The principle of data compression of text, graphic and moving image data is based on the elimination of redundant characters or a dynamic allocation of data bits depending on or frequency of a character.
Sophisticated compression algorithms allow data compression of arbitrary data streams, even if they are computer programs or random sequences. The aim of data compression is to reduce the size of files in order to achieve faster data throughput during data transmission and to require less storage space during archiving. In this context, this is also referred to as source coding.
Lossless and lossy compression
In data compression, a distinction is made between lossless compression and lossy compression. While in lossless compression the reconstructed signal is exactly the same as the original signal, in lossy compression it is no longer identical to the original signal. In the case of image and video compression, the reconstruction is aimed at the viewer's perception, while in lossy audio compression it is aimed at the listener's perception of sound. Lossy compression methods have much higher compression rates in contrast to lossless ones. The different lossy compression techniques use different algorithms and often only capture state changes from the previous state.
Lossy compression methods are particularly important for image compression( graphics, photos), audio and video compression. In the compression of images and graphics, one relies on the redundancy ofcolors, areas and lines, in audio on the irrelevance ofaudio frequencies and levels, and in video compression on the differences between successive images. Known compression methods for images and graphics are run-length coding, Tagged Image File Format( TIFF) and JPEG, for voice compression they are GSM 06.10, the A-Law method according to H.321 and adaptive delta pulse code modulation( ADPCM); for audio compression, these include MPEG, MP3, Audio Code Number 3( AC-3), Adaptive Transform Acoustic Coding( ATRAC) and Advanced Audio Coding( AAC); and for video and motion picture compression, H .261 and MPEG. Methods for text compression include run-length encoding( RLE), word coding, Huffman coding, and the LZW algorithm.
In modem technology, the V.42bis compression method is frequently used for data compression, which can be regarded as the successor technology to Microcom Networking Protocols( MNP) MNP4 and MNP5.