BUT-CZAS: Korpus kvalitních nahrávek české řeči pořízených v bezodrazové komoře

Autor(en)

Vojtěch Hájek, Pavol Hárar, Jiři Schimmel, Radim Burget

Abstrakt

The paper introduces a novel database of human voice recordings named BUT-CZAS (Brno University of Technology, Czech Anechoic Speech), acquired in the anechoic chamber of the university. In total, the database consists of 405 mono-recordings of the text reading task acquired using the bit depth of 24 b and the sampling frequency of 48 kHz. Next, 18 speakers aged between 16–76 years old were involved in the data acquisition (9 women, 9 men). The overall duration of the recordings is approximately 315m (comprising more than 40 000 versions of 1 711 unique words). The database is designed with the special focus on the quality of recordings, gender-and-age balanced group of speakers, and the same environment during the acquisition. Finally, the full transcript is available for all the recordings.

Organisation(en)

Externe Organisation(en)

Brno University of Technology

Journal

Elektrorevue

Band

20

Seiten

48-52

Anzahl der Seiten

5

Publikationsdatum

04-2018

Peer-reviewed

Ja

ÖFOS 2012

202037 Signalverarbeitung

Link zum Portal

https://ucrisportal.univie.ac.at/de/publications/f5a3ec35-a80e-4df8-9133-be2d25235ef8