Audio Compression Coding Using Sign-Separated Residual Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional prediction methods are not applicable to all types of input signals, particularly those with larger dynamic ranges and white noise-like spectra, leading to poor compression efficiency in voice and audio signals.
Innovation Solution
The method involves extracting sign information to obtain an absolute value signal, determining a prediction coefficient based on the signal characteristics, and entropy-coding the residual signal along with the sign information and coding parameters to generate a coding code stream.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional prediction methods (LPC, LTP) are used for voice and audio coding, then compression efficiency is improved for signals with predictable characteristics, but compression gain is lost for signals with large dynamic ranges and white noise-like spectra
Solution Approach 1:
The patent applies dynamics by making the prediction method adaptable to different signal characteristics. The system dynamically selects between different coding approaches (conventional prediction vs. sign coding) based on the actual signal properties, allowing the coder to optimize for each specific signal type rather than using a fixed prediction method
Solution Approach 2:
The patent changes the fundamental parameter being coded by separating the sign information from the magnitude information. Instead of directly coding the prediction residual, the system codes the sign bit separately and applies prediction only to the absolute values, which fundamentally changes how the signal is processed and enables better performance for signals with large dynamic ranges
2Reliability
If lossless coding is used to ensure reconstructed signal consistency with original signal, then voice quality is ensured, but compression rate is lower
Solution Approach 1:
The patent segments the signal processing into distinct components: sign information extraction, absolute value processing, and residual coding. This segmentation allows each component to be optimized independently, achieving both high fidelity reconstruction and improved compression by treating sign and magnitude differently
3Device complexity
If conventional prediction coding is applied to signals with large dynamic ranges and white noise-like spectra, then coding complexity is maintained, but compression gain is severely reduced
Solution Approach 1:
The patent extracts the sign information from the signal and processes it separately from the magnitude information. This extraction allows the prediction mechanism to focus only on the absolute values, which have more predictable characteristics, while the sign information is handled through a simpler separate coding path, thereby maintaining low complexity while achieving compression gain for difficult signal types
Data Source
AI summary
The embodiments of the present invention relate to a compression coding and decoding method, a coder, a decoder and a coding device. The compression coding method includes: extracting sign information of an input signal to obtain an absolute value signal of the input signal; obtaining a residual signal of the absolute value signal by using a prediction coefficient, where the prediction coefficient is obtained by prediction and analysis that are performed according to a signal characteristic of the absolute value signal of the input signal; and multiplexing the residual signal, the sign information and a coding parameter to output a coding code stream, after the residual signal, the sign information and the coding parameter are respectively coded, so as to improve compression efficiency of a voice and audio signal.