TTCI, pt2

(There was somethin’ in the city that he said he couldn’t breathe
And there was somethin’ in the country that he said he couldn’t leave)

As laid-back and beautiful as it was, it also seemed kind of.. too small, dead-end and going nowhere fast.

I don’t know why I’m thinking in the context of Places I Want To Plop Myself Down And Start A Family And Stuff. Because I’m nowhere near that.

My inner being screams at the thought of plunking down in the middle of a big city, where I can’t even see the stars, or a tree without a ring of asph(gra(mulch)ss)alt around it, taking up space in a median strip. But as much as I’d like to claim I’d be perfectly content to live naked in a cave out in the middle of some untouched wilderness, I don’t think I really would. And the middle ground, suburbia, is…well, suburbia. Yechh. One thing that really stood out (not so much in Pueblo, but in Denver and points in between) was the sheer number of identical tickytack townhouses, creeping up on the interstate and lurking at just about every residentially-zoned grid coordinate. Anonymous little dwellings stacked against each other, 3,4, in some places even 10 alongside. Even the ritzy stuff, rows of what looked like million-dollar mansions sharing walls.

Houses with different colored doors / So you can tell which one is yours

Maybe I’m just hard to please. Maybe I just don’t like that idea of commitment to any particular place. Anyway, maybe I’ll drop this train of thought for now and come back to it when I figure that shit out :-)

So anyway, more trip stuff. A few Rockies (and some more), some flat stuff, and a Big Ass Burger (this rare species discovered at a small Pueblo diner). We hauled down DOT Road and back once more, at relativistic speeds (look ma, straight deserted road and no cops!), recovered our equipment and the data thereof…everything looked good, so we didn’t really need to stick around there to record a second day of the moving and the shaking. So we cancelled the last night of our wifi fortress and headed back to Denver, where we bummed around most of the day sightseeing, checked out the Red Rock amphitheater, and grabbed some dinner at a place that serves buffalo hamburgers (mmmmm). Somewhere along the line we came back to the house we hung at the first day to see what was left of our snowcastle. I was shocked! Here it is after 2 days.

Oh yeah, the testing. We got data…lots and lots of data. There’s still a glitch somewhere in my hardware/firmware (small life lesson on eating your own dogfood, when it hasn’t been tested on the dogs yet, so to speak), but it only manifested an average of a few times per hour, so for our purposes we weren’t screwed. (It means more nights of hair-pulling in the lab, but that comes later.) We even got energy, even on our totally untuned harvester (the purpose of all this vibration recording stuff is to figure out how to tune these things). The actual amount is comparable to the feeble glow of a lightning bug’s ass, but that should soon change. Today was the day of grovelling through all these piles of numbers, and cooking the occasional glitches out (the nature of the glitch was to randomly drop a byte or two, so the fix was to identify this in software, and basically just drop more bytes until both the channels and endian-ness synced up again, even though it means losing a few samples here & there), and just for fun, performing parlor tricks like piping raw accelerometer data from one of the channels through SoX and playing back the surprisingly clear and audio-like sounds from the resulting .WAV file. If anyone asked, the purpose of this (besides showing off general l33tness and explaining “what is this data?” to everyone who can’t tell just by looking at the pretty graphs, which is, well, everyone) was to use that big, heavily optimized, massively multiparallel lump of signal processor between my ears to detect which accelerometer channels were saturating heavily, which isn’t really visible by staring at a plot of raw data that’s been through crazy 8th-order antialiasing filters. (Surprisingly, this ‘signal processing’ method actually worked pretty well.) I was a bit too busy to see if I could extract our voices on it during the first couple minutes of data before testing began…(besides, that’s getting too close to just plain fucking off, and sooner or later I actually have to deliver something!), but I was able (in the name of data sanity-checking) to identify the length of the test track and each of its various materials-test segments by their unique noise (or lack thereof) profiles, and roughly recreate the shape of the test track right down to the sharpness of every curve, assuming a constant speed kind of a bogus assumption, but not *too* far off.

Sorry, I just totally geeked out. It won’t happen again. (Okay, yes it will…)

The money shot! All right, a little piece of money shot, anyway. (Heh, I just realized, given where the expression “money shot” comes from… this is engineer porn.) Dedicated to Dr. R.J.W., presumably still dreaming in Fourier (more than I do).


Leave a Reply