Audio Signal Encoding with Amplitude Sorting for Low-Bitrate Quality

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional audio coding methods face challenges in maintaining signal quality at low bit rates, particularly with lossy coding methods that introduce perceptible artifacts and require significant computing power, while lossless methods struggle with data reduction and efficient encoding.

Innovation Solution

A device and method for encoding and decoding audio signals that involves sorting the input signal by amplitude, followed by prediction and entropy coding, allowing for effective coding of sorted data and re-sorting rules, which can include permutation and inversion tables, to achieve efficient data representation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If lossy coding methods with psychoacoustic models are used, then data reduction is achieved, but signal quality deteriorates with perceptible coding artifacts at low bit rates

Engineering Contradiction:
Improvedata volumeVSAvoidsignal quality
Core Design Contradiction:
Quantity of substanceVSManufacturing precision

Solution Approach 1:

The patent applies preliminary sorting of spectral coefficients by amplitude before coding. This preliminary action reorganizes the data so that energy is concentrated in fewer coefficients, enabling more efficient encoding. By sorting beforehand, the system can achieve better data reduction while maintaining signal quality, as the most significant coefficients are encoded with higher precision while less significant ones can be coarsely quantized or discarded.

Inventive Principle:
Principle #10Preliminary action

2Loss of information

If conventional transformation-based coding is used, then frequency domain representation is achieved, but computing power consumption increases due to quantization and downsampling operations

Engineering Contradiction:
Improvefrequency domain representationVSAvoidcomputing power
Core Design Contradiction:
Loss of informationVSUse of energy by moving object

Solution Approach 1:

The patent extracts and encodes only the most significant spectral coefficients after sorting, rather than processing all coefficients equally. By taking out only the essential information (top N sorted coefficients), the system reduces computing power requirements for quantization and encoding operations while maintaining the frequency domain representation quality. This selective extraction avoids unnecessary processing of less significant coefficients.

Inventive Principle:
Principle #2Taking out (Extraction)

3Manufacturing precision

If lossless coding methods are used, then perfect reconstruction is achieved, but data reduction efficiency decreases compared to lossy methods

Engineering Contradiction:
Improvereconstruction accuracyVSAvoiddata reduction ratio
Core Design Contradiction:
Manufacturing precisionVSQuantity of substance

Solution Approach 1:

The patent applies different coding strategies to different portions of the sorted spectral coefficients. The most significant coefficients (at the beginning of the sorted sequence) are encoded with lossless precision to ensure perfect reconstruction of important signal components. Less significant coefficients (toward the end of the sequence) can be encoded with lossy compression or discarded entirely. This local differentiation of quality levels enables both perfect reconstruction of critical information and overall data reduction.

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP2054884B1Device for encoding and decoding
Publication Date: 2011.10.05 FRAUNHOFER GESELLSCHAFT ZUR FORDERUNG DER ANGEWANDTEN FORSCHUNG EV
  • EP2054884B1 patent drawingFigure 1A
  • EP2054884B1 patent drawingFigure 1B
  • EP2054884B1 patent drawingFigure 2A

AI summary

The invention relates to a device (100) for encoding a sequence of scanning values of an audio signal, wherein each scanning value within the sequence has an original position. The device (100) comprises a system (110) for sorting the scanning values according to size in order to obtain a sorted sequence of scanning values, wherein each scanning value has a sorting position within the sorted sequence. Furthermore, the device (100) comprises a system (120) for encoding the sorted scanning values and information about a relationship between the original and sorting positions of the scanning values.