BASS-dB: the Blind Audio Source Separation evaluation database
Terms of use
The use of the database and its content is subject to a set of licenses.
Should any change occur in the terms of use at any time in the future, the licenses that will apply to the content you may have obtained from the database at a given time are the licenses that were in force at the time you obtained the content. Please register to our discussion list if you want to be notified of such changes.

Scientifical use
You may exploit the database for a non-commercial scientifical purpose provided you mention it in any written work or software you derive from its use. Within a published article, paper or report, the database must appear in the bibliographical references as
    E. Vincent, R. Gribonval, C. Févotte and al.. BASS-dB: the Blind Audio Source Separation evaluation database. URL: http://www.irisa.fr/metiss/BASS-dB/
If you use the evaluation toolbox, you muse also cite it in the bibliography section as
    C. Févotte, R. Gribonval and E. Vincent. BASS_EVAL Toolbox User Guide. IRISA Technical Report 1706, 2005.

Use of audio data
Most of the audio data are made available under the Creative Commons Attribution-NonCommercial-ShareAlike 2.0 license. This means you can redistribute original or modified audio data provided they mention the original authors and follow the same license. Commercial use of audio data is prohibited. Some audio data are distributed under other licenses. Read the license of each dataset for details.

Use of the software tools
The evaluation toolbox is available under the GNU Public License. You can redistribute modified version of it provided they mention the original authors and follow a GPL-compatible license.
Blind source separation algorithms are distributed by their authors under specific licenses. Read the license section of each algorithm for details.

Authors
Authors of the database are E. Vincent, R. Gribonval, C. Févotte, A. Nesbit, M.D. Plumbley, M.E. Davies and L. Daudet. Authors acknowledge partial financial support from GdR ISIS (France) and EPSRC (United Kingdom).
Authors of audio data and software tools are mentioned separately within each dataset and each algorithm.