Using IIIF for images and audio/video at the Swedish National Archives
Mats Berggren - Swedish National Archives (Sweden)
Abstract: The Swedish National Archives has a public web service (http://sok.riksarkivet.se) that enables researchers to search the National Archival Database (NAD) and other publicly available databases and also to view images from digitized Swedish archival collections. During 2015 and 2016 new functionality for viewing digital images has been developed. The new system is based on the open source project “Universal Viewer” (https://github.com/UniversalViewer). The National Archives has participated in the project together with the British Library, the Wellcome Library and others.
The Universal Viewer based application for viewing images retrieves the images from an IIIF-server. The IIIF-server has been implemented in house by the National Archives following the IIIF-standard. A major reason for implementing IIIF for image access was a desire to make the image viewing application independent of image formats. The IIIF-server accesses images both in JPEG- and DjVU-format. The DjVU-format has been in extensive use within the National Archives but because the format has lost support in web browsers there was a need to build a solution that solved this problem. The application DPSS from Cuminas is used to instantly convert DjVU-files to JPEG when they are requested through the IIIF-server. This solution has made it possible to implement a new application for image viewing that can be used on all devices without having to do a major conversion of all existing DjVu-images to JPEG.
On the 1st of February 2018 the National Archives digital images became accessible free of charge. Before that date researchers had to pay a fee to access images. Because of the expected increase in image access the architecture for image access had to be upgraded. Currently the search application runs on an MSWindows webserver. The image viewing application accesses a pool of 13 IIIF-servers working in parallel fronted by a load balancing cluster consisting of two servers. The images (both JPEG and DjVu) are stored in an object store based on the Hitachi Content Platform (HCP). The search engine is implemented with SOLR and MSSQL Server is used as the database engine.
The National Archives has about 200 million digitized images of which about 100 million are publicly accessible. The new image viewing system based on Universal Viewer and IIIF has worked very well. It withstood the onslaught on the 1st February when the images became free to view. Currently the work load is approximately three times the work load before the 1st of February.
The National Archives is currently digitizing analogue audio- and video-recordings in collaboration with the Royal Library. During 2018 the possibility to show and play audio- and video-recordings will be added to the online search service. The same basic architecture will be used, Universal Viewer as the client application and the IIIF-servers for file access. The IIIF-implementation has to be modified to allow retrieval of audio-and video-files. We are very interested in other efforts to use IIIF for audio and video.
Presentation type: 20 minute presentations (plus 5 mins questions)
- IIIF and archival collections,
- IIIF synergies with regional and national ongoing digital libraries, museums and archives initiatives,
- Audio/Video IIIF implementations,
- Emerging use cases for IIIF technical specifications
- Universal Viewer,
- Open source,