Audio Gain Interpolation Decoding for Continuous Waveforms

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing audio coding techniques, such as MPEG AAC, face issues with discontinuous temporal waveforms in audio signals due to uniform gain correction across frames, leading to degradation in sound quality when significant gain changes occur between frames.

Innovation Solution

A decoding method that reads encoded gain values and inclination values at sample positions within an audio signal, using linear or non-linear interpolation to calculate gain values between sample positions, allowing for precise volume correction and smoother waveform reproduction.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of information

If uniform gain correction is applied across all samples in one frame, then the code amount is reduced and processing is simplified, but the temporal waveforms become discontinuous between frames causing degradation in sound quality

Engineering Contradiction:
Improvecode amountVSAvoidsound quality
Core Design Contradiction:
Loss of informationVSManufacturing precision

Solution Approach 1:

The frame is divided into multiple gain application regions, each with its own gain value. This segmentation allows different gain values to be applied to different temporal segments within a frame, enabling smooth gain transitions while maintaining coding efficiency through region-based processing rather than sample-by-sample processing.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Gain values are determined in advance for each gain application region before actual audio processing. The gain values are calculated based on intersection points of straight lines representing gain changes, and these pre-determined gain values are then applied to corresponding regions, ensuring smooth transitions while reducing real-time computational complexity.

Inventive Principle:
Principle #10Preliminary action

2Manufacturing precision

If gain values are determined for each sample position through interpolation, then the temporal waveform continuity is improved and sound quality is enhanced, but the processing complexity and computational load increase

Engineering Contradiction:
Improvetemporal waveform continuityVSAvoidprocessing complexity
Core Design Contradiction:
Manufacturing precisionVSDevice complexity

Solution Approach 1:

Instead of processing each sample individually, the audio frame is segmented into gain application regions. This reduces the number of gain determination operations from thousands of samples to a manageable number of regions, significantly lowering computational complexity while maintaining waveform continuity through appropriate region-based gain application.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The approach changes from determining gain for every sample to determining gain for each region. By changing the parameter granularity from sample-level to region-level, the computational complexity is reduced while still achieving smooth gain transitions through the strategic placement of gain application regions and use of straight line intersections for gain value determination.

Inventive Principle:
Principle #35Parameter changes

3Measurement precision

If the number of gain application regions is increased, then the gain control precision is improved and sound quality is enhanced, but the code amount and processing overhead increase

Engineering Contradiction:
Improvegain control precisionVSAvoidcode amount
Core Design Contradiction:
Measurement precisionVSLoss of information

Solution Approach 1:

Rather than applying gain to every sample or using a very large number of regions, the method uses a selective number of gain application regions that are strategically positioned. This partial action approach achieves sufficient gain control precision for high-quality sound reproduction while avoiding the excessive code amount and processing overhead that would result from too many regions.

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentEP3089161B1Decoding device, method, and program
Publication Date: 2019.10.23 SONY GROUP CORP
  • EP3089161B1 patent drawingFigure 1~2
  • EP3089161B1 patent drawingFigure 3
  • EP3089161B1 patent drawingFigure 4

AI summary

The present technology relates to a decoding apparatus, a decoding method and a program which make it possible to obtain sound with higher quality. A demultiplexing circuit demultiplexes an input code string into a gain code string and a signal code string. A signal decoding circuit decodes the signal code string to output a time series signal. A gain decoding circuit decodes the gain code string. That is, the gain decoding circuit reads out gain values and gain inclination values at predetermined gain sample positions of the time series signal and interpolation mode information. An interpolation processing unit obtains a gain value at each sample position between two gain sample positions through linear interpolation or non-linear interpolation according to the interpolation mode based on the gain values and the gain inclination values. A gain applying circuit adjusts a gain of the time series signal based on the gain values. The present technology can be applied to a decoding apparatus.