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 "This door is too small (for a bear)" Needcompany / Grace Ellen Barkey, "Talking Head" Liquid Loft, "GRIND" - Jefta Van Dinther | Minna Tiikkainen | David Kiers