Metadata creation as an accessibility and user value element
In this viewpoint I will share how the Cultural Heritage Project worked with ‘production’ of metadata as a process generated through our collaborations and dissemination lab.
Metadata is one of the most complex areas I have come across during my work with archives. Metadata secures future use and accessibility and is in this respect a cornerstone in digital archiving. But technology and the rapid changes in our digital possibilities are challenging our former ways while creating new demands and opportunities.
In DR (Danish Broadcasting Company), the Cultural Heritage Project was organized under the Director of Broadcasting, which has shaped a non-traditional metadata focus in the project. Archiving and digitisation is not a traditional task in the broadcasting department.
Ability to sign in
We have strategic responsibility for DR’s programs and products, and focus on creating and developing attractive offers for our users. The objective is to engage and create value for our users.
One of my main inspirations for the metadata ’production’ was ReCAPTCHA. The service requires the user to identify letters from a small picture to prove we are a human, not a computer. The picture is from a non-readable OCR-scan of a book and through our input (to identify the letters) we get something we want (the ability to sign in) while we help digitise books (secure our heritage). It is a fast service, easy and helpful for all involved parties – in my mind a really valuable concept.
In 2007 we spent a lot of time contem plating how we would be able to create value with the soon-to-be digital 400- 500,000 hours of content. The ideas were numerous, but when we started analysing the collections we often found a lot of inconsistent metadata.
The best and the worst
Our first concept BONANZA was launched in 2008. BONANZA was about activating 1,000 programs and shows and engaging the users. The concept was a joint project between DR’s Interactive Production Department and The Cultural Heritage Project. We created an online gallery and invited the users to vote for 10 favourites within 10 different categories.
The winners were digitised and made available in a high quality streaming version. The users loved seeing and engaging with our shared digital heritage.
Bonanza taught us that we had to use a lot of resources in order to secure a proper product. Both contextualization and the editorial task were huge. It became obvious that we had to experiment with other concepts where fewer resources were needed and preferably more metadata were created. In order to do this we took part in several research consortiums.
Creating a platform
Our vision is that the archive has to be used! In order to do this we activated our most complete part of the archive – the DAT collection. The DAT collection holds all aired radio from 1989-2005 (360,000 hours).
In 2010 the LARM consortium received a grant to establish an infrastructure and a number of different research areas. A part of the project was to create a front-end where students and researchers could access more than 800,000 hours of radio, recorded or digitised by The State and University Library and DR.
One of the most interesting results was the creation of CHAOS (Cultural Heritage Archive Open System). CHAOS was an Open Source platform, which could handle a massive amount of objects and related metadata. CHAOS was built to support technology development and reuse.
The platform helped us create www.larm.fm where the users could search and listen to the entire collections.
Users are engaged with the content through a number of tools or services. One of these was that researchers and students could annotate in all programs.
The annotations were linked to the program’s timeline and through this, the users created a relation between their knowledge and the content. Their work could be shared in their research community while enriching our shared metadata set.
New technology and user input
We are currently exploring many of our insights and developed methods in a new research consortium: CoSound. In Co-Sound we are combining some of the newest technology and user input to create new metadata, insights and products. Our objective is to create knowledge on how we can create an engaging experience using different game mechanics.
In CoSound we are working with the Technical University of Denmark (DTU). DTU has developed a set of services which are able to analyze sound. We are currently working on gameplays that will make it interesting for the users to use and engage with our archive.
We have tested some of our ideas and gameplays through a WC football game where the users competed on their knowledge about old WC football matches. They had to guess a game from the radio host’s excited commentary when a team scored.
Our actual focus is on a set of services, which work around the idea of a ’sound fingerprint’. The technology is sophisticated and is able to create a ’sound fingerprint’ from a human voice. The voice fingerprint should be able to tell us that person A is speaking here and here. But the service will not know who speaker A is.
The goal is to let users create metadata by interacting with content. In the game the users will be able to compete on their ability to identify the voices with a certain name of a host, a guest or participants in DR’s radio archive.
Whatever the result may be, I am truly looking forward to the forthcoming 12 months and to explore whether we can achieve some of the same benefits for our users and AV collections as ReCAPTCHA helped create for its users and scanned books.