Gravesend Inn-frasound: Creating and Measuring Infrasound

In Part II of this series, I laid out the infrasound research study we recently completed at City Tech; this entry details the issues of generating and measuring the infrasound itself.  This gets a bit technical, so if you're not interested, please feel free to check back later--I'll be posting results of the study here as soon as we have them.

For our experiment, we needed to generate sub-sonic “infrasound”, and, as a sound engineer who has frequenty worked with large subwoofers and who feels pretty comfortable patching up weird systems, I thought this would be a pretty straightforward process.  Fortunately, though, I long ago learned that when I'm doing something I've never done before, I need to leave plenty of time for unexpected developments, and that was certainly a good strategy in this case, as I detail here.

We picked a 19 Hertz (Hz, or cycles per second) sine wave as our infrasound source, since that’s the frequency Tandy and Lawrence had found in their initial investigation of a “haunted” space (see Part I of this series for details). For our initial tests last spring, we used an Audio Toolbox as our signal generator, and connected it up to our ancient Apogee P-10 subwoofer processor and a Crown K2 amp, which drove an Apogee AE-10 double 18” subwoofer (these units work just fine, but haven’t been made in years.) We got the sound going, but after a minute or so, it would cut out. I figured out that the processor was apparently protecting the subs against the “bad” infrasound, and was cutting out. So, I bypassed the processor and its protection circuits, and (carefully) drove the subwoofers directly from the amp. This worked just fine, but I was a bit concerned about damaging the speakers themselves.  Fortuitously, over the summer I managed to get for our department a massive, modern Meyer Sound 650-P powered subwoofer.  And after some tests in late summer we determined that the unit could generate quite a bit of 19Hz without clipping, and we figured the level was pretty good because by the time we pushed the 650-P up to its limits, the effects of the infrasound would be obvious to anyone in the room, which wouldn’t have been acceptable for our purposes. 

To control the experiment, I used Medialon Manager show control software, which, while common in the world of show business, has not likely been used in the world of psychological research before. I will talk more about Manager's role in the experiment in the next entry, but I planned to use Manager's built-in “local media player” to play back both Todd’s voiceover out the left channel of the computer output, and the infrasound on the right channel.  This way, I could make the system (which would have to be setup and struck repeatedly) as small as possible, and I could enable/disable Manager's infrasound output through the software code.  To test this, I used Adobe Audition’s “generate” function to create a 19Hz sine wave mono .wav file (you can download a 10 second sample of a 19Hz sine wave here).  I planned to play this out of the computer via a Sound Devices USB Pre professional audio interface, but when I tested it, I got no infrasound, and instead a horrible digital popping sound. The USB Pre is really designed for high quality input into a system, and only has unbalanced, consumer level outputs, which were incredibly noisy.  I figured this might be the problem, so I switched to my personal M-Audio Fast Track Pro audio interface. Switching to this interface cleaned up the interfacing noise, but the 19Hz signal would still not play--just popping.  I did a bunch more testing on this, and my hypothesis was that there is some sort of brick wall filter in the signal chain in the computer, and my 19Hz signal was beating right up against it.  

Running out of time, I abandoned the idea of playing infrasound out of the computer and went to the idea an external generator feeding an older Yamaha 01V Mixer that I could control from Medialon Manager.  The Audio Toolbox works great, but its user interface is clunky, and if it loses power briefly or anything happens it defaults back to a mode that takes about 10 steps to reconfigure. So, I searched around online and found a Mastech SFG-1002 function generator. I was a bit dubious about the unit since it was so cheap, but I figured I would be measuring the output anyway and could verify its operation.  When the unit arrived I immediately plugged it in, and it immediately blew up (power supply capacitor). I got a replacement, and (surprisingly to me) the unit was rock solid through our whole experiment, and you can see below that it generated a very clean 19Hz signal throughout the experiment. So now, we had a workable system: a laptop running Manager controlling everything and playing back Todd's voiceover through an older Meyer UM1 (non powered) speaker (our equipment choices were limited since these trials took place during the Gravesend Inn, which uses a large part of our audio inventory); the Mastech signal generator generating 19Hz, and the Yamaha 01V mixer switching the infrasound on and off and setting the level of Todd's voiceover under the control of Manager; and the Meyer 650-P acoustically generating the infrasound signal. Here's the white board block diagram: 

