main-video
You might wonder why we don’t videotape the lectures.
Its easy enough to stick an HD camcorder in the back of the room. Unfortunately, this yields uneadable results (you can’t see what’s on the chalkboard).
It turns out, the problem isn’t resolution. Here’s an example - we took a video of a fake lecture with a Panasonic SD9 video camera (a decent HD camcorder). We then cropped out just the blackboard.
This picture was taken with a DSLR (a Canon, I believe). The original picture was a lot more pixels, but we downsampled it to be the same resolution as the video, and cropped it similarly
Other than the fact that the color balance being completely different (that’s a balancing issue with the fluorescent lights in the room - and easy to fix), you will notice the the DSLR image is easily readable. There are a number of reasons that its better:
- it has a sharper lens
- it has a better sensor
- it doesn’t have video compression
You could solve problems #1 and #2 with an expensive video camera. But problem #3 will be a problem since we need to compress the video to put it on the web.
Probably, what would work best would be to use a still camera (with a good lens and sensor) and have it take a picture every second or two. It wouldn’t be smooth motion video, but it would convey what was going on in the class. Unfortunately, this would be a hassle, since we’d need to synchronize it with the audio, and somehow find a delivery mechanism that didn’t ruin it with video compression.
In the idea world, we’d use a video camera (to capture the instructors actions) and a still camera (to capture what was written on the board) and somehow mix the two. But this would require a lot of work.
Next year, I might try to take the still images. This year, I am just audio recording.