Using standardised metadata delivered continuously from broadcasting companies, media files can be connected to searchable metadata through an automated process. To manually input the metadata that TV and radio broadcasters provide in their programme listings, the Swedish Media Database (SMDB) would need almost 300 full-time employees. Fortunately, however, the metadata can be automatically collected.
SMDB is a search service that has been covering Swedish broadcasts and publications since 1979, and includes some older material as well. It contains TV broadcasts, radio transmissions, videos, films, music and multimedia material such as computer games. All in all, it contains almost 8 million hours of recorded material. It is only available in Swedish and researchers, authors and journalists have access to the audio-visual collections. SMDB’s system is built to automatically import files and metadata. When this is done the files are ready to be played vthrough the interface.
“The SMDB system automatically connectsmedia files and associated metadata through structured file-naming conventions. This allows metadata to be entered before or after the actual media files are delivered. When the metadata is in place, the media can be discovered through the SMDB web interface. Once the media files are on the SMDB servers, the media can be played directly by the researcher on the premises,” says Olle Johansson, Bibliographic Expert at the National Library of Sweden.
Metadata in different forms
For phonograms, films, videos and multimedia the metadata must be entered manually, but radio and TV has automated input. The metadata is delivered from the broadcasters or purchased from a news agency.
“Using standardised formats such as XMLTV for the exchange, the metadata is easy to process systematically. Sveriges radio and SVT deliver the metadata in their programme listings, which is the basis for the TV and radio guides in newspapers. The good thing is that these companies deliver their legal deposits through a system that automatically connects the file with the metadata in the programme listing”, Olle Johansson says and adds, that the automated metadata needs to be monitored, but that it works really well.
Asking for files
“We could never catalogue all this information manually; we would need nearly 300 full-time employees to catalogue all these broadcasts on a national biblio graphic level, which is almost as many people as work at the National Library.”
The automatic connection between the file and the metadata is only made when the legal deposits are delivered by file.
“The majority of the TV and radio companies deliver their broadcasts on files via FTP but a small number still use physical data carriers, mostly CD-Rs and DVD-Rs. In those cases we have to describe the carrier and connect that description manually to the metadata already delivered. According to the law, we cannot demand delivery by files, but we always ask for it, and I am certain that file delivery will be standard procedure in a near future”, Olle Johansson says.