Asymmetric Windowing for Multichannel Audio Decoding

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Multichannel audio coding systems face challenges in reducing decoding delay and improving sound quality, particularly due to mismatches in encoder and decoder windowing schemes, which can lead to audio artifacts.

Innovation Solution

The implementation of asymmetric windowing schemes in both the encoder and decoder, with unevenly weighted smoothing to offset the effects of unequal window overlaps, allows for reduced decoding delay and improved sound quality by interpolating stereo parameters between frames.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Object-affected harmful factors

If look ahead portions are used in windowing schemes, then audio artifacts are reduced, but decoding delay increases

Engineering Contradiction:
Improveaudio artifactsVSAvoiddecoding delay
Core Design Contradiction:
Object-affected harmful factorsVSLoss of time

Solution Approach 1:

The patent applies asymmetric windowing schemes where the encoder and decoder use different window configurations. The encoder uses a first windowing scheme with specific look-ahead portions, while the decoder uses a second windowing scheme that compensates for the delay introduced by the encoder's look-ahead, thereby reducing overall decoding delay while maintaining audio quality.

Inventive Principle:
Principle #4Asymmetry

Solution Approach 2:

The patent modifies windowing parameters (window length, overlap portions, look-ahead portions) to optimize the balance between artifact reduction and delay. By adjusting these parameters in the asymmetric windowing schemes, the system achieves reduced audio artifacts without proportionally increasing decoding delay.

Inventive Principle:
Principle #35Parameter changes

2Loss of time

If encoder and decoder use different windowing schemes, then decoding delay is reduced, but audio artifacts increase

Engineering Contradiction:
Improvedecoding delayVSAvoidaudio artifacts
Core Design Contradiction:
Loss of timeVSObject-affected harmful factors

Solution Approach 1:

The patent incorporates feedback mechanisms where the decoder's second windowing scheme is designed to compensate for the encoder's first windowing scheme. This feedback loop ensures that the asymmetric windowing does not introduce significant audio artifacts, as the decoder adjusts its processing to match the encoder's output characteristics.

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The patent performs preliminary planning of the asymmetric windowing schemes during system design. The encoder and decoder windowing parameters are pre-coordinated to ensure compatibility, allowing the system to achieve reduced delay without artifacts through carefully designed complementary windowing strategies.

Inventive Principle:
Principle #10Preliminary action

3Object-affected harmful factors

If stereo parameters are interpolated between frames, then sound quality is improved, but processing complexity increases

Engineering Contradiction:
Improvesound qualityVSAvoidprocessing complexity
Core Design Contradiction:
Object-affected harmful factorsVSDevice complexity

Solution Approach 1:

The patent applies partial interpolation of stereo parameters between frames rather than full interpolation. By selectively interpolating only the necessary parameters and using uneven weighting, the system improves sound quality while avoiding the full processing complexity of complete parameter interpolation.

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentEP3577647B1Multi channel decoding
Publication Date: 2020.12.16 QUALCOMM INC
  • EP3577647B1 patent drawingFigure 1
  • EP3577647B1 patent drawingFigure 2
  • EP3577647B1 patent drawingFigure 3

AI summary

A method includes generating a windowed time-domain mid channel by applying two first asymmetric windows to a first frame of a time-domain mid channel and applying two second asymmetric windows to a second frame of the time-domain mid channel. The method includes transforming the windowed time-domain mid channel to a transform domain to generate sets of transform-domain mid channel data including first transform-domain mid channel data corresponding to a first mid channel window of the first frame and second transform-domain mid channel data corresponding to a second mid channel window of the first frame. The method includes performing an up-mix operation using the sets of transform-domain mid channel data, stereo parameters from the bit stream, and an interpolated parameter determined using an unevenly weighted interpolation between a first stereo parameter value associated with the first frame and a second stereo parameter value associated with the second frame.