Well spotted on the 2:17. It wasn't auto exposure, it was me adjusting the iris because I noticed zebras on the screen were increasing a bit. I haven't ever used auto anything on my camera, perhaps I should give it a go some day, I don't know if I'm a bit manual mode fixated.
I recorded the sound with a Sennheiser ME66/K6 direct into the camera, which you can see at 3:12. I also used the zoom as a backup but it wasn't needed. I didn't tweak the sound except for a bit of volume balance.
Although she is a very bright kid, she is still a 2 year old and it took us hours to get her to "perform" for the camera. So I don't have as much other footage as I would like but I am dealing with the long clips right now. I think you must be stood over my shoulder because I'm doing exactly as you suggested.