reply to post by CLPrime
Made it through the experiment with no cops called. Didn't even wake my parents up, I don't think. And, unfortunately, there were no explosives.
So...
the results:
It comes in quite handy that this is winter, because our furnace through the registers just happens to sound very similar to the sound people have
been hearing. In fact, based on recordings of both, it's pretty darn close.
I took a spectrograph of the furnace near a register, as a control:
In case you don't know how to interpret that, the real-time frequency spectrum is along the bottom (though that's not what we're concerned with).
The area above, with the blue stuff, is the important part. This is a spectral analysis over time, with the time axis going down the screen (real-time
at the bottom, and about 1 minute ago at the top). The frequency spectrum runs across the screen, perpendicular to time (as the Hz labels show), and
the colours denote amplitude. A brighter blue is a higher amplitude (a louder sound).
So, that's the sound itself, recording with the laptop mic next to the register to get the pure sound with no effect of distance. It's a
predominantly 120-400 Hz sound, diminishing smoothly beyond that with increasing frequency. This is very similar to the fracking sound, which was also
grouped in a specific range and fell off gradually from there.
Also, the red line running up the left side marks the 8 Hz frequency (I marked it just in case anything was "heard" at the 7ish Hz frequency, so I
could see it).
Now, part 1 of the experiment:
This is the spectral analysis of the furnace sound as heard from about 6 feet away.
Still paying attention to just the top of the image, with the blue lines, the top half (above the obvious horizontal blue line in the middle) shows
how the sound is now focused at two frequencies - 120 and 240 Hz.
The bottom half is to establish the effect of confining this sound. I took two blankets and used them to cover the distance from the register to the
laptop. As you can see, the effect was that the 120 Hz frequency stayed the same, while the 240 Hz frequency was significantly reduced, as was the
more faint region in between (albeit to a lesser degree).
This is exactly what I expected to see. This indicates that the confinement caused higher frequency and lower amplitude sound to be absorbed and/or
dispersed, leaving a localized low-end frequency. This would explain why the spread of the fracking sound became localized to a couple low
frequencies.
And, finally, part 2:
This was to test the effect of temperature, and it's a normalized spectrum to make any changes more obvious.
Again, the colourful part is divided by a horizontal line. The top half is the control - a recording with me holding the laptop and standing next to
the register.
For the bottom half, I opened our very handy patio door. Just for the record, it's -6°C (21°F) outside and 28°C (82°F) inside. I believe that
qualifies as a significant temperature differential.
The effect would have been too subtle to notice without the normalization, but it's pretty clear here. The sound, below 400 Hz, is amplified.
Again, I'm not surprised. kdog's link supports this.
So, we have two things happening:
- we have a localization of low frequency sound in confinement, accounting for the frequency spread of the fracking sound becoming localized frequency
peaks after having travelled some distance within some sort of confinement;
- and we have an amplification of this sound by a significant temperature differential (for instance, a temperature inversion), giving us a suspect
for what is confining the sound and allowing it to travel such distances.
Put these together and you get an otherwise uninteresting fracking sound being turned into an ethereal roar over some distance, under the right
conditions.