Multichannel Stereo Filling for Quantization Spectral Holes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing audio signal coding technologies struggle with quantization-induced spectral holes, particularly in multichannel coding scenarios, as they fail to optimally exploit time-variant inter-channel dependencies and result in noise artifacts for tonal signals.
Innovation Solution
A decoding apparatus and method that selects pairs of decoded channels based on multichannel parameters, identifies frequency bands with zero spectral lines, and fills these bands with noise generated from a mixing channel using previous audio output channels, adapting to time-variant dependencies.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If quantization is applied to compress audio signals, then data rate is reduced, but spectral holes are created causing degradation in audio quality
Solution Approach 1:
The patent converts the harmful spectral holes created by quantization into beneficial filled spectral regions by generating noise from previous frame's spectral information. This noise filling technique transforms the degradation caused by quantization into an opportunity to maintain audio quality while preserving compression efficiency.
Solution Approach 2:
The patent performs preliminary action by storing and utilizing spectral information from previous frames before the current quantization process. This pre-computed spectral data is then used to fill spectral holes, anticipating and preventing quality degradation before it occurs.
2Manufacturing precision
If noise filling is applied to fill spectral holes, then audio quality is improved, but noise artifacts are introduced for tonal signals
Solution Approach 1:
The patent applies local quality by selectively filling only those spectral holes that meet specific criteria (e.g., adjacent filled regions, below threshold levels) while leaving other regions untouched. This localized approach ensures that noise filling is applied only where beneficial, avoiding artifacts in tonal signals.
Solution Approach 2:
The patent changes parameters by dynamically adjusting the noise filling threshold and selection criteria based on signal characteristics. By modifying these parameters adaptively, the system optimizes the balance between filling spectral holes and avoiding noise artifacts in different signal contexts.
3Device complexity
If fixed channel pairing is used for joint stereo coding, then coding complexity is reduced, but time-variant inter-channel dependencies cannot be exploited
Solution Approach 1:
The patent applies dynamics by transitioning from fixed channel pairing to dynamic channel selection based on inter-channel correlation analysis. The system adaptively identifies and pairs channels with highest correlation for each frame, allowing the coding scheme to respond to time-variant dependencies while maintaining manageable complexity through systematic selection criteria.
4Manufacturing precision
If spectral information from all previous channels is used for noise generation, then filling effectiveness is improved, but computational complexity increases
Solution Approach 1:
The patent applies segmentation by dividing the multichannel spectral information into manageable pairs or groups for processing. Instead of treating all channels simultaneously, the system processes channels in segmented units, reducing computational complexity while maintaining filling effectiveness through systematic combination of segmented results.
Data Source
AI summary
An apparatus for decoding an encoded multichannel signal of a current frame to obtain three or more current audio output channels is provided. A multichannel processor is adapted to select two decoded channels from three or more decoded channels depending on first multichannel parameters. Moreover, the multichannel processor is adapted to generate a first group of two or more processed channels based on the selected channels. A noise filling module is adapted to identify for at least one of the selected channels, one or more frequency bands, within which all spectral lines are quantized to zero, and to generate a mixing channel using, depending on side information, a proper subset of three or more previous audio output channels that have been decoded, and to fill the spectral lines of frequency bands, within which all spectral lines are quantized to zero, with noise generated using spectral lines of the mixing channel.