Optimisation of variable-length code for data compression of memoryless Laplacian source

Abstract

In this paper we present the efficient technique for compression and coding of memoryless Laplacian sources, which uses variable-length code (V LC). That technique is based on combination of two companding quantizers in the first case, and three companding quantizers in the second case. These quantizers have disjoint support regions, different number of representation levels and different compressor functions. The closed-form expressions are obtained for the distortion, average bit rate and signal to quantization noise ratio (SQN R). Presented numerical results point out the effects of rate-distortion (R-D) optimization on the system performances. Since our model assumes the general case of Laplacian distribution, it has a wide applications like coding of speech and images. It is shown that the difference of SQN R of our model and classical

2-Step Scalar Deadzone Quantization for Bitplane Image Coding

“…We note that the proposed 2SDQ scheme may also be seen as a compander [33]- [35] that employs the piecewise functions expressed in (8) and (9) as the compression and expanding functions, respectively. Appendix A expands this point of view further.…”

Design of forward adaptive hybrid quantiser with Golomb–Rice code for compression of Gaussian source

“…Thus, more sophisticated speech compression has become an essential component in telecommunications. This has prompted our extensive research in the area of speech compression during the past years [14][15][16][17][18].…”