Pointer Fun with Binky video.">
This is a short history of the making of the Pointer Fun with Binky video. The original idea for Pointer Fun was a cheesy Pascal video I made with the help of Brian Skinner. Brian aimed the camera, and I acted out a little scene with the Plah-Doh Binky about allocating and dereferencing pointers. We shot it in one day in the living room of the BHH. The video was used in our introductory courses which were taught in Pascal.
A few things about that video stuck with me...
The recognition of the importance of distribution is part of the motivation for the whole Stanford CS Education Library project.
At some point in 1996, I played around with a QuickCam digital camera and QuickTime multimedia editing software. The combination make it pretty easy to shoot stop-action video and add sound and text to it. It seemed like just the thing to make an improved, digital version of the Pointer Fun video. I was spending most of my spare time working on CS Education Library materials, but I decided that at some point, I would take a break from conventional documents and do the Pointer Fun video.
Finally the available spare time arrived over Winter break 1998. The basic steps were..
I planned what pointer concepts should be in the video. I used concepts so basic that they are true in all languages.
I wrote a basic plot between the offscreen voice and Binky. I figured out who talks when, and roughly how long in seconds their speeches are -- 10 hours of planning to get this far.
I played around with different materials for the video -- string, Plah-Doh etc. I made some little screen tests, play around with lighting and logistics.
I used two desktop halogen lamps to illuminate the "set" which is a little table next to my computer. The camera is taped to a pile of books about 1 foot up and 2 feet over from the set floor. A gray T-shirt pinned to the wall behind the stage is the background.
I shot a half-length preliminary version which I ended up throwing out, but where I got better at the planning and logistics for getting the video right.
I shot all the final video in one night. For this step, I give great credit to the QuickCam camera and QuickTime software. The QuickCam is a little camera which sends what it sees to a host computer in digital form. It's included software supports a "stop action" mode where each time you click a button, it snaps the current scene and adds it to the end of a growing movie. The video is about 200 seconds long, so at 5 frames per second that's about 1000 individual frames. I had everything pretty well set up and planned before hand. My chair hovered right over the set and I had most of the Plah-Doh and props pre-built. I could reach in, adjust things slightly, lean out, and click off a frame. Repeat.
Right before the final shooting, I realized that the easiest way to cover time during the speeches was to move Binky's eyes and eyebrows around rather than moving Binky -- I could move just the head features and click off a frame pretty quickly .
I made errors a few times during production. An eyebrow would fall off, but I would not notice right away and so produce video with a wrong looking Binky. I had to delete the bad frames, back the whole set arrangement up to the last good point, and shoot forward from there. The QuickCam software (which is itself based on QuickTime) is very basic, but with practice it can navigate these situations nicely. There were maybe four problems like that. In any case, the shooting went until 3:00 a.m. having started at 10:00 p.m.
At this point I had lots of video and no sound. Up to this point, things had proceeded pretty quickly -- maybe 40 total hours of preparation and shooting to get the video.
Next I worked on the script more.
Then I worked on the logistics of dubbing the speech over the right parts of the video. This required reading the text many times. Partly to get the voicing right. Partly to tune the text itself. This part took a surprisingly large amount of time with many adjustments and repetitions of little parts of the voicing. In places I would adjust the duration of parts of the (silent) video to better match the timing needed by voicing. Overall the sound took another 30 hours or so.
At some point, I decided to add the last "summary" section -- it just uses video from the earlier section, with new text and voice overs.
All of the video and sound editing was done with the free version 2.5 of Apple's QuickTime software. It's simple, but very capable. With version 3.0, the free version only plays, and Apple charges $30 or so for the "Pro" editing capabilities I used. QuickTime did a great job for me, although I think the part of the 3.0 version which pesters you to upgrade to QuickTime Pro "now" or "later" is unnecessarily irritating for an otherwise fine product.
All of the above work was done on an ordinary Apple PowerMac 4400 -- A 200 Mhz PowerPC 603e -- a pretty lame machine compared to the current G3's, but perfectly adequate for the job.
For the last part of the content, I needed to edit together the text for the bottom of the screen, and work on overlaying all the different versions -- 15 hours.
I experimented with the many different formats that QuickTime can export, before finally setting on using the Video codec for the video and the IMA sound codec. These worked on both Macs and PCs.
Finally, I created the companion text.
The total project time was probably around 100 hours spread out in "spare time" between late Dec 1998, and mid February 1999. Shooting the video was pretty easy. Most of the time seemed to go into editing and tweaking the video, sound, and text to all fit together. It might have been easier to record all the sound first, and then do the video knowing exactly how long each part should take (I believe this is how the Simpsons is done). It probably would have gone faster with a professional package like Adobe premier, but it costs a fortune compared to QuickTime 2.5 which was free and installed on my machine.
I was impressed how quickly many aspects of the video came together. The work was fun and the technology was well suited for the problem in most cases. Like any large project, there was work to get all the details right. With a little time and effort, it seems like anyone could produce one of these. To make things go quicker in the future, I might try to use a script where there were not too many timing constraints between the video, sound, and text.
Up to the Pointer Fun home
Up to the Stanford CS Education Library home