I was really impressed by this introductory course for machine learning by Google. Not only is it well designed and put together, but as someone who is extremely hard of hearing, I took special interest in the subtitles for the video section.
The closed captioning is incredibly well done. It's more than just subtitles, it shows a time-stamped list of the spoken part that scrolls automatically. This is much better than subtitles for learning material, since the viewer can see not only what is currently being spoken, but quickly review what was just said as well as what's coming next.
They're even clickable so that one can skip ahead to salient parts of the video. Oh my god -- a way to skim and basically page-flip in a video. I'm on cloud nine.
I'd be happy just with subtitles, but this interface knocks it out of the park. Videos in place of documentation is more and more common, but even simple subtitles are a luxury for most learning material. Autodesk's Fusion 360 training materials, for example, are almost entirely video based and have no subtitles whatsoever.
Millions of people in North America suffer from some level of hearing loss, and half of them are still in the workforce. If your documentation is all on video where the subject matter is unfamiliar to the viewer and the speaker (and their mouth) is not visible, it makes learning far, far harder than it needs to be.
Anyway, not only is the video and captioning done extremely well in the course I linked, the course itself is really well done. If you have any interest in this area at all, give it a look.