Audio Compression Using Low-Band Matching for High Frequencies
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current audio coding techniques, such as spectral-band-replication (SBR), face challenges in accurately representing the high frequency region of audio signals, leading to dissimilarities between original and reconstructed high frequencies, and require additional noise and sinusoids to maintain quality at low bit rates.
Innovation Solution
The method involves dividing the audio signal into low and high frequency bands, determining signal sections within the low frequency band that best match high frequency sub-bands, and generating parameters to model these sections, allowing for accurate reconstruction of high frequency sub-bands by transposing and scaling low frequency signal sections, while reducing the number of quantization bits through envelope normalization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If spectral-band-replication (SBR) is used to code the high frequency region, then the bit rate is reduced, but the similarity between the original high frequencies and their reconstruction deteriorates
Solution Approach 1:
The patent copies signal sections from the low frequency band to represent the high frequency sub-bands. Instead of blindly transposing the entire low frequency band, it identifies and copies specific signal sections that best match the characteristics of each high frequency sub-band, thereby maintaining higher similarity while reducing bit rate
Solution Approach 2:
The patent applies local quality by selecting different signal sections from the low frequency band for different high frequency sub-bands based on their specific characteristics. Each high frequency sub-band is represented by the most suitable local section from the low frequency band, rather than using a uniform approach for all sub-bands
2Measurement precision
If more quantization bits are allocated to high frequency sub-bands, then the quality of high frequency representation is improved, but the overall compression ratio deteriorates
Solution Approach 1:
The patent changes the parameter representation by using indices to point to signal sections in the low frequency band instead of directly quantizing and transmitting high frequency data. This parameter transformation allows high quality representation with fewer bits, as the indices require minimal quantization bits compared to direct high frequency signal transmission
Data Source
AI summary
For audio encoding and decoding, in order to enhance coded audio signals, the audio signal is divided into at least a low frequency band and a high frequency band, the high frequency band is divided into at least two high frequency sub-band signals, and parameters are generated that refer at least to the low frequency band signal sections which match best with high-frequency sub-band signals.