Developing a Dynamic Cluster Quantization based Lossless Audio Compression (DCQLAC)
- PDF / 3,328,040 Bytes
- 24 Pages / 439.642 x 666.49 pts Page_size
- 68 Downloads / 257 Views
Developing a Dynamic Cluster Quantization based Lossless Audio Compression (DCQLAC) Uttam Kr. Mondal1
· Asish Debnath2
Received: 13 September 2019 / Revised: 29 July 2020 / Accepted: 16 September 2020 / © Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract In this paper, an approach has been made to produce a compressed audio without losing any information. The proposed scheme is fabricated with the help of dynamic cluster quantization followed by Burrows Wheeler Transform (BWT) and Huffman coding. The encoding algorithm has been designed in two phases, i.e., dynamic cluster selection (of sampled audio) followed by dynamic bit selection for determining quantization level of individual cluster. Quantization level of each cluster is selected dynamically based on mean square quantization error (MSQE). Bit stream is further compressed by applying Burrows Wheeler Transform (BWT) and Huffman code respectively. Experimental results are supported with current state-of-the-art in audio quality analysis (like statistical parameters (compression ratio, space savings, SNR, PSNR) along with other parameters (encoding time, decoding time, Mean Opinion Score (MOS) and entropy) and compared with other existing techniques. Keywords Quantization · Lossless audio compression · Sampling · Entropy · Mean Square Quantization Error (MSQE) · Burrows Wheeler Transform (BWT) · Huffman encoding
1 Introduction Audio compression is a process for reducing the size of an audio file [28]. It is required for efficiently utilizing storage and transmission perspective [22]. Audio compression can be categorized to lossless and lossy. In lossless compression, exactly identical original audio can be restored from compressed representation. It reduces a file’s size with no loss of data quality. Lossy audio compression on the other hand sacrifices audio quality by omitting less important parts of audio data and achieves higher compression rates. Since that lossless Uttam Kr. Mondal
uttam ku [email protected] Asish Debnath [email protected] 1
Department of Computer Science, Vidyasagar University, Midnapore 721102 WB, India
2
Tata Consultancy Services Ltd. Newtown, Kolkata WB, India
Multimedia Tools and Applications
compression does not discard any audio data, it is used where any deviation from the original data is undesirable. It frequently used in different applications like social media, defence, biomedical, forensic, etc [28]. In this paper, audio compression is designed in three steps. First, sampling process is applied on the audio signal and the sampled data is produced considering the Nyquist criteria. Generated sampled values are dynamically distributed among clusters. This dynamic selection is performed using predefined rules. Varying size clusters are generated as an outcome of the uneven distribution. In the second step, entities in each cluster are processed and quantized sequentially. For each cluster, quantization bits are chosen dynamically. The sampled values of a cluster are quantized initially with a le
Data Loading...
