Home | Digitisation (Summary)
Full Text: Page 1 | Page 2 | Page 3 | Page 4

GOOD PRACTICE GUIDELINES

The process of digitisation is superficially simple, since an inexpensive PC and scanner can be used with standard software to create a digital copy of any human-readable artefact. However, when issues such as the quality of the resulting image, the format used to store it, its description, its intended use and its preservation are considered, the process becomes much more complex. The very simplicity of the basic process can lead librarians and policy makers into a false belief that a digitisation programme will be straightforward and inexpensive. It will be neither.

The delivery of digitised materials can be relatively straightforward, since the development of the Internet and the World Wide Web has provided the infrastructure, software and technical standards needed. Again, however, the very ease with which the Web can be accessed masks some of the underlying difficulties. For example, if images are not simply to be given away, complex software and procedures may be needed to collect revenue and to guard against infringement against intellectual property rights.

It is important to recognise that digitisation is often taken to mean only the creation of digital images of originals. More complex processes which seek to capture the meaning of the text (or image or whatever) – such as Optical Character Recognition (OCR) – may also be used and these are considered briefly below.

Planning
It is first necessary to be clear as to why digitisation is felt to be appropriate, the precise selection criteria and the audience being addressed. It will be noted while none of these issues are first and foremost strictly technical, they are all essential to a successful project. They lead on to a decision on what should be digitised. (See also diverse cultural content)

For an example of digitisation selection criteria see University of California Selection criteria for digitisation. Lee (see below) summarises the necessary questions and suggests the use of a ‘decision matrix’. Colet’s ‘Planning an Imaging Project’ will be found useful.

The scope of the project will have to be considered carefully – it is easy to underestimate the effort needed to digitise traditional materials, and it is almost inevitable that only parts of most collections can be handled. Cost, staffing and other issues (see below) will contribute to these decisions.

IPR (Intellectual Property Rights) (see also copyright)

There are two aspects to this:

• Does the library hold the right to make digital copies? If it will be necessary to secure the rights to the material, this will inevitably be a long drawn out and probably costly process.

• What rights to the digitised images does the library intend to impose? Will all images be made freely available, or will thumbnails be displayed on the web and a charge be made for higher definition copies?

File formats
The output from a digitisation process is an image, or other information object, in digital format. However, there are many different formats in use and that chosen will have to be selected with care. Bear in mind that it is usual to hold both a master and delivery copies, and it is often appropriate for these to be in different formats. Common formats used for digitisation projects include TIFF (for high quality masters) and JPEG or GIF (for delivery copies), but remember that there are different versions of all formats. The following criteria need to be considered (based on C. W. Brown and B. J. Sheperd. ‘Graphics File Formats, Reference and Guide’. Greenwich, Connecticut, USA: Prentice Hall, 1995, as described in Franziska Frey ‘File Formats for Digital Masters’ in the US Guides to Quality in Visual Resource Imaging series,).

Although a large number of formats are available it is likely that the practical choice will be fairly limited. Taking a lead from other projects, and taking expert advice, is recommended.

Digitisation hardware and software
The hardware choices, broadly, are:

• Scanners. It should be noted that simple flat-bed scanners are not likely to be suitable for bound volumes and can cause considerable damage. Where it is possible to split originals into separate pages a document feeder will be found to be advantageous, as it can speed up processing and improve accuracy. Specialised scanners are available for digitising slides and microforms. Don Williams has written a guide to scanner selection in the US Guides to Quality in Visual Resource Imaging series.

• Digital cameras. Where a book cannot be opened flat, or where material is delicate, a camera is much more suitable than a scanner since it avoids physical contact between the equipment and the original object. It will be necessary to build or acquire a book cradle and to provide suitable, low temperature lighting.

For non-visual media, such as audio recordings, equipment can be purchased to provide digital output from analogue media.

The software used in the digitisation process needs to be chosen with equal care. There are three processes to consider: capture, processing and delivery – the last of these is considered separately below. Image capture software is often bought with the hardware, and this can be a good solution. It must support common standards (e.g. TWAIN) and should provide basic management features such as image preview – this enables adjustments to be made without a full scan. The chosen file formats must, of course, be supported. Image processing software may range from the simple to the highly sophisticated, depending on the intended use. At one end of the spectrum, a project may decide not to process captured images at all. Another project, however, may decide to ‘clean’ the images to remove scratches or stains which appear on the original and to adjust colour balance etc. manually.

The choice of appropriate hardware and software will be dictated by a number of other considerations, including purchase and maintenance costs, ease of use (need for special training), throughput requirements, adaptability and extensibility, ability (quality and choice of output), standards compliance, reliability and so on.

Home | Digitisation (Summary)
Full Text: Page 1 | Page 2 | Page 3 | Page 4