AllThatSounds: Associative Semantic Categorization of Audio Data

Abstract

Finding appropriate and high-quality audio files for the creation of a sound track nowadays presents a serious hurdle to many media producers. As most digital sound archives restrict the categorization of audio data to verbal taxonomies, this process of retrieving suitable sounds often becomes a tedious and time-consuming part of their work. The research project AllThatSounds tries to enhance the search procedure by supplying additional, associative and semantic classifications of the audio files. This is achieved by annotating these files with suitable metadata according to a customized systematic categorization scheme. Moreover, additional data is collected by the evaluation of user profiles and by analyzing the sounds with signal processing methods. Using artificial intelligence techniques, similarity distances are calculated between all the audio files in the database, so as to devise a different, highly efficient search algorithm by browsing across similar sounds. The project’s result is a tool for structuring sound databases with an efficient search component, which means to guide users to suitable sounds for their sound track of media productions.