videos in AudioSet
A list of about a million ten-second YouTube clips, each labelled with the sounds in it, that Google published in 2017. Labs use it to train and test general audio and music models.
Including Canon in D (Pachelbel) for Harp Combo (Harp, Violin, Cello), When I'm 64 (The Beatles) for New Orleans Trio (Trumpet, Tenor Sax, Trombone), St. Thomas (Rollins) for Jazz Quartet (Tenor Sax, Trumpet, Piano, Bass)