All my videos are just taken with my camera, and about 30 seconds long. I think 30 seconds is a good amount of time to get a point across, especially when it comes to baby development.
Her speech patterns are so interesting to me. I need to get a video of her babbling these days, because she seems to "work" on a different syllable every other week or so. At 2-3 months, she was mostly just making vowel sounds. Then she got the occasional G down (a-guuuuh) and now she's on to Bs but without a vowel at the end, just ah-bbbb. I'm fascinated by how humans learn language from scratch.
I'm also fascinated by how easily we look around and just take stuff in, but that's because of my grad school work in image processing. I mean, we were trying to teach computers to recognize images like we do, it's the last thing that computers really aren't good at. I could hand you a fistful of photographs and you'd charge through them with no effort, you could tell me "this one is of a beach", "this one has desks and a person in front by a whiteboard, it must be a classroom", "this one is a family because the people sort of look alike but are different ages". It's no effort.
We watched a national geographic special on infant development, and one of the things they said we learn to do as humans is to focus in on what's important. Like, they learned that these tiny babies could recognize the faces of apes they'd seen before. But months later, the babies could no longer tell the difference between one ape and another, they were all just apes. The babies were trained to focus on human differences, to distinguish familiar people. For my graduate work we did some of that, writing programs so the computer would recognize the moving parts of a video because that's important information. There was a whole team trying to write a program to detect highlights of sports games, based on motion and finding a ball. But it's so much effort there, and so easy for our brains, even when they're very small.