Voice Activity Detection: In this first assignment, we will create a dataset that simulates speech in every-day scenarios. We train a classifier on this dataset for distinguishing voiced from non-voiced sections, a task called voice activity detection, VAD for short. This, of course, requires a ground truth in terms of VAD annotations.