Audio Decoding Frequency Regeneration to Reduce Low-Bitrate Artifacts

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current audio codecs face challenges in maintaining perceptual quality at low bitrates due to limitations in bandwidth extension techniques, which result in artifacts like dissonance, beating, and increased computational complexity, especially in mobile devices.

Innovation Solution

A signal-adaptive approach that performs frequency regeneration by analyzing the decoded core signal to detect and attenuate tonal components near transition points, applying cross-over filtering to reduce ringing artifacts, and modifying frequency borders to minimize artifacts during the reconstruction process.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If bandwidth extension techniques are used to maintain audio quality at low bitrates, then audio quality is improved, but artifacts like dissonance and beating occur

Engineering Contradiction:
Improveaudio qualityVSAvoiddissonance and beating artifacts
Core Design Contradiction:
Manufacturing precisionVSObject-affected harmful factors

Solution Approach 1:

The patent applies cross-over filtering to attenuate tonal components near transition points between low-frequency and high-frequency spectral regions. This converts potentially harmful tonal components that cause dissonance and beating into beneficial attenuated components, eliminating artifacts while preserving audio quality at low bitrates

Inventive Principle:
Principle #22Blessing in disguise (Convert harm into benefit)

Solution Approach 2:

The patent applies localized attenuation only to tonal components near transition points (cross-over frequencies) rather than uniformly across the entire spectrum. This selective approach maintains audio quality in critical regions while eliminating dissonance and beating artifacts only where they occur, preserving overall perceptual quality

Inventive Principle:
Principle #3Local quality

2Adaptability or versatility

If spectral patching is used to reconstruct high-frequency regions, then audio bandwidth is extended, but computational complexity increases

Engineering Contradiction:
Improveaudio bandwidth extensionVSAvoidcomputational complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent extracts and removes tonal components from the spectral patches before they are synthesized in the high-frequency region. By taking out the problematic tonal components that would require complex processing to manage, the system simplifies the overall computational complexity while maintaining bandwidth extension capability

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent performs preliminary attenuation of tonal components near transition points before the spectral patching and synthesis process. This preliminary action prevents the formation of artifacts that would otherwise require complex post-processing, thereby reducing overall computational complexity

Inventive Principle:
Principle #10Preliminary action

3Productivity

If frequency regeneration is performed without signal analysis, then processing speed is improved, but artifacts like ringing occur

Engineering Contradiction:
Improveprocessing speedVSAvoidringing artifacts
Core Design Contradiction:
ProductivityVSObject-affected harmful factors

Solution Approach 1:

The patent performs preliminary analysis of the decoded core signal to identify tonal components near transition points before frequency regeneration. This preliminary action allows the system to attenuate problematic components in advance, preventing ringing artifacts without requiring complex post-processing, thus maintaining processing speed

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3723091B1Apparatus, method and computer program for decoding an encoded audio signal
Publication Date: 2024.09.11 FRAUNHOFER GESELLSCHAFT ZUR FORDERUNG DER ANGEWANDTEN FORSCHUNG EV
  • EP3723091B1 patent drawingFigure 1A~1B
  • EP3723091B1 patent drawingFigure 2A
  • EP3723091B1 patent drawingFigure 2B

AI summary

Apparatus for decoding an encoded audio signal comprising an encoded core signal and parametric data, comprising: a core decoder (600) for decoding the encoded core signal to obtain a decoded core signal; an analyzer (602) for analyzing the decoded core signal before or after performing a frequency regeneration operation to provide an analysis result (603); and a frequency regenerator (604) for regenerating spectral portions not included in the decoded core signal using a spectral portion of the decoded core signal, the parametric data (605), and the analysis result (603).