IntMDCT-based lossless audio encoding method and decoding method

Abstract

The invention provides an IntMDCT-based lossless audio encoding method and decoding method. The encoding method includes the following steps that: (1) audio time domain signals x (k) are framed, and a single frame of time domain signals x (k) can be obtained, wherein the k=0,..., N-1, wherein the N is frame length; (2) IntMDCT transformation is performed on the time domain signals, so that frequency domain signals X (m) can be obtained, wherein the m=0,..., N-1; and (3) Rice encoding is performed on the frequency domain signals, so that code streams can be obtained. The decoding method is an inverse process of the encoding method and includes the following steps that: firstly, Rice decoding is performed on an inputted code stream file, so that the frequency domain signals X (m) can be obtained, wherein the m=0,..., N-1; and secondly, IntMDCT inverse transformation is performed, so that the time domain signals x (k) can be restored, wherein the k=0,..., N-1. With the methods of the invention adopted, encoding parameters can be obtained adaptively, and the value of compression ratio is lower than that of block adaptive Rice, and the methods of the invention have a wider applicable range.

Description

technical field [0001] The invention relates to the field of lossless audio encoding and decoding, in particular to an IntMDCT-based lossless audio encoding method and decoding method. Background technique [0002] Lossless audio coding refers to coding audio signals to reduce the amount of data and obtain a low bit rate. Regardless of any information loss, the receiving end can decode it without error. [0003] The lossless audio coding framework commonly used in the world mainly includes two steps: prediction / transformation and entropy coding. Prediction / transformation essentially removes the temporal and spatial correlation of audio signals; while entropy coding removes the statistical correlation between audio signals. [0004] At present, the two mainstream decorrelation techniques mainly include prediction and transformation. Among them, prediction is the more commonly used method. Predictive techniques have certain limitations for audio coding. Many audio clips pe...

AI Technical Summary

Problems solved by technology

Rice coding is indeed an optimal code when the Laplacian distribution and the geometric probability parameter ρ are known, but in fact, Rice coding still has problems: the probability parameter ρ must be known, and the coding of the optimal code The parameter s can be determined, but in fact, it is not possible to fully estimate the entire information source to obtain the most suitable ρ, and then obtain the optimal Rice coding parameter s, so the determination of ρ directly affects the effect of Rice coding

[0007] (2) It is difficult to choose an appropriate block size. If the block size is too large, the statistics will change drastically within the block. If it is too small, the decoder must be informed which parameter is used for encoding, making the extra overhead unaffordable

[0009] It can be seen that the lossless audio encoding and decoding methods commonly used in the world have defects, so it is necessary to further study them

Images