I would say it's likely. There are all sorts (pun intended) of optimizations you can do for each algorithm, but at the textbook level, I could say they seemed typical as far as speed expectancy goes. My rough estimates for the times were (in seconds): insert: ~10, bubble: ~27, selection: ~18, merge: ~17, and gnome: ~19. I have only written inserts, bubbles, and selections; with bubble being slowest and insert being the fastest, my limited experience agrees with the video. Note: the merge algorithm appeared to have sorted a different data set.

Given that the width of the elements being sorted changes (therefore there are more/less elements in different visualisations), I'd say no.

I'm not sure, but the visual part of the videos looks a lot like visualizations I've seen in algorithms classes where the speed does represent the speed of the sort. So I'd guess that it does here as well.

These are really great. As a musical person it engrained these concepts that much deeper for me. My wife, who is not much of a computer science person like the merge sort the best.