Audio Masking Model for Accurate Bit Allocation at Dynamic Signal Levels
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing audio encoding technologies fail to accurately estimate masking thresholds based on known properties of human hearing, leading to inefficient bit allocation and reduced audio quality, particularly for signals with dynamic level changes.
Innovation Solution
A masking model that calculates masking thresholds using energy values and hearing thresholds in quiet, combined with sensitivity values to better emulate human hearing, allowing for more accurate bit allocation and improved audio quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If traditional psychoacoustic models are used to estimate masking thresholds, then the encoding process can be simplified, but the accuracy of bit allocation and audio quality deteriorate
Solution Approach 1:
The patent changes the parameters used in masking threshold estimation from traditional psychoacoustic models to a simplified model based on energy values and hearing thresholds in quiet. This parameter change reduces encoding complexity while improving accuracy for signals with dynamic level changes, as the new model directly uses measurable energy values rather than complex psychoacoustic calculations.
Solution Approach 2:
The patent replaces complex psychoacoustic modeling mechanisms with a simpler physical measurement-based approach. Instead of using complex auditory model calculations, the system uses direct measurements of energy values and hearing thresholds in quiet, substituting a simpler measurement-based mechanism for a complex theoretical model.
2Productivity
If masking threshold estimates exceed the JND noise level, then fewer bits are allocated, but audio quality deteriorates due to insufficient bit allocation
Solution Approach 1:
The patent implements a feedback mechanism where the masking threshold estimation is continuously refined based on the actual energy values and hearing thresholds measured from the audio signal. This feedback loop ensures that bit allocation is optimized, preventing both over-allocation (which wastes compression efficiency) and under-allocation (which degrades audio quality), by adjusting the masking threshold estimate to match the actual signal characteristics.
3Reliability
If masking threshold estimates are below the JND noise level, then more bits are allocated, but compression efficiency decreases due to unnecessary bit allocation
Solution Approach 1:
The patent introduces dynamics into the bit allocation process by making the masking threshold estimate adaptive to the actual signal energy levels and hearing thresholds. Rather than using a fixed or overly conservative threshold, the system dynamically adjusts the masking threshold based on real-time measurements of energy values and hearing thresholds, allowing optimal bit allocation that varies with signal characteristics to maximize both quality and compression efficiency.
4Ease of manufacture
If heuristic rules are used in masking threshold models, then model development is simplified, but the accuracy based on known properties of human hearing deteriorates
Solution Approach 1:
The patent segments the masking threshold estimation process into distinct, measurable components: energy value calculation and hearing threshold in quiet determination. This segmentation allows each component to be independently optimized and measured against known properties of human hearing, replacing vague heuristic rules with discrete, measurable steps that can be directly validated against auditory perception data.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The present disclosure relates to the field of audio coding, in particular, it relates to a method for encoding audio signals through a masking model based on a hearing threshold of frequency intervals of the audio signal and a measured energy of the audio signal for the corresponding frequency intervals. The disclosure further relates to an encoder that is capable of carrying out the audio encoding method.