User Interfaces for Speech-Based Retrieval of Lecture Recordings

Despite recent efforts in indexing and retrieving of audio data, common speech and audio search engines do return not only relevant results but also lots of documents that have no importance to the users. In addition, relevant information can be distributed over different parts of a file. Therefore, it is essential to represent the retrieval result in a way that enables users to easily filter relevant from irrelevant documents and to allow for comfortable and flexible skimming, selection, and access of interesting parts. In this paper we discuss a particular example for this problem, namely audio retrieval of lecture recordings. Six different interfaces are introduced, an evaluation of their usability is given, and a final interface design is presented.