Data compression: techniques and applications, hardware and software considerations (2nd ed.)
Data compression: techniques and applications, hardware and software considerations (2nd ed.)
Data Compression: The Complete Reference
Data Compression: The Complete Reference
A multi-resolution model of vector map data for rapid transmission over the Internet
Computers & Geosciences
Hi-index | 0.00 |
Large gridded data sets useful in the Earth sciences are routinely made available on the Internet. In most cases the data files are provided using 2-byte (16 bit) binary integer values; these are often compressed using general-purpose utilities such as gzip or bzip2 to reduce file sizes prior to transmission over the Internet. However, even files compressed in this way can be hundreds of megabytes, contributing to net congestion and excessive transmission times. This paper presents a command line tool (grdzip) that takes advantage of the structure of gridded data sets to achieve better net compression. The data are first reduced to double differences and packed using a variable width bit-packing algorithm before being written via the general-purpose compression routines in the bzip2 library. The process is fully reversible. Testing of grdzip on eight commonly used gridded data sets reveals that grdzip compresses better and faster than either gzip or bzip2 alone, typically resulting in compressed files that are 25% smaller than bzip2 output and less than 50% of gzip output. Depending on original file size, these savings can be considerable and significantly reduce both transmission time and data server disk space.