Lossless audio codec based on CNN, weighted tree and arithmetic encoding (LACCWA)

In this paper, an integrated lossless audio codec with improved compression ratio has been proposed using traditional methods as well as convolutional neural network (CNN) architecture without losing any audio information. The method applies adaptive arithmetic encoding followed by binary weighted t...

Full description

Saved in:
Bibliographic Details
Published in:Multimedia tools and applications 2024-05, Vol.83 (16), p.48737-48759
Main Authors: Debnath, Asish, Mondal, Uttam Kr
Format: Article
Language:English
Subjects:
Adaptive sampling
Arithmetic
Artificial neural networks
Audio data
Codec
Codes
Coding
Compression ratio
Computer Communication Networks
Computer Science
Data integrity
Data Structures and Information Theory
Methods
Multimedia
Multimedia Information Systems
Neural networks
Special Purpose and Application-Based Systems
Citations: Items that this one cites
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In this paper, an integrated lossless audio codec with improved compression ratio has been proposed using traditional methods as well as convolutional neural network (CNN) architecture without losing any audio information. The method applies adaptive arithmetic encoding followed by binary weighted tree based transform and convolutional neural network (CNN) respectively. The first level of compression uses adaptive arithmetic encoding to process audio samples block by block, reducing the size of the data to the number of blocks. Subsequently, each of the arithmetic encoded values for every block is encoded by the binary weighted tree based encoding. Estimating a dynamic weighted path, it transforms the arithmetic-encoded data into an equivalent binary pattern. The binary stream is further compressed with latent space representations using a proposed CNN architecture. The analysis of the simulation results are performed using various statistical and robustness characteristics and comparison with other existing methods.
ISSN:1573-7721
1380-7501
1573-7721
DOI:10.1007/s11042-023-17393-4