Subject: IOWA Sound Samples Preparation for Instrument ID From: Joey McKay <joey.mckay@xxxxxxxx> Date: Tue, 11 Nov 2008 15:21:35 -0000 List-Archive:<http://lists.mcgill.ca/scripts/wa.exe?LIST=AUDITORY>
Dear all, I'm wondering has anybody automatically separated the notes in the IOWA music samples? In the area of automatic musical instrument identification (AMII), these samples have been used as the testing and/or training corpus in many approaches, but I've yet to find literature which details how the individual notes have been separated. For the purposes of my own research in musical instrument identification, manual note separation using a sound editor is feasible yet time consuming. I've spent some time trying to automatically separate the notes, using onset detection algorithms developed in house, but the algorithms are imperfect. For the purposes of AMII, the onsets of notes have been determined as important in the literature. So as not to introduce imperfections into my system, I'm looking for the 'ideal' solution, if it exists? Without sounding averse to the manual task, are the individually separated notes for the IOWA samples available publicly online anywhere? Maybe I'm over complicating this issue, but if there is somebody out there who has preprocessed the IOWA samples and/or has thought this task through, then maybe they can assist? Thank you, Joey McKay.