I think that it is about time that I summarized some of the steps I have taken to help improve the PiP (picture-in-picture) webcam video included in my screencasts, with illustrative samples from the G14FUN Functional Analysis screencasts.
Thanks to Alan Mintey for lots of helpful advice.
I am trying out two different USB webcams, and recording the PiP video feed at 320×240 resolution, with other settings set to defaults (mostly automatic settings). The more recent recordings have been made using my own Creative Livecam Optia rather than the Logitech webcam I have borrowed from the school.
- Use mains power, not batteries
This was, of course, the most important lesson I learned from my early work with screencasting using Camtasia. On battery power, the laptop’s processing power is insufficient to cope with the demands on it. This leads to a variety of problems, including poor sound recording quality (popping), unsynchronized audio and video, and strange effects when writing in Windows Journal (e.g. unexpected spikes or long straight lines appearing, presumably when the tablet loses track of the pen for a moment). For some examples of what happens on battery power when everything else is otherwise OK, see G14FUN Lecture 14, part a (recap of preliminary discussion relating to Theorem 3.8, laptop running on batteries). It starts OK, but gets progressively worse. Scroll through it to the later sections to see what happens!
This is almost certainly the single most important thing when it comes to the video quality (sharpness and smoothness). You don’t want too much light shining on the data projector screen, as that will spoil the lecture for those present (though it does not affect the laptop screen capture, of course). But you do need sufficient on the lecturer, as otherwise the webcam can not cope. Compare results with poor lighting such as Lecture 1 and Lecture 2 with more recent videos such as Lecture 13, part b and Lecture 14, part b
Generally I try to arrange for the room lighting to be as high as possible without spoiling the data projector screen. I also take a desk lamp with a standard 40 watt pearl bulb with me to lectures. The lamp should not be too close to the lecturer, or else you get results such as those in Lecture 10 .
Once these two main issues have been addressed, the remaining issues are perhaps less essential. Still, there are many improvements that can be made.
If you compare Lecture 13, part b with Lecture 14, part b you can see that my shirt makes a big difference! Alan Mintey recommends avoiding clothing that is either too dark or too bright, as this will cause problems with the white balance/colour balance. Thus the colours have come out best in the second of these.
- Camera position and angle
Alan Mintey recommends (very sensibly!) that the lecturer, tablet and camera should be in a straight line, and that the camera should be tilted so that the lecturer is still visible when writing on the tablet, as well as when stepping back and slightly to one side. In this respect Lecture 13, part b is superior to Lecture 14, part b . I now also take a box with me to put the webcam on: you can see what happens otherwise in, for example, Lecture 4 , which was otherwise unexpectedly successful, given that I didn’t have a desk lamp with me. I think I was just lucky with the light that time!
- Leaning over the tablet
It is a shame that my tablet ends up so low down, so that I have to lean so far forward when writing on the tablet. Alan suggests that any (safe) means of raising the height of the tablet would help here. Ideally there would be a (safe!) sloped lectern to place the laptop on when writing. However, since I am looking for a “portable solution”, it may be too much to expect for every lecture room to have something appropriate. I am already carrying a lot of kit with me into every lecture (laptop, microphones, webcam, desk lamp, as well as a box of handouts and a box to put the webcam on), and I don’t think that adding a portable lectern to my list is going to be practical! I will be looking for some compromise here.
- PiP position
When the lectern is on one side of the room, I often end up looking towards the audience on the other side of the room. This means that I may appear to be looking out of the edge of the presentation instead of into it. Thus, where possible, I should position the PiP video on the appropriate side of the screencast. (This is an option during Camtasia production.) In other words, the PiP video should be on the same side as I am standing.
By chance I had this wrong for Lecture 13, part b and half-right for Lecture 14, part b