11. Future Image Formats and Compression

Future formats for discussion only. This project was focussed on finding the most appropriate ways to scan a testbed of 10,000 Federal Theatre Project images with readily available, broadly applicable, reasonably standard tools and techniques. Thus the consultants and the Document Digitization Evaluation Committee removed from immediate consideration several new technologies but note that these should receive more study in the future.

Object wrapper file formats. Several new file format registration schemes are developing at this time. One interesting format is Bento, developed at Apple Computer. Bento provides a descriptive metadata wrapper around objects of all sorts which unambiguously defines their type. One key type attribute for an image or other multimedia object is the software needed to view, decompress or activate it. New standards for downloading executable content through a network, like Java, could allow the (platform-independent) viewer to be stored with the object, mitigating the preservation worry about data orphaned by obsolete platforms, but raising parallel concerns about the long-term executability of the stored program. As run-anywhere languages such as Java evolve and change, it will be essential for future Java Virtual Machines to know to what version of the language an ancient viewing program adheres.

Wavelet compression. Wavelet compression is a technique based on a class of pulse-like functions, unlike the cosine functions that form the basis for the JPEG baseline compression algorithm. Printed matter involves foreground information which was formed by dragging or impressing a dark, narrow marking device on a light background, producing narrow, pulse-like strokes. Because wavelets more closely resemble the types of data seen in manuscripts and other printed matter, a compression scheme based on them could be expected to offer better compression. Wavelet compression typically offers a lower level of artifacts at a given compression ratio than JPEG does. Unfortunately, no internationally sanctioned standard existed at the time the committee was meeting and thus it did not consider the use of wavelet compression in this project.

Next Section | Previous Section | Contents