Spectral Band Replication for High-Frequency Audio Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current audio coding technologies face challenges in effectively extending the bandwidth of high-frequency bands, particularly in efficiently encoding and decoding signals above 6.4 kHz, due to limitations in bit allocation and hearing resolution of human perception.
Innovation Solution
A coding and decoding apparatus that generates a base signal from a low-frequency signal to extend the high-frequency bandwidth through spectral band replication, using techniques like core coding, frequency transformation, and bandwidth extension coding, with energy quantization and vector quantization methods to optimize bit allocation across frequency bands.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a large number of bits are assigned to the low-frequency band to increase coding efficiency, then the coding efficiency is improved, but the bandwidth extension capability deteriorates
Solution Approach 1:
The patent segments the audio signal into low-frequency band (0-6.4 kHz) and high-frequency band (6.4-16 kHz), applying different coding strategies to each segment. The low-frequency band receives detailed coding with more bits, while the high-frequency band uses bandwidth extension techniques that require fewer bits, thus resolving the contradiction between coding efficiency and bandwidth extension capability
Solution Approach 2:
The patent applies local quality by providing different coding precision to different frequency bands. The low-frequency band is coded with higher precision (more bits) where human hearing is more sensitive, while the high-frequency band uses lower precision (fewer bits) with spectral band replication, optimizing the overall coding efficiency without sacrificing perceived audio quality
2Quantity of substance
If spectral band replication is used to extend high-frequency band, then the bandwidth extension capability is improved, but the audio quality for complex signals deteriorates
Solution Approach 1:
The patent dynamically adapts the bandwidth extension process by detecting signal characteristics (tonal vs. non-tonal) and adjusting the synthesis parameters accordingly. For tonal signals, harmonic generation is enhanced, while for non-tonal signals, noise-shaped synthesis is used, thus maintaining audio quality across different signal types while achieving bandwidth extension
Solution Approach 2:
The patent changes synthesis parameters based on the input signal characteristics. The gain factor, spectral shape, and excitation type are adjusted dynamically to match the original high-frequency signal properties, ensuring that the synthesized bandwidth extension maintains audio quality fidelity for both speech and music signals
3Productivity
If hearing characteristics are used to allocate bits to high-frequency band, then the coding efficiency is improved, but the resolution and audio fidelity deteriorate
Solution Approach 1:
The patent introduces an intermediary process of spectral band replication that generates high-frequency content from low-frequency information. Instead of directly coding high-frequency signals with limited bits, the system uses the low-frequency signal as an intermediary to synthesize the high-frequency band, preserving perceived audio quality while improving coding efficiency
Data Source
Figure 1~2B
Figure 2C~2D
Figure 3~4
AI summary
Disclosed are an apparatus and a method for encoding/decoding for high-frequency bandwidth extension. The encoding apparatus may downsample an input signal, perform core-encoding of the downsampled input signal, perform frequency conversion of the input signal, and perform bandwidth-extension encoding using a basic signal of the input signal of a frequency domain.