Workbook on Digital Private Papers > Administrative and preservation metadata > Preservation metadata
Preservation metadata
Metadata specific to object types
PREMIS provides a generic preservation model applicable to all object types; it does not support the detailed technical metadata that is particular to object types. To give an indication of the nature of the metadata needed for this purpose, a number of metadata schemes being used by digital repositories for technical metadata are listed below. This list is by no means complete, as more schemes for audiovisual material are starting to emerge owing to the growth of web services in this area. It should be noted that some of the more generic elements in these schemes overlap with PREMIS elements, and the repository must decide whether to record them redundantly, or in a single place. More work is needed in understanding the specific technical metadata requirements and how they may be produced before an application profile can be determined.
Metadata for still images
MIX for Images (Z39.87)
Drafted by the National Information Standards Organization, the Data Dictionary - Technical Metadata for Digital Still Images is extensive. Containing more than 100 elements, it relates mainly to images produced via digitisation projects, where metadata capture can be embedded in the process, but much of the standard is applicable to digital images entering repositories via digital archives collections. MIX is an XML implementation of the Data Dictionary that is maintained by the Library of Congress. The Jhove tool is capable of outputting some of the MIX metadata set from the GIF, JPEG, JPEG 2000 and TIFF image formats, and tools have been written to extract the MIX XML from the Jhove XML.
The areas covered by the NISO Data Dictionary are as follows:
- Basic Digital Object Information.
- Basic Image Information.
- Image Capture Metadata.
- Image Assessment Metadata.
- Change History.
Harvard Digital Repository Service Image
Metadata set to record details about image attributes and image production (20 elements) developed for Harvard’s digital repository, which records metadata about image attributes and the image production process.
National Library of Australia Image
Limited metadata set (10 elements) for still images developed at the National Library of Australia.
Digital Images Archiving Study
The Digital Images Archiving Study by Sheila Anderson et al. (March 2006) is a useful overview of issues relating to the preservation of raster and vector images.
Exchangeable image file format (Exif)
Exif is a metadata specification for digital images and was created by the Japan Electronic Industry Development Association (JEIDA). It includes descriptive metadata, such as date and time, and technical metadata about camera settings at image capture.
Metadata for moving images
National Library of Australia Video
Limited metadata set (7 elements) for moving images developed at the National Library of Australia.
Library of Congress VideoMD Data Dictionary
Developed as part of the Digital Audio-Visual Prototyping Projects at the Library of Congress, this scheme comprises 16 elements.
MPEG-7
A multimedia content descriptions standard for moving pictures and audio developed by the Moving Pictures Expert Group (MPEG).
Harvard Digital repository service Audio
Metadata set to describe the attributes of an audio file. Comprises 26 elements.
Library of Congress Audio (Source) Data Dictionary
Developed as part of the Digital Audio-Visual Prototyping Projects at the Library of Congress, this scheme comprises 18 elements.
National Library of Australia Audio
Metadata set comprises 7 elements.
MPEG-7
A multimedia content descriptions standard for moving pictures and audio developed by the Moving Pictures Expert Group (MPEG).
ID3
ID3 is a container for metadata stored in MP3 format. Many of the semantic units relate to descriptive metadata, such as title and performer, but there are also units relating to useful administrative metadata, such as file type, size and length.
Metadata for text files
Schema for Technical Metadata for Text by Jerome McDonough
Metadata for text files, with an XML schema, developed by Jerome McDonough at New York University. Responsibility for the maintenance of this schema is unclear. Comprises 19 elements.
National Library of Australia Text
Simple text metadata scheme comprising 5 elements.
Metadata for databases
National Library of Australia Database
Six elements for database-specific metadata.
Metadata for executables
National Library of Australia Executables
Metadata to record the code type and version of an executable file.
Deciding what metadata is useful for particular content types
Few archivists will have had cause to consider the technical properties of digital content types and may find it difficult to understand the importance and potential usage of the metadata elements chosen by the initiatives listed above. Very few explain the reasoning behind the elements chosen, and the archivist may therefore need to look to other sources to gain an understanding of these elements. The Sustainability of Digital Formats web page, hosted by the Library of Congress, is a useful place to learn more about the specifics of formats commonly found in digital collections; the site includes assessments of qualities and function important to content types that should inform metadata application profiles. Implementing many of these metadata is presently difficult as tools which support the extraction of the metadata are not widely available. JISC has initiated a set of studies on the significant properties of software, vector images and moving images, which may provide useful information for developing technical metadata profiles.
