BUT Speech@FIT Reverb Database

This is the first pre-release of BUT Speech@FIT Reverb Database. The database is being built with respect to collect a large number of various Room Impulse Responses, Room environmental noises (or "silences"), Retransmitted speech (for ASR and SID testing), and meta-data (positions of microphones, speakers etc.).

The goal is to provide speech community with a dataset for data enhancement and distant microphone or microphone array experiments in ASR and SID.

The database has Apache 2.0 license and you can download it here:
http://www.fit.vutbr.cz/~szoke/speech/ReverbDB/BUT_ReverbDB_rel_18_06.tgz [17,3 GB]

The BUT Speech@FIT Reverb Dataset consists of 2 rooms: a middle size office room and a small size hotel room. We placed 31 microphones in both rooms. The source (a hi-fi loudspeaker) was placed on 6 positions in the office and 5 positions in the hotel room. We measured RIRs (using exponential sine sweep method) for each speaker position (so we have 11 times 32 RIRs). Next we recorded environmental noise (silence). There was a radio at background playing in one speaker position in the office.

We also retransmitted LibriSpeech Test-clean dataset for 2 positions of speaker in the office. This data is freely available from our web-pages along with the RIRs. We also retransmitted a portion of NIST Speaker recognition evaluation 2010 dataset, the availability of this data is limited to sites that have valid LDC license to the original data.

All microphone positions are measured and stored in meta-files. We pre-calculated positions of microphones and speakers in Cartesian and polar coordinates as absolute and relative (to the speaker).

Please see attached README.txt for more detailed description of data

More rooms and environments will come soon (actually we have 7 rooms "measured"). If you want to publish a paper using this dataset, please refer to this page. The speaker recognition work making use of this data has been described in

Mošner Ladislav, Matějka Pavel, Novotný Ondřej and Černocký Jan. Dereverberation and Beamforming in Far-Field Speaker Recognition. In: Proc. ICASSP 2018. Calgary, 2018. http://www.fit.vutbr.cz/~imosner/pubs.php?id=11717


Ondřej Novotný, Oldřich Plchot, Pavel Matějka, Ladislav Mošner, Ondřej Glembek: On the use of X-vectors for Robust Speaker Recognition, in Proc. Odyssey 2018, les Sables d’Olonne, 2018. https://www.isca-speech.org/archive/Odyssey_2018/abstracts/54.html

We are currently writing a technical report and more papers about it.

Feel free to provide us with your feedback to szoke@fit.vutbr.cz with a subject mentioning BUT-ReverbDB.