@incollection{BodmerMoryFachSchmidtetal.2015, author = {Franck Bodmer Mory and Marcus L. Fach and Rudolf Schmidt and Wilfried Sch{\"u}tte}, title = {Von der Tonbandaufnahme zur integrierten Text-Ton-Datenbank. Instrumente f{\"u}r die Arbeit mit Gespr{\"a}chskorpora}, series = {Romanistische Korpuslingustik: Korpora und gesprochene Sprache}, editor = {Claus D. Pusch and Wolfgang Raible}, publisher = {Narr}, address = {T{\"u}bingen}, isbn = {3-8233-5436-1}, url = {https://nbn-resolving.org/urn:nbn:de:bsz:mh39-43256}, pages = {209 -- 243}, year = {2015}, abstract = {The development of tools for computer-assisted transcription and analysis of extensive speech corpora is one main issue at the Institute of German Language (IDS) and the Institute of Natural Language Processing (IMS). Corpora of natural spoken dialogue have been transcribed, and the analogue recordings of these discourses are digitized. An automatic segmentation system is employed which is based on Hidden Markov Models. The orthographic representation of the speech signal is transformed into a phonetic representation, the phonetic transcription is transformed into a system-interĀ­nal representation, and the time alignment between text and speech signal follows. In this article, we also describe the retrieval software Cosmas II and its special features for searching discourse transcripts and playing time aligned passages.}, language = {de} }