Several months later (after the Gravesend Inn scenery had been loaded out), we got the subwoofer (large black monolith in the picture at right) positioned in our scene shop, in the corner above our tool cage, where it should get a boost from the hard walls in the corner and the hollow, drum-like platform, while simultaneously being out of the way for daily shop operations and out of view of  the research subjects. Using my M-Audio Fast Track pro, I measured the output of the system using Rational Acoustics’ SMAART measurement software, and I got all kinds of crazy readings, showing what looked like harmonic distortion well up into the audible range. So, running out of time on a night where we brought in our research volunteers for training, we placed the audience in the spot where we thought the 19Hz was loudest. We did a trial run for the volunteers, and even though the sound was quite apparent to me, none of the volunteers thought it was even on, so we felt good about the level we had chosen.

Some time later, after having further problems in an unrelated class demonstration with SMAART, I finally got time to go back and test out the M-Audio Fast Track Pro to try and get to the bottom of the apparent distortion evident on the screen. Through some careful testing, it turned out that the Fast Track Pro was giving me completely bogus measurements, but the USB Pre worked fine (at least for input).  So, the laptop for the experiment actually has two audio interfaces on it--the USB Pre for SMAART measurements, and the M-Audio pro for playback of Todd’s voiceover. Here’s a little video I shot demonstrating the M-Audio errors:


Finally, with a working measurement system, we went back to the shop and set up things again, and I discovered something very interesting walking out the shop to take a break.  In one spot by the entrance, the "infrasound" became inaudible, but it felt very weird standing there.  Putting the measurement mic there, it turned out that at that spot there was very intense 19Hz build up there, but very little audible harmonics. It turns out that we had thought all along that we could “hear” the 19Hz signal, but what we were actually hearing was its third harmonic: 57Hz (this had shown up in the measurements, but was buried by the garbage data).

We then moved the measurement mic all around the shop, and found that the 19Hz level was reasonably consistent in most of the shop (not surprisingly, given its roughly 60 foot wavelength), but that the 57Hz level varied quite a bit. To ensure that the research subjects were exposed to as much infrasound as possible and as little of the audible harmonics as possible, we searched around the shop to find a zone with good 57Hz cancellation large enough to accomodate our research subject group (boring video documentation here). This caused us, shortly before a "dress rehearsal" to completely move our planned audience area (more boring video documentation here).

Each night of the experiment, we measured the infrasound using SMAART with the mic in the center of the research subject area (SMAART data captures available on request).  Here's a composite measurement I captured last night showing three traces (I'm using spectrum measurments because transfer functions don't make much sense with a single sine wave input):

The orange colored trace is the output of the mixer to the subwoofer (which measured out to about -16 dBV), which is clean 19Hz sine wave (documentation video here of the signal chain).  The blue trace shows the actual measurement of the room during a typical trial condition; the green trace shows the same but when the subway comes by (the shop is over the A/C subway line).  Interestingly, we discovered that the subway noise is well above the infrasound frequency.

This trace shows an "infinite" average during the period of one trial:

The 19Hz level, on average, was approximately 40dB above Todd's voiceover, which you can see in the graphic in the range of roughly 200 Hz up to 1K. Todd's voice at the listening position averaged about 70dB SPL, A weighted, and surprisingly to me, the 19Hz was also quite measurable in C weighting (which doesn't de-emphasize the low frequencies as strongly as A weighting) at about 80dB SPL C weighting at the listening position (video here).  This is about 15dB SPL above the background noise (I have more boring videos of all this if you're interested, and lots more pictures here).

So, I feel pretty confident that we delivered a strong 19Hz sine wave to the research subjects, and through careful control of information, as as far as we know, none of our research subjects (even my sound engineer friends who went through) figured out that we were actually testing for infrasound.

As for the experience of the infrasound itself, I personally can see how it might cause people to have "unusual" experiences.  I experienced a lot of it, often late at night by myself in "spooky" places, and while I never thought I saw a ghost, I did find the 19Hz infrasound to be quite irritating. Also, it might be coincidence, but Dr. Riggins got a terrible headache during our first experiments with the new subwoofer in a smaller space in our basement where the sound levels were pretty intense.  And when Toija, Andrew and I were having a discussion during the process and had left the infrasound running, I often just had to go over and turn it off, because I found that it just grates on me.

In the Part IV of this series, I describe the show control aspects of the experiment.