Central and Eastern European Survey

---

NL and Speech Resources available at the organisation: BABEL Multilingual Speech Database-Hungarian, LIAS-Language Independent Automatic Speech Segmentiser, MTBA Telephone Speech Database, SPECO Multilingual Multimodal Speech Training System.

Name: BABEL Multilingual Speech Database-Hungarian
Nature: speech
Language: Hungarian
Size: 2,5 hours
Format: SAM format
Coverage: prose
Medium: CD-ROM
Availability: ERLA

Software description: BABEL is a multilingual speech database for five Eastern Europien languages. These are Hungarian, Polish, Romanian, Hungarian, Bulgarien and Estonian. The database contains clear and read speech. The collection formatting of the data conforms to the protocols established by the ESPRIT SAM project and the resulting EUROM databases.

Name: LIAS-Language Independent Automatic Speech Segmentiser
Nature: speech
Language: Language independent
Format: SAM format
Coverage: prose
Availability: free for research purposes

Software description: A neural network based automatic segmentation technique was developed. While the segmentation method uses the so-called broad phonetics classification, it gives the opportunity of developing a system, which is good for many languages. Thus, if a phoneme set of a language is transcribed into the international SAMPA characters, and SAMPA transcription of a sentences or paragraphs are given, the automatic segmentation works, and gives good result for English, German, Estonian, Hungarian, Bulgarian, Polish and Rumanian. There is no omission or addition of labels. The obtained boundary shifting from the hand made one are between 81-90% within the (25 ms from the hand made one. The segmentation result is the best for clear read speech. The method gives help in the segmentation of the clear speech and noised speech, too.

Name: MTBA Telephone Speech Database
Nature: speech
Language: Hungarian
Size: 3 hours
Format: SpeechDat format
Coverage: prose and spoken dialogues
Medium: CD-ROM
Availability: commercial product

Software description: The database contains the voice of 500 speakers (300 wireless speech and 200 mobile speech). The text and the format of the database are equivalent of Speechdat. The phonetically balanced sentences are segmented on phoneme level.

Name: SPECO Multilingual Multimodal Speech Training System: English, Swedish, Slovenian and Hungarian version
Nature: speech
Language: Hungarian
Size: 3 hours
Format: SpeechDat format
Coverage: prose and spoken dialogues
Medium: CD-ROM
Availability: commercial product

Software description: The SPECO System is a new Speech Teaching and Training software system for four languages: Hungarian, English, Swedish and Slovenian. This SPECO System (SPEech COrrector) aims to help to develop or correct the speech (articulation, intonation, loudness, rhythm etc.) of children with speech disabilities. It is very important in our system that we present the speech parameters in a way that is understandable and interesting for young children, while remaining correct from the acoustic-phonetic point of view. The program presents the important cues for different phonemes in sound pictures and emphasizes important parts with amusing drawings to make the pictures understandable for 5 to 6-year-old children. The system is based on up-to-date technology, but we follow the steps of traditional speech therapy in both modules. These are sound preparation, sound development, followed by training in words and automation (meaning the achievement of a reliable production not requiring further instruction). Specific tasks have been constructed in a specific order involving the teaching experiences of the teachers of a given language. The base of the SPECO system is a general language-independent measuring tool, a database editor and the database. The database editor made it possible to construct modules for all participant languages and for different sound groups.

Name: MULTIVOX text-to-speech synthesizer
Language: 10 languages

Software description: text-to-speech multilingual system: multilingual grapheme sound conversion, prosody modelling, formant synthesis supporting 10 languages

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