"And cut!" Exploring textual representations for media content segmentation and alignment