In Part III of this series, I laid out the infrasound generation issues of our research experiment; this entry details the show control system we used to run the experiment.
Medialon Manager is a powerful and widely used development environment for show control; we use it to run the Gravesend Inn haunted hotel, I teach it in classes at City Tech and Yale, and I've done a number of commercial projects with it as well. I’ve long felt that Manager is well suited for a wide variety of control tasks, and our experiment really is, at its core, a show, so it made perfect sense to use Manager to control it.
Manager's strength is that it offers control support for a wide variety of show devices, and offers the ability to edit the code of the program while the show is running (something you could never do with a compiled programming language). From a control perspective, this is a very simple show, controlling only a few devices: a “local media player” (built into Manager) to play back the Todd Robbins voice over (described in Part II of this series); a MIDI-controlled lighting dimmer to control the lighting in the room, and a Yamaha 01V mixer (see Part III for details on its use) which turned the infrasound on or off. I timed out the lighting on/off times from the voiceover, and wrote a very simple timeline in Manager to control everything:
Here's a little video showing how the system worked:
I started out on the project thinking that I would use Manager’s random number generator to determine whether or not the infrasound would be on for any given trial. Later, though, I realized that I would have to massage the random seed to ensure that an equal number of trials did and did not have infrasound. Testing that kind of code to ensure that it was working properly would have taken more time than I had avaialable, and I thought that having an equal number of trials with and without infrasound was more important than having it be completely random. So, I programmed the system to simply turn the infrasound on and off for each trial, and then instructed Andrew Puccio, our volunteer technician, to periodically disrupt the on/off sequence so that the research attendants would not be able to detect a pattern.
Given the nature of the setup, Andrew was going to know whether or not the infrasound was on anyway, because it’s definitely obvious when it switches on or off, and the 57Hz third harmonic is quite audible outside our test subject area (see Part III for details), which is where Andrew was positioned. Because of this, to minimize any potential bias, we made sure that Andrew had no interaction with the research subjects.
One other useful feature of Manager for this kind of work is the ability to easily create date and time-stamped event log files, which I used to log normal startups, abnormal shutdowns, and also the infrasound status of each trial. Below is an excerpt; you can download a complete log file for all the research trials here.
[ 29/10/2010 - 20:00:21 - Message ] Trial Number 34 started with an infrasound status of 0
[ 29/10/2010 - 20:02:26 - Message ] Trial Completed Successfully
[ 29/10/2010 - 20:04:13 - Message ] Trial Number 35 started with an infrasound status of 1
[ 29/10/2010 - 20:06:18 - Message ] Trial Completed Successfully
The system worked flawlessly for the trials! And now, head onto Part V for the results!