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 Moog Theremini Tutorial with thereminist Lydia Kavina, Shlomo beatboxing on the RODE M1-S Dynamic Microphone [HD], Eventide Pedal Effects Demo - Pitchfactor, Timefactor, Modfactor