Continuing our experiments from yesterday creating representative thumbnails of a television news broadcast, let's look at how the two approaches (time-based fixed sampling vs algorithmic sampling) work on a few more broadcasts. Here we see that the fixed sampling strategy incorporates "upcoming broadcasts" preview clips and excessive advertisement clips, whereas the algorithmic sampler remarkably excludes them.
A discussion on human rights and Taliban Rule in Afghanistan on CSPAN. Note how both of the fixed sampling strategies incorporated an "upcoming broadcasts" preview clip at the start of the broadcast, whereas the algorithmic samplers skipped over it:
Fixed 5
Algorithmic 5:
Fixed 10:
Algorithmic 10:
Or this MSNBC broadcast. Note how the 10-frame fixed sample includes three advertising frames, whereas the algorithmic excludes them.
Fixed vs Algorithmic 5:
Fixed vs Algorithmic 10:
