Multiplet-Based Matrix Mixing for Multichannel Audio Downmixing

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current downmixing and upmixing techniques for multichannel audio struggle to preserve precise panning behavior and localization, often resulting in artifacts and limited ability to separate independent signals that overlap in time and frequency, especially in surround sound systems with higher channel counts.

Innovation Solution

The multiplet-based spatial matrixing codec reduces channel counts by downmixing non-surviving channels onto multiplets of surviving channels, using pairwise, triplet, and quadruplet matrix rules to enable effective upmixing and preserve spatial accuracy, while also being backward compatible with legacy decoders.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Power

If constant-power panning is used in downmixing and upmixing, then signal power is maintained across channels, but precise panning behavior and localization are not preserved and artifacts occur

Engineering Contradiction:
Improvesignal powerVSAvoidpanning behavior and localization precision
Core Design Contradiction:
PowerVSMeasurement precision

Solution Approach 1:

The patent segments the downmixing and upmixing process into distinct matrixing operations with separate coefficient sets for different channel configurations. By dividing the audio signal into multiple channel groups and applying specific matrixing rules to each, the system preserves precise panning information while maintaining power balance, avoiding the artifacts caused by uniform constant-power approaches.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent changes the parameters of the matrixing coefficients based on the specific channel configuration and signal characteristics. Instead of using fixed constant-power coefficients, the system dynamically adjusts the matrixing parameters to preserve spatial localization accuracy while maintaining power balance, thereby eliminating the precision loss and artifacts associated with rigid constant-power panning.

Inventive Principle:
Principle #35Parameter changes

2Loss of substance

If matrixing is used to reduce channel count for storage and transmission, then file size and bandwidth are reduced, but the ability to separate and recover original channels is limited

Engineering Contradiction:
Improvestorage space and bandwidthVSAvoidchannel separation and recovery accuracy
Core Design Contradiction:
Loss of substanceVSLoss of information

Solution Approach 1:

The patent applies preliminary matrixing operations during encoding that preserve separable signal components. By pre-organizing the channel mixing using specific matrixing rules and coefficient sets, the system enables accurate channel separation and recovery during decoding, overcoming the limitations of traditional matrixing that loses channel information.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent incorporates feedback mechanisms where the decoder uses knowledge of the original channel configuration and matrixing coefficients to accurately separate and recover the original channels. This feedback approach allows the system to reverse the downmixing process with high fidelity, maintaining channel separation accuracy despite the reduction in channel count for storage and transmission.

Inventive Principle:
Principle #23Feedback

3Productivity

If high-channel count audio is processed with traditional downmixing techniques, then channel count is reduced, but spatial accuracy and signal separation are compromised

Engineering Contradiction:
Improvechannel count reduction efficiencyVSAvoidspatial accuracy and signal separation quality
Core Design Contradiction:
ProductivityVSManufacturing precision

Solution Approach 1:

The patent implements dynamic matrixing operations that adapt to the specific characteristics of high-channel count audio signals. By using dynamic coefficient selection and adaptive matrixing rules based on signal analysis, the system achieves efficient channel count reduction while preserving spatial accuracy and signal separation quality, overcoming the static limitations of traditional downmixing techniques.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP3444815A1Multiplet-based matrix mixing for high-channel count multichannel audio
Publication Date: 2019.02.20 DTS INC(US)
  • EP3444815A1 patent drawingFigure 1
  • EP3444815A1 patent drawingFigure 2
  • EP3444815A1 patent drawingFigure 3

AI summary

A method performed by a computing device for matrix downmixing an audio signal having N channels is provided, the method comprising selecting which of the N channels are surviving channels and which are non-surviving channels such that the surviving channels total M channels, where N and M are non-zero positive integers and N is greater than M; downmixing each of the non-surviving channels onto multiplets of the surviving channels using the computing device and multiplet pan laws to obtain panning weights, downmixing further comprising: downmixing some non-surviving channels onto surviving channel doublets using a doublet pan law; downmixing some non-surviving channels onto surviving channel triplets using a triplet pan law; downmixing some non-surviving channels onto surviving channel quadruplets using a quadruplet pan law; and encoding and multiplexing the surviving channel doublets, triplets, and quadruplets into a bitstream having M channels and transmitting the bitstream for rendering in a playback environment.