Audio frequency lossless compression coding and decoding method and system based on basis pursuit

Abstract

The invention discloses an audio frequency lossless compression coding and decoding method and system based on basis pursuit. The coding method comprises the following steps: 1) framing an input audio frequency signal, and inputting each frame of signal into a sparse coding module; 2) carrying out sparse transformation to each frame of signal by an overcomplete dictionary basis function via the sparse coding module, and selecting the most sparse transformation mode from a sparse transformation result by using a basis pursuit method; 3) getting a predictor parameter and an updater parameter required in integer transformation via the sparse coding module by using a base vector combination corresponding to the most sparse transformation mode, carrying out sparse integer transformation to the frame of signal to obtain a sparse coefficient and transfer the sparse coefficient to an entropy coding module for coding, and sending side information generated by the sparse transformation of the frame of signal to a bit stream formation module to be coded; and 4) integrating the coding of the frame of signal together to be output as compressed coding by a code stream integration module. According to the invention, the coding and decoding efficiency of lossless coding is greatly improved.

Description

technical field [0001] The invention belongs to the field of information source encoding and decoding, and in particular relates to an audio lossless compression encoding and decoding method and system. Background technique [0002] Audio lossless codec technology is a technology that can compress and store digital audio signals and restore them without distortion. In recent years, the continuous development of technologies such as large-scale storage devices, high-speed Internet connections, and broadband wireless networks has provided a huge stage for the promotion of audio lossless codec technology, high-quality Internet audio downloads, and symphony coding with a large frequency dynamic range. , the editing and archiving of classic audio files, etc. are constantly emerging. With the improvement of people's audio quality requirements, traditional lossy audio codec methods can no longer meet such needs. In order to solve this type of problem, it is necessary to apply to L...

AI Technical Summary

Problems solved by technology

Linear predictive coding has a good predictive ability for quasi-stationary signals such as speech, but for music signals with richer high-frequency components, greater dynamic range and greater rate of change, its predictive ability is greatly limited, although it can be achieved through Increase the order of the predictor to overcome this problem to a certain extent, but at the same time, in order to ensure lossless decoding, more predictor coefficient information must be introduced, which reduces the coding efficiency

Images