Språkbanken Speech is part of the Division of Speech, Music and Hearing at KTH (Royal Institute of Technology). The department was founded in the early 1950s and has been active in the fields of speech communication and speech technology for almost 70 years. Research conducted at the department is interdisciplinary and includes linguistics, phonetics, visual and auditory perception, cognitive science and experimental psychology. This research provides a basis for the development of multimodal systems where speech, music, sound, gestures and other phenomena are combined to create human-like communication. The division is part of the Department of Intelligent Systems at the School of Electrical Engineering and Computer Science.
Within CLARIN, Språkbanken Speech comprises a K-centre with a focus on speech analysis: CLARIN Knowledge Centre for Speech Analysis, CLARIN SPEECH. Through the K-centre, Språkbanken Speech provides expertise in the collection, processing, annotation and exploitation of large multimodal speech corpora. Studies of such speech and dialogue corpora enable and facilitate a large number of related research activities in the humanities and social sciences, ranging from speech and gesture research to studies of behaviour and social signals in human communication and interaction.
Within Swe-Clarin, Språkbanken Speech focuses on developing and improving speech technology resources and tools for Swedish. The main applications are found, for example, in the areas of speech synthesis and speech recognition. Activities in these areas are currently being pursued in terms of recording speech data and developing analytical methods to handle the large amounts of speech and video data stored at Swedish public authorities and in Swedish archives. Machine learning and data-driven methods are crucial for managing these recordings and for developing the speech technology that underlies both application research and basic research in a number of areas. Språkbanken Speech has collaborations with most Swedish universities, and also conducts many research projects together with partners in industry, the public sector, and non-profit organizations, such as Wikimedia Sweden, SVT (Swedish Television), the Swedish Agency for Accessible Media, Bonnier Publishing House, the National Library of Sweden, and the Institute for Language and Folklore.
Jens Edlund, edlund@speech.kth.se