Build an audio description pipeline that actually fits the gaps

TL;DR We're building an audio description pipeline in Python: find the silent gaps with...