Audio Decoder Tool Selection for Packet-Loss Concealment

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing audio decoders face challenges in selecting the most suitable loss concealment method for different audio signal characteristics during packet loss, leading to noticeable artifacts in the reconstructed audio signal.

Innovation Solution

An audio decoder that assigns loss concealment tools based on spectral centroid and temporal predictability measures to choose the most appropriate method for recovering lost audio signals, using either pitch-based or tonal frequency domain techniques, ensuring minimal annoyance in the reconstructed audio.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If a single loss concealment method is used for all audio signals, then the device complexity is reduced, but the quality of reconstructed audio deteriorates due to noticeable artifacts

Engineering Contradiction:
Improveloss concealment mechanism complexityVSAvoidreconstructed audio quality
Core Design Contradiction:
Device complexityVSManufacturing precision

Solution Approach 1:

The patent implements dynamic selection of loss concealment methods based on real-time analysis of audio signal characteristics. The system adapts the concealment approach (pitch-based, frame repetition, or tonal component prediction) according to the detected signal type, transforming a static single-method system into a dynamic multi-method system that optimizes reconstructed audio quality while managing complexity through conditional logic

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system changes operational parameters by analyzing signal characteristics (spectral properties, temporal structure) and selecting different concealment methods accordingly. This parameter-based selection allows the system to adjust its behavior based on the specific audio content, improving reconstructed quality without requiring a permanently complex multi-method architecture

Inventive Principle:
Principle #35Parameter changes

2Manufacturing precision

If multiple loss concealment methods are implemented for different signal types, then the reconstructed audio quality improves, but the device complexity increases

Engineering Contradiction:
Improvereconstructed audio qualityVSAvoidloss concealment mechanism complexity
Core Design Contradiction:
Manufacturing precisionVSDevice complexity

Solution Approach 1:

The patent segments the loss concealment process into distinct specialized methods (pitch-based for speech, frame repetition for noise-like signals, tonal component prediction for polyphonic music). Each method is optimized for specific signal types, and the system divides the overall task by analyzing signal characteristics and routing to the appropriate segment, managing complexity through functional decomposition

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system introduces an intermediary signal analysis mechanism that evaluates audio characteristics and mediates between multiple concealment methods. This intermediary layer (analyzing spectral centroid, signal stationarity, tonal properties) selects the most suitable concealment approach, allowing multiple methods to coexist without requiring the system to manage all methods simultaneously in a complex manner

Inventive Principle:
Principle #24Intermediary (Mediator)

3Manufacturing precision

If pitch-based PLC techniques are used for speech signals, then the reconstructed speech quality improves, but the method becomes ineffective for non-periodic noise-like signals

Engineering Contradiction:
Improvespeech reconstruction qualityVSAvoidconcealment method applicability
Core Design Contradiction:
Manufacturing precisionVSAdaptability or versatility

Solution Approach 1:

The patent applies local quality by matching specific concealment methods to specific signal types: pitch-based methods for periodic speech signals, frame repetition for non-periodic noise-like signals, and tonal component prediction for polyphonic music. Each method has localized optimization for its target signal type, and the system routes signals to the appropriate method based on local signal characteristics, achieving both specialization and versatility

Inventive Principle:
Principle #3Local quality

4Manufacturing precision

If frame repetition with sign scrambling is used for noise-like signals, then the concealment effectiveness improves, but the method produces artifacts for periodic speech signals

Engineering Contradiction:
Improvenoise signal reconstruction qualityVSAvoidartifacts in periodic signals
Core Design Contradiction:
Manufacturing precisionVSObject-affected harmful factors

Solution Approach 1:

The system implements feedback by analyzing the characteristics of the signal before applying concealment and evaluating whether the chosen method produces acceptable results. The signal analysis (detecting periodicity, spectral properties) provides feedback that guides the selection of an appropriate concealment method, preventing the application of frame repetition to periodic signals where it would create harmful artifacts

Inventive Principle:
Principle #23Feedback

Data Source

PatentEP3707705B1Audio decoder supporting a set of different loss concealment tools
Publication Date: 2022.10.19 FRAUNHOFER GESELLSCHAFT ZUR FORDERUNG DER ANGEWANDTEN FORSCHUNG EV
  • EP3707705B1 patent drawingFigure 1
  • EP3707705B1 patent drawingFigure 2
  • EP3707705B1 patent drawingFigure 3

AI summary

An assignment of one of phase set of different loss concealment tools of an audio decoder to a portion of the audio signal to be decoded from a data stream, which portion is affected by loss, that is the selection out of the set of different loss concealment tools, may be made in a manner leading to a more pleasant loss concealment if the assignment/selection is done based on two measures: A first measure which is determined measures a spectral position of a spectral centroid of a spectrum of the audio signal and a second measure which is determined measures a temporal predictability of the audio signal. The assigned or selected loss concealment tool may then be used to recover the portion of the audio signal.