How Files Are Compressed

[juju]

One compression algorithm is not meant for every type of file. Moreover, some file types simply are better suited to compression than others. That’s why there are many compression schemes. Apple Macintosh machines often use SIT and RAR file formats while PCs gravitate heavily to ZIP formats. Used primarily to compress and/or archive groups of files, these compression schemes are known as being “lossless” because no data is lost during compression or decompression. In contrast, compressed still image formats, such as JPEG (Joint Photographic Experts Group); motion video formats, such as MPEG-4 (Moving Pictures Experts Group-4); and audio formats, including MP3 and WMA (Windows Media Audio), are all great not only at shrinking files but also in letting users control how much compression to apply. The trade-off is that these schemes work by eliminating data, making them “lossy” formats. Clearly this wouldn’t work with text documents. Imagine extracting a file and finding half of its letters gone. Traditional compression technologies, such as LZW and ZIP, perform very poorly with video and audio data because these programs don’t understand multimedia content. Audio formats, on the other hand, are designed specifically to recognize and manipulate audio signals based on models built around the capacities of the human ear. Why save the data for a tiny bell when a crashing cymbal buries that sound? Similarly, if you know your audience’s speakers can only play frequencies up to 16KHz, there’s no point in preserving all the data from 16KHz to 20KHz, much less anything above that, which is beyond the range of most human hearing.


From Smart Computing