Thoughts

I had a quick brainstorm and I want to put it here before I forget it. It’s one of those ‘nice and simple ideas that I’ll probably never get around to coding,’ so maybe you will!

Proposal:

The advent of internet video has opened the doors to many phenomenal lectures from world renown professors around the globe. Unfortunately, many of these professors tend to speak at length. They also spend a good amount of time scribbling on a blackboard writing long words or equations. They tend to pause while doing so, or worse, continue talking in a tangent direction! The former is time wasted in my opinion.
I propose a very simple video-viewer application that allows one to listen to the audio and video on-line lectures in a faster way… by speeding up the sound-track, but also by reducing the pitch of the sped up voice, so that the voice still sounds natural (instead of having an Alvan-And-The-Chipmonks-like quality). This can potentially allow people to “absorb” the big picture of what the lecturer is saying in half the time it would normally take. Since people usually only retain the big picture of a lecture, this could let people absorb twice as much as they would watching through it at normal speed. For fun, I’ll call this “Pedagogy Optimization.”

Problems with the Idea:

This of course ignores other modalities associated with learning. Speeding up the video would probably be too disorienting, though this should be experimentally verified. Also, professors couple blackboard or whiteboard writing with their lecturing. These visuals are key tools that enable the learning process, and lessening their display time may negatively impact perception of the ideas at hand. Another subtle issue may arise in time-sensitive lectures. These are lectures that only make sense in the context of normal time passage. For example, the demonstration of gravity in a physics lecture wouldn’t be natural in a framework of increased time speed.

Solutions to problems:

Allow for dynamic adjustment of playback. Default it to 2x speed, then, when the user decides they ‘missed’ something, with very simple and natural gestures, they ought to be able to ’slow down’ or ‘back up’ the feed, and change it continuously. They can do this through something like a digital touch surface on which different areas represent different parameters for playback, and simple slips of the fingertips in these areas adjust the parameters.

More Problems:

Modern day video compression is usually keyframe-based, meaning that in a compressed video like mpeg4, skipping to an arbitrary location takes processing time, and I believe reverse-playback isn’t possible with certain formats. Many algorithms assume a forward-playing invariant and exploit it during their lossy optimization.

More Solutions:

To get reversal, one can replicate the original source frames of the movie and make two versions, one compressed moving forward through time and another compressed moving backward through time. They can then be correlated via software, and toggled between when the user decides to ‘reverse.’

Slow-down and skipping in the videos can be achieved in the typical way done with today’s compressed video.