Adaptive Transition Frequency for Audio Spectrum Recovery
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing audio coding methods face challenges in efficiently filling spectral holes across the entire frequency spectrum, leading to annoying artefacts in decoded audio signals, especially when transitioning between different audio content types like clean speech and full-band music.
Innovation Solution
Adaptive determination of a transition frequency based on the audio signal's spectral content to differentiate between noise filling and bandwidth extension, allowing for efficient suppression of perceptual artefacts and delivery of high-quality, full-bandwidth audio signals.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If a fixed transition frequency is used between low band and high band in SBR, then the decoding process is simplified, but the quality of reconstructed high frequencies deteriorates when the audio content varies (e.g., clean speech vs. full-band music)
Solution Approach 1:
The transition frequency is made dynamic and adaptive rather than fixed. The decoder determines the transition frequency adaptively based on the decoded low band signal characteristics, allowing the system to optimize reconstruction quality for different audio content types while maintaining a relatively simple decoding architecture.
Solution Approach 2:
The transition frequency parameter is changed adaptively based on the audio signal characteristics. By analyzing the decoded low band signal and adjusting the transition frequency accordingly, the system achieves better reconstruction quality across varying audio content without requiring complex pre-processing or fixed configurations.
2Device complexity
If spectral holes are filled using a single method across the entire spectrum, then the processing is simplified, but annoying artefacts persist in the decoded audio signal
Solution Approach 1:
The frequency spectrum is segmented into different regions (below and above the transition frequency) with different spectral hole filling methods applied to each. This segmentation allows noise filling to be used where appropriate (reducing artefacts) while avoiding its application in regions where it would be ineffective or harmful, thus reducing overall artefact levels.
Solution Approach 2:
Different spectral hole filling strategies are applied to different frequency regions based on local signal characteristics. Below the transition frequency, noise filling is applied to fill spectral holes, while above the transition frequency, other methods are used. This local adaptation of processing quality reduces artefacts without requiring complex global processing.
Data Source
AI summary
A method for spectrum recovery in spectral decoding of an audio signal, comprises obtaining of an initial set of spectral coefficients representing the audio signal, and determining a transition frequency. The transition frequency is adapted to a spectral content of the audio signal. Spectral holes in the initial set of spectral coefficients below the transition frequency are noise filled and the initial set of spectral coefficients are bandwidth extended above the transition frequency. Decoders and encoders being arranged for performing part of or the entire method are also illustrated.