EvaldioData 1.0

  Authors

Rysová, Kateřina ; Novák, Michal ; Rysová, Magdaléna ; Polák, Peter and Bojar, Ondřej

  Item identifier

http://hdl.handle.net/11234/1-5731

 Project URL

https://ufal.mff.cuni.cz/automated-speech-scoring-czech

 Demo URL

https://lindat.mff.cuni.cz/services/teitok-live/evaldio/index.php

 Date issued

2024-10-31

 Type

audio, corpus

 Size

4773 utterances, 33267 tokens, 198 minutes

 Language(s)

Czech

 Description

EvaldioData 1.0 is the language corpus of spoken performances by non-native speakers of Czech. It includes recordings capturing the oral part of the Czech Language Certificate Exam. The recordings consist of dialogues between the examiner (a native speaker) and the candidate (a non-native speaker). In addition to the recordings, the corpus also contains their transcriptions, which are richly linguistically annotated. Some recordings are accompanied by multiple transcriptions from different annotators, allowing for comparisons of various transcripts of the same recording and evaluations of the degree of consistency in converting spoken language into written text. The current version focuses on the A2 level (according to the CEFR), which is required for the granting of permanent residency in the Czech Republic.

 Publisher

Charles University, Faculty of Mathematics and Physics, Institute of Formal and Applied Linguistics (UFAL)

 Acknowledgement

Ministry of Culture

Project code: DH23P03OVV037

Project name: Automatické hodnocení mluveného projevu v češtině

 Subject(s)

non-native speakers transcriptions

 Collection(s)

LINDAT / CLARIAH-CZ Data & Tools

Show full item record