Lossy and Lossless Graphical Formats

Multimedia data is, in this context, any collection of data from images / graphics, video, audio and text and since their widespread use it has been important to use compression/decompression techniques to enable the fastest possible transfer speed over limited (and variable) speed connections whilst maintaining either full quality or acceptable quality to the recipient person or system.

Graphic formats present challenges in respect of compression as experienced in the publishing industry as the more advanced and professionally used graphic software packages are vector based and their target print systems are raster based. Vector based graphical files describe the resultant multimedia type (the most widely used being Encapsulated PostScript (EPS)) in terms of mathematics; points, lines and other geometric entities (PC Magazine Encyclopedia, n.d.). Raster based graphical files, such as those used widely on the Internet (GIF, JPG, PNG) and bitmaps (BMP), describe files in terms of pixels (i.e. every pixel in the file) and as a result have to store a large amount of data when taking into account size, colour and other pixel attributes. As a result those working with graphics professionally tend to work in vector but have to “rasterise” or render their image into a workable format for print or web and the resulting files are compressed.

An interesting format regarding compression is the Tagged-Image-File-Format (TIFF) as although it has not been updated for some years, it is still accepted by most, if not all, professional print houses however the provider of the TIFF has some choices to make regarding compression depending on their quality objective. Their choice of compression must first decide whether it is acceptable to use lossy or lossless compression. A TIFF can have JPEG, Lemple-Zif-Welch (LZW) or ZIP compression. JPEG is a lossy compression whilst the others are lossless.

Lossy compression, as implied by the name, loses some of the information in the file whereas lossless compression reduces the file size without losing information. This does not imply that lossless compression does not reduce file size. Lossless compression uses techniques (such as run-length or frequency dependent encoding) to remove redundant information from within the file without affecting the content of the file itself (Glenn, J. Computer Science: An Overview, 10th Edition. Addison-Wesley, Page 80).

Usually in the printed world, the best suited compression for TIFFs is lossless (LZW or ZIP compression), i.e. there is no effect on image quality and the resultant printed file, assuming that all other factors outside compression that could affect printed image quality (and there are many) have unintelligible impact, is the best it possibly can be. The deciding two factors on which type of lossless compression to use are the makeup of the file itself and the system being used by the recipient.

For the makeup of the file we would need to mathematically examine the blocks of single colour, use of text and spacing of objects. The resulting file size with be governed by the compression algorithm used by LZW or ZIP as they each consider these mathematical properties differently.

For the system being used by the recipient we only have to consider the age of the software for until the patents expired, software developers had to pay a royalty to the owners of LZW compression to include it, therefore old or inexpensive software tended not to include the capability to read such compressed TIFFs, despite the compression results being better, i.e. smaller files, no loss in quality.

Comment: By default on low to mid-range digital camera devices the default JPEG compression is set, in order to store more images on internal memory or memory card, to use lossy compression (i.e. baseline). Therefore quality is lost as a result and cannot be regained by the user after the picture is stored (or taken). Users can set options to lowest compression ratios but there are not that many amateur photographers that realise this resulting in an acceptance of what users see on the small screen on-device preview (low resolution) leads to disappointment when images are transferred onto higher resolution devices such as a monitor or above 600 dot per inch (dpi) printer.

References

PC Magazine Encyclopedia: Vector Graphics Definition [Online]. Available at: http://www.pcmag.com/encyclopedia_term/0,2542,t=vector+graphics&i=53728,00.asp (Accessed 8 November 2009).

Wikipedia: Tagged Image File Format [Online]. Available at: http://en.wikipedia.org/wiki/Tagged_Image_File_Format (Accessed 8 November 2009).

Wikipedia: Data Compression [Online]. Available at: http://en.wikipedia.org/wiki/Data_compression (Accessed 8 November 2009).

Sue Chastain: The Pitfalls of JPEG compression [Online]. Available at: http://graphicssoft.about.com/cs/digitalimaging/qt/overcompression.htm (Accessed 9 November 2009).

Wikipedia: Lossless JPEG [Online]. Available at: http://en.wikipedia.org/wiki/Lossless_JPEG (Accessed 8 November 2009).