| Dear Robert, I do not know of such database, but I am generally very impressed with the advantages of machine learning/pre-trained models to separate vocal tracks out of full audio mixes. There are many commercial solutions (such as Moises.ai), and also Logic Pro as a widely used and quite affordable digital audio workstation has a feature they call “separate stems” which does exactly that. You might be able to obtain acapella rap/hip hop tracks in much less time than one might have thought two years ago or so. Best wishes, Jonas Prof. Dr. Jonas Obleser Chair in Physiological Psychology and Research Methods Department of Psychology University of Lübeck MFC VIII, Maria-Goeppert-Straße 9a 23562 Lübeck, Germany Phone +49 (0)451 3101 3620 jonas.obleser@xxxxxxxxxxxxxx Meeting room https://uni-luebeck.webex.com/meet/jonas.obleser — sent from mobile device – Prof. Dr. Jonas Obleser Chair in Physiological Psychology and Research Methods Department of Psychology University of Lübeck MFC VIII, Maria-Goeppert-Straße 9a 23562 Lübeck, Germany Phone +49 (0)451 3101 3620 jonas.obleser@xxxxxxxxxxxxxx Meeting room https://uni-luebeck.webex.com/meet/jonas.obleser — sent from mobile device – On 7. Mar 2025, at 06:32, Robert J Zatorre, Dr <robert.zatorre@xxxxxxxxx> wrote:
|