Central and Eastern European Survey

---

Type of Organisation: public
Number of Employees: Less than 10
Activities developed: Research, Education.

The Laboratory for Digital Signal Processing carries out various national and international research projects and is engaged in education and training in undergraduate and postgraduate study. Research Projects: The Laboratory for Digital Signal Processing is currently engaged in various research projects in the field of automatic speech recognition funded by Ministry of Science and Technology Republic of Slovenia -- project ARTES (Continuous Slovene Speech Recognition over the Telephone), European Union -- COPERNICUS-ONOMASTICA, COST 249 - Continuous Speech Recognition over the Telephone, SpeechDat II, and by industry -- SIEMENS Corporate Research and Development Laboratories. The laboratory has also developed three spoken dialogue systems: O-tel (reverse directory), HOBIS (home banking system) and VEDAMA (automatic correspondence system). In the framework of projects activities two Slovenian speech databases were developed -- SNABI (80 speakers, 200 sentences per speaker, telephone quality and studio quality speech) and SeechDat2 (1000 speaker, 5 minutes of speech per speaker, telephone quality speech). Education and training: The laboratory caries out undergraduate courses Digital Signal Processing and Man-machine communication as well as postgraduate study in the field of pattern recognition, automatic speech and speaker recognition, and speech synthesis (courses: Selected topics in pattern recognition and Fundamentals of language technologies). Computing facilities consists of 15 HP 9000/7xx workstations with HP-UX 10.20, 6 PCs with Linux o.s., programming language C++, different hardware components for digital communication with audio components DAT-link+ -- Digital Audio Interface, connection with telephone line -- TeleLINK 2S, laserprinters, scanner, etc. Software resources: HTK 2.1, Siemens Toolkit, HP Softbench for C++, KHOROS Pro, DSPlib for speech processing, Xwave for speech analysis. Speech databases: TIMIT; DARPA Resource Management; SNABI - Slovenian speech database -- 25 hours of read speech recorded in studio environment and over the telephone line, 80 different speakers where each read in average 200 sentences; SpeechDat II for Slovenian language (1000 speakers, telephone quality speech); SpeechDat for German language (1000 speakers, digits, continuous speech); Voice Mail (German and American English -- digits, command words). Members of the laboratory are members of the ESCA and IEEE.

This page is no longer maintained. Please visit http://www.elsnet.org/survey/quests to find out how to update your organisation profile or to find information about this organisation