Perception-Based Audio Parameter Quantization for Bitrate Reduction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing parametric stereo coding methods face challenges in balancing bit stream compression and audio quality, as high compression leads to unacceptable sound quality due to information loss, while low compression results in capacity issues.

Innovation Solution

Implementing a non-uniform scalar quantization scheme for audio parameters, where smaller step-sizes are used in sensitive ranges and larger step-sizes in less sensitive ranges, along with a scaling function to adjust quantization of the second parameter, reducing average bit consumption without compromising sound quality.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If high compression is applied to the bit stream, then bit stream capacity is improved, but audio quality deteriorates due to information loss

Engineering Contradiction:
Improvebit stream capacityVSAvoidaudio information
Core Design Contradiction:
Quantity of substanceVSLoss of information

Solution Approach 1:

The patent applies non-uniform quantization where different step-sizes are used for different ranges of the first parameter based on human sound perception sensitivity. Smaller step-sizes are used in ranges where perception is most sensitive, and larger step-sizes are used where perception is less sensitive. This local differentiation of quantization quality allows efficient bit stream compression while preserving perceptually important information.

Inventive Principle:
Principle #3Local quality

2Loss of information

If low compression is applied to the bit stream, then audio quality is maintained, but bit stream capacity deteriorates due to excessive data volume

Engineering Contradiction:
Improveaudio informationVSAvoidbit stream capacity
Core Design Contradiction:
Loss of informationVSQuantity of substance

Solution Approach 1:

The patent transforms the quantization approach by introducing a scaling factor that modifies the step-sizes of the second parameter based on the value of the first parameter. This parameter change allows the system to adapt the quantization precision dynamically, using coarser quantization (larger step-sizes) when appropriate and finer quantization (smaller step-sizes) when necessary, thereby reducing overall bit stream capacity while maintaining audio quality.

Inventive Principle:
Principle #35Parameter changes

3Device complexity

If uniform quantization is used for all parameter ranges, then device complexity is reduced, but measurement precision deteriorates in sensitive perception ranges

Engineering Contradiction:
Improvequantization scheme complexityVSAvoidparameter quantization precision
Core Design Contradiction:
Device complexityVSMeasurement precision

Solution Approach 1:

The patent implements non-uniform quantization with perception-based step-sizes for the first parameter and scaling-factor-based step-sizes for the second parameter. This local quality approach maintains relatively simple device complexity while significantly improving measurement precision in human sound perception sensitive ranges compared to uniform quantization.

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP3044788B1Non-uniform parameter quantization for advanced coupling
Publication Date: 2017.08.09 DOLBY INTERNATIONAL AB
  • EP3044788B1 patent drawingFigure 1
  • EP3044788B1 patent drawingFigure 2
  • EP3044788B1 patent drawingFigure 3

AI summary

The present disclosure provides methods, devices and computer program products for non-uniform quantization of parameters relating to parametric spatial coding of audio signals. The disclosure further relates to a method and apparatus for reconstructing an audio object in an audio decoding system taking the non-uniformly quantized parameters into account. According to the disclosure, such an approach renders it possible to reduce bit consumption without substantially reducing the quality of the reconstructed audio object.