Lecture 1 - More 'Easy listening'
Speaker Dr. Willemijn Heeren (UT/UL)
Date Thursday, 27 august 2009
Location Room H03-02 (Woudestein, Rotterdam)
Over the past century, millions of hours of spoken audio recordings have been collected with great potential for e.g., research, educational purposes and new creative audiovisual productions. The actual (re-)use of these collections, however, is severely hindered by their generally limited access. This is mainly caused by insufficiently accurate annotations, at the level of programs or tapes (i.e. durations up to hours) instead of user-manageable chunks, such as fragments of only a few minutes or less.
The HMI speech group is concerned with improving access to multimedia archives through the application of speech and language technology (SLT): within the project CHoral-access to oral history the focus lies on improved disclosure of spoken heritage collections.
In this lecture an overview is given of how SLT is being used to generate time-stamped, multi-layered annotations that support fragment retrieval from audiovisual heritage collections, how it can be used to enrich data from spoken word archives, and how the user experience may be enhanced accordingly both for professionals and for the general audience.
This Lecture is not open to the public!