., ambisonic or surround domain) for quantization. This parameter change in the transformation domain allows for more effective quantization noise management when the inverse transformation is applied, resolving the contradiction between simplicity and quality.
2Device complexity
If quantization is performed in the transformed domain, then device complexity is reduced, but quantization noise increases in the listening domain
Solution Approach 1:
The inverse transformation acts as an intermediary that maps quantized components from the transformed domain back to the listening domain. This intermediary step ensures that quantization noise is distributed and masked by the audio signals in the listening domain, reducing the perceptibility of harmful quantization noise while maintaining the simplicity of quantization in the transformed domain.
Solution Approach 2:
The patent converts the potentially harmful quantization noise generated in the transformed domain into a beneficial situation by applying the inverse transformation. The noise is redistributed and masked by the audio signals in the listening domain, transforming the harmful quantization noise into an imperceptible artifact, thus converting harm into benefit.
3Manufacturing precision
If psychoacoustic masking thresholds are used for quantization, then audio quality is improved, but bit rate increases
Solution Approach 1:
The patent changes the domain parameters by performing quantization in the transformed domain rather than directly in the listening domain. This parameter change allows psychoacoustic masking to be effectively applied in the transformed domain, and the inverse transformation ensures that the masking效果 is preserved in the listening domain, achieving high audio quality without excessive bit rate increase.
Data Source
Figure 1~2
Figure 3
AI summary
The invention relates to a method for quantifying components ((yj)1=j=r ), wherein certain components are each determined based on a plurality of audio signals ( (sj)11=j=N) and can be calculated by the application of a linear conversion on the audio signals, said method comprising: determining a quantification function (Qm) to be applied to the components by testing a condition relative to an audio signal ( Si ) and depending on a comparison made between a psycho-acoustic masking threshold (Mm t(s,i) ) relative to the audio signal and a value determined based on the reverse linear conversion and quantification errors of the components by the function.