Psychoacoustic Masking Model for Dynamic Audio Bit Allocation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing audio encoding models fail to accurately capture the masking behavior of the human hearing system, leading to inefficient bit allocation and reduced audio quality, particularly for signals with dynamic level changes.
Innovation Solution
A psychoacoustic model that calculates masking thresholds based on sensitivity values and energy values for each frequency band, adjusting the excitation function to better match human hearing characteristics, thereby improving bit allocation and audio quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional psychoacoustic models are used to estimate masking thresholds, then the encoding process can proceed with standard compression techniques, but the bit allocation is inefficient and audio quality deteriorates for signals with dynamic level changes
Solution Approach 1:
The patent applies dynamics by transitioning from static masking threshold estimates to dynamic estimates that adapt to the actual signal characteristics. The model continuously adjusts masking thresholds based on measured signal energy and frequency content, allowing the encoder to optimize bit allocation in real-time according to the specific audio signal being processed, thereby improving both quality and efficiency.
Solution Approach 2:
The patent changes key parameters in the masking threshold calculation by incorporating signal energy measurements and frequency band analysis. Instead of using fixed psychoacoustic models, the system dynamically adjusts masking thresholds based on actual signal parameters (energy, frequency distribution), enabling optimal bit allocation that adapts to varying audio conditions.
2Productivity
If masking thresholds are overestimated, then fewer bits are allocated to frequency bands, but this results in audible distortion and reduced audio quality
Solution Approach 1:
The patent implements feedback by measuring the actual signal energy and using this information to adjust masking threshold estimates. The system continuously monitors signal characteristics and modifies masking thresholds accordingly, creating a closed-loop system that prevents both over- and under-allocation of bits while maintaining optimal audio quality.
3Reliability
If masking thresholds are underestimated, then more bits are allocated to frequency bands, but this wastes bits that could be allocated to neighboring frequency bands
Solution Approach 1:
The patent applies local quality by analyzing and allocating bits independently to each frequency band based on its specific characteristics. Instead of using uniform masking thresholds across all bands, the system calculates localized masking thresholds for each frequency band, considering its specific energy content and acoustic properties, thereby optimizing bit allocation locally while maintaining overall efficiency.
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.