Inter-channel Coding Graph for Multi-channel Audio

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current inter-channel coding methods for multi-channel audio signals face challenges in achieving high coding gain while maintaining computational efficiency, especially under the constraint of lossless coding where encoder side operations must be invertible on the decoder side.

Innovation Solution

A method involving the construction of a directed acyclic graph (DAG) using a basic graph with nodes representing audio channels and edges indicating prediction relationships, where the graph is optimized to minimize cumulated cost and ensure invertibility, utilizing techniques like Edmonds' algorithm to determine a minimum directed spanning tree for inter-channel encoding.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of information

If inter-channel coding is used to exploit dependencies among channels, then coding gain is improved, but computational complexity increases

Engineering Contradiction:
Improvecoding gainVSAvoidcomputational complexity
Core Design Contradiction:
Loss of informationVSDevice complexity

Solution Approach 1:

The patent segments the inter-channel coding process into two distinct phases: an encoder-side optimization phase that determines the optimal directed acyclic graph structure, and a decoder-side reconstruction phase that simply follows the stored graph structure. This segmentation allows complex computations to be confined to the encoder only, while the decoder maintains low complexity.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent performs preliminary action by pre-determining and storing the optimal directed acyclic graph structure and channel prediction relationships during encoding. This pre-computed structure is then transmitted to the decoder, which only needs to perform simple reconstruction operations without re-computing the optimal prediction relationships, thereby reducing decoder computational complexity.

Inventive Principle:
Principle #10Preliminary action

2Manufacturing precision

If lossless coding constraint is applied to ensure invertible operations, then decoding accuracy is improved, but encoding complexity increases

Engineering Contradiction:
Improvedecoding accuracyVSAvoidencoding complexity
Core Design Contradiction:
Manufacturing precisionVSDevice complexity

Solution Approach 1:

The patent introduces an intermediary element: the directed acyclic graph structure that mediates between the original multi-channel signal and the reconstructed signal. This graph structure serves as a blueprint that guides the reconstruction process, ensuring that the decoding operations are invertible and accurate while organizing the encoding complexity into a structured framework.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent changes the parameter representation by using a directed acyclic graph structure to describe channel relationships, where nodes represent channels and edges represent prediction relationships. This parameter transformation allows the system to maintain lossless reconstruction capability while managing encoding complexity through structured graph optimization.

Inventive Principle:
Principle #35Parameter changes

3Loss of information

If multiple source channels are used to predict a target channel, then coding gain is improved, but prediction order complexity increases

Engineering Contradiction:
Improvecoding gainVSAvoidprediction order complexity
Core Design Contradiction:
Loss of informationVSDevice complexity

Solution Approach 1:

The patent applies dynamics by making the prediction order flexible rather than fixed. The directed acyclic graph structure allows each channel to have a different prediction order depending on its relationship with other channels in the graph. This dynamic adaptation enables the system to use higher prediction orders only where beneficial, while maintaining lower complexity elsewhere.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The patent transitions from considering prediction order as a single-dimensional parameter to a multi-dimensional structure using directed acyclic graphs. The graph structure adds dimensional context by showing not just how many channels predict a target channel, but also the specific relationships and dependencies among channels, allowing for more efficient prediction strategies.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Data Source

PatentEP3467824B1Method and system for inter-channel coding
Publication Date: 2021.04.21 DOLBY LABORATORIES LICENSING CORP
  • EP3467824B1 patent drawingFigure 1A~1B
  • EP3467824B1 patent drawingFigure 1C~1D
  • EP3467824B1 patent drawingFigure 2A~2B

AI summary

A method (800) for performing inter-channel encoding of a multi-channel audio signal (501) comprising channel signals for N channels, with N being an integer, with N>1, is described. The method (800) comprises determining (801) a basic graph (210) comprising the N channels as nodes (111) and comprising directed edges (112) between at least some of the N channels. Furthermore, the method (800) comprises determining (802) an inter-channel coding graph (220) from the basic graph (210), such that the inter-channel coding graph (220) is a directed acyclic graph, and such that a cumulated a cumulated cost of the signals of the nodes (111) of the inter-channel coding graph (220) is reduced.