JPEG2000: the solution to mass digitisation?

A number of libraries have begun major digitisation projects using JPEG2000 images. Why is this of interest? It's a question that Simon Tanner, from King's College, tackled at the opening of today's JPEG2000 seminar at the Wellcome Library. His short answer was, bluntly, that cultural heritage institutions are seeing the JPEG2000 format as increasingly attractive for mass digitisation because they can afford it; essentially, it's the only reason the format is considered.

This was borne out in subsequent presentations - that you can do more with less money - but there are other reasons that JPEG2000 can seem enticing. Basically, the cost savings are based on smaller file sizes - the Wellcome Library reported a 89% reduction in file size after compression before loss of 'visual' resolution in comparison to TIFF format for their needs. A smaller file size also means faster processing - high definition, zoomable images of large maps or entire newspapers, visually indiscernible to uncompressed TIFFS, are displayed in real time and rapidly integrate with other image software - two notable projects were the National Library of Norway's digitisation programme powered by Geneza, and the Old Maps Online and Georeferencer projects. Each library that has chosen to commit to JPEG2000 is therefore principally concerned with access: new online collections where more material is delivered faster.

Another notable thread to the disussion was the relative ambiguity of the format's preservation credentials. Richard Clark, the UK head of delegation to JPEG, queried why the digital preservation community hadn't been more involved in feeding back to the developers (compared to the commercial sector, notably the film and security industries). I suspect that communication will be on the rise, as it was noted today that future migrations of JPEG2000 images may result in the loss of ICC colour profiles. This, coupled with concerns about failing image resolution down the road, makes for some pretty fundamental preservation problems.

As observed by Johan van der Knijff of the Koninklijke Bibliotheek, just because JPEG2000 has been taken up by major institutions doesn't mean that the format is tried and tested in the long term context of digitisation. If anyone still thinks that digitisation is going to save the world's collections, this is some of the best evidence I've seen so far that this is a fantasy - a piece of paper, properly stored, will always outlast a digital file. So why bother? There is a balance between long-term preservation and access here, and while you making every effort to keep the stuff as long as possible (saving you the need, at least, to re-scan), these are programmes driven by access. While a secondary outcome may be the relief of pressure on your physical collections, digitisation projects largely reflect the new order of information retrieval.