Adaptive Spectral Tile Selection for Low-Bitrate Audio Coding

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current audio codecs face limitations in maintaining audio quality at low bitrates due to restricted bandwidth extension techniques, which fail to accurately align tonal harmonics and introduce artifacts, especially in non-tonal signals, and require complex domain transformations increasing computational complexity and memory requirements.

Innovation Solution

The implementation of an adaptive frequency tile filling scheme that identifies the best matching source region for a target region, transmitting matching information to the decoder for spectral reconstruction, allowing for signal-dependent spectral tile selection and whitening to enhance audio quality without the need for additional domain transformations.

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 computational complexity and memory requirements increase due to complex domain transformations

Engineering Contradiction:
Improveaudio qualityVSAvoidcomputational complexity
Core Design Contradiction:
Manufacturing precisionVSDevice complexity

Solution Approach 1:

The patent extracts only the essential spectral information needed for bandwidth extension by identifying and transmitting matching information between source and target spectral regions. Instead of performing complex domain transformations, the system extracts spectral similarity metrics and uses these to guide the bandwidth extension process, significantly reducing computational complexity while maintaining audio quality.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent uses spectral copying by identifying matching source spectral regions and copying their characteristics to target high-frequency regions. This copying approach avoids complex transformations by directly utilizing spectral similarity, thereby reducing computational requirements while preserving perceptual audio quality through accurate spectral replication.

Inventive Principle:
Principle #26Copying

2Manufacturing precision

If bandwidth extension techniques are used to maintain audio quality at low bitrates, then audio quality is improved, but device memory requirements increase

Engineering Contradiction:
Improveaudio qualityVSAvoidmemory requirements
Core Design Contradiction:
Manufacturing precisionVSVolume of stationary object

Solution Approach 1:

The patent extracts only the critical matching information between spectral regions rather than storing complete spectral data. By identifying and transmitting only the essential similarity metrics and matching parameters, the system significantly reduces memory requirements while maintaining the capability to perform accurate bandwidth extension for high audio quality.

Inventive Principle:
Principle #2Taking out (Extraction)

3Manufacturing precision

If spectral patching is used to reconstruct high frequency regions, then bandwidth extension is achieved, but artifacts are introduced especially in non-tonal signals

Engineering Contradiction:
Improvebandwidth extensionVSAvoidartifacts
Core Design Contradiction:
Manufacturing precisionVSObject-generated harmful factors

Solution Approach 1:

The patent implements dynamic spectral tile selection that adapts to the local spectral characteristics of the signal. By dynamically identifying matching source regions for each target region based on spectral similarity metrics, the system adjusts the patching process to suit tonal and non-tonal signals differently, thereby reducing artifacts while achieving effective bandwidth extension.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The patent changes the parameters used for spectral matching by introducing spectral similarity metrics and adaptive matching criteria. This allows the system to differentiate between tonal and non-tonal regions and apply appropriate matching strategies, reducing artifacts in non-tonal signals while maintaining bandwidth extension effectiveness.

Inventive Principle:
Principle #35Parameter changes

4Manufacturing precision

If high resolution encoding is applied to tonal portions, then audio quality is improved, but bitrate consumption increases

Engineering Contradiction:
Improveaudio qualityVSAvoidbitrate
Core Design Contradiction:
Manufacturing precisionVSLoss of information

Solution Approach 1:

The patent applies local quality by encoding different spectral regions with different resolutions based on their perceptual importance. Tonal portions that are perceptually critical are encoded with high resolution, while non-tonal or less important regions use lower resolution encoding. This selective approach improves overall audio quality while controlling bitrate consumption.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The patent applies partial action by selectively applying high-resolution encoding only to the most perceptually important spectral portions rather than encoding the entire spectrum at high resolution. By identifying and prioritizing tonal regions for high-resolution encoding, the system achieves improved audio quality with controlled bitrate usage.

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentEP3025343B1Apparatus and method for decoding and encoding an audio signal using adaptive spectral tile selection
Publication Date: 2018.02.14 FRAUNHOFER GESELLSCHAFT ZUR FORDERUNG DER ANGEWANDTEN FORSCHUNG EV
  • EP3025343B1 patent drawingFigure 1A~1B
  • EP3025343B1 patent drawingFigure 2A
  • EP3025343B1 patent drawingFigure 2B

AI summary

An apparatus for decoding an encoded signal, comprises: an audio decoder (1102) for decoding an encoded representation of a first set of first spectral portions to obtain a decoded first set of first spectral portions (1101); a parametric decoder (1104) for decoding an encoded parametric representation of a second set of second spectral portions to obtain a decoded representation of the parametric representation (1103), wherein the parametric information includes, for each target frequency tile, a source region identification as a matching information; and a frequency regenerator (1106) for regenerating a target frequency tile using a source region from the first set of first spectral portions (1101) identified by the matching information.