Minimally Guided Bandwidth Extension for Low Bitrate Audio
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current audio coding systems face challenges in achieving acceptable perceptual quality at very low bitrates, particularly for sounds with low correlation between high-frequency (HF) and low-frequency (LF) content, such as fricatives, due to limitations in bandwidth extension (BWE) techniques, which are either computationally complex or require excessive side information.
Innovation Solution
A minimally guided bandwidth extension (BWE) approach that adapts to the instantaneous characteristics of the input signal by using small amounts of side information to selectively enhance the spectral envelope, distinguishing between different speech sounds and controlling the statistical model to improve perceptual quality, particularly for difficult-to-code sounds like sibilants and fricatives.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If guided bandwidth extension is used to improve perceptual quality, then audio quality is improved, but side information rate increases
Solution Approach 1:
The patent extracts only the most critical spectral envelope parameters needed for bandwidth extension, transmitting a minimized set of side information that captures essential HF characteristics without the overhead of full guided BWE. This selective extraction resolves the contradiction by taking out only what is necessary for quality improvement.
Solution Approach 2:
The patent applies different processing strategies to different spectral regions, focusing side information transmission on the critical spectral envelope parameters that most impact perceptual quality while omitting less critical details. This local quality approach ensures quality improvement where it matters most while minimizing overall side information rate.
2Loss of information
If blind bandwidth extension is used to reduce side information rate, then side information rate is reduced, but perceptual quality deteriorates
Solution Approach 1:
The patent transmits a partial amount of side information - enough to guide the statistical model for spectral envelope estimation but not enough for full guided BWE. This partial action provides just sufficient guidance to avoid blind BWE quality degradation while keeping side information rate low.
Solution Approach 2:
The patent changes the parameter representation from full spectral details to a compact statistical model description. By transforming the side information into statistical parameters that capture essential HF characteristics, the system achieves better perceptual quality than blind BWE while using minimal side information rate.
3Manufacturing precision
If complex statistical models are used to handle low-correlation sounds, then perceptual quality is improved, but computational complexity increases
Solution Approach 1:
The patent employs a dynamic statistical model that adapts its parameters based on the instantaneous characteristics of the input signal. The model dynamically adjusts to handle different speech sounds (vowels, fricatives, plosives) with varying HF-LF correlations, improving perceptual quality for challenging sounds while maintaining computational efficiency through adaptive rather than universally complex processing.
Data Source
Figure 1~2
Figure 3~4
Figure 5
AI summary
A decoder for generating a frequency enhanced audio signal (120), comprises: a feature extractor (104) for extracting a feature from a core signal (100); a side information extractor (110) for extracting a selection side information associated with the core signal; a parameter generator (108) for generating a parametric representation for estimating a spectral range of the frequency enhanced audio signal (120) not defined by the core signal (100), wherein the parameter generator (108) is configured to provide a number of parametric representation alternatives (702, 704, 706, 708) in response to the feature (112), and wherein the parameter generator (108) is configured to select one of the parametric representation alternatives as the parametric representation in response to the selection side information (712 to 718); and a signal estimator (118) for estimating the frequency enhanced audio signal (120) using the parametric representation selected.