IKA CI Pop Music Dataset

<p><strong>IKA CI Pop Music Dataset</strong></p> <p>The <strong>IKA CI Pop Music Dataset (IKA-CI-PMD)</strong> is a dataset of music excerpts that has especially compiled to evaluate different music signal preprocessing strategies for CI listeners.<br> It contains 15 pop music excerpts with a duration of approximately 12 s each that are taken from the MedleyDB dataset (<a href="https://medleydb.weebly.com/">https://medleydb.weebly.com/</a>) curated by <a href="https://steinhardt.nyu.edu/marl/research/resources/medleydb">Rachel Bittner et. al.</a>.<br> Like MedleyDB, it is licensed MedleyDB is licensed under a <a href="http://creativecommons.org/licenses/by-nc-sa/4.0/">Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License</a>.<br> The excerpts have been selected such that they contain short self-contained phrases where vocals, bass, drums and other accompaniment are likewise present.<br> They also cover a range of different pop music subgenres and songs with both male and female lead vocals.</p> <p>The signals have been modified from the original MedleyDB versions. They have been remixed to equal loudness of vocals stem and all accompaniment stems where the mixing ratios between the accompaniment stems (bass, drums, other accompaniment) was retained.<br> Furthermore, the excerpts included in the dataset are normalized to a loudness level of -27 LUFS.<br> The signals are stored in the lossless FLAC format.<br> The files contain stereo signals, where both channels are identical.</p> <p>The dataset has been compiled at the Ruhr University Bochum <a href="https://www.ruhr-uni-bochum.de/ika/index_en.html">Institute of Communication Acoustics</a> in 2022 by Johannes Gauer (<a href="mailto:mailto:johannes.gauer@rub.de">johannes.gauer@rub.de</a>) in collaboration with the fellow researchers Anil Nagathil, Benjamin Lentz, and Rainer Martin.</p>