From the Citiot Desk

Signal Processing


Another catchy title from the citiot’s desk, right? My last offering to this audience was “Math” in which I hoped to make some room in your heart for a topic that often frustrates. This time I’ll ask you to become an engineer since you’re so good at math now.

So, what is signal processing? Broadly put, it’s the act of separating the wheat from the chaff, pulling important information from all the junk that surrounds it — signal from noise.

Think of your friend, the audiophile, who insists that vinyl is the only way to really hear music. Maybe there are a few pops or a hiss that distract you, but doesn’t it feel warmer, more real? What he (it’s probably a guy, maybe even an engineer) hears is better, more of what he wants to hear. I believe him but for me, the medium is pretty irrelevant, it just reminds me of the music that’s already in my head. We’re processing signals differently, no right or wrong about it.

In my career, I had the honor of working with engineers, mostly in electronics, optics and software. In making a particular kind of instrument, I was the biologist who helped them decide which signals were important for clinicians and researchers in various fields of life science. In prioritizing, of course first the signals had to be real, with no feedback from the amplifiers, no wonky connections, and no software gremlins, but beyond that the fascinating challenge was about what matters. What do we want to hear?

Early in a project, you want to hear it all. Humbly, you build a device that lets you hunt around, wander in virgin territory, and see what’s what. When you decide on your signal, then it’s about never missing it, no one could miss it, and it always means the same thing. Broad horizons, then robust insight, both of those designs are tricky, but we have signal processing technologies that are powerful, not just for me or my projects, but to manage our brave new world of data.

Our phones are the most immediate demonstration. In your hand, signals from space. Come on. Video chat with your friend in Cleveland, or Buenos Aires, anytime, kind of for free. The engineering in that better shock you. And let’s think about the signals going the other way, from you to the world.

Every click, every pause, every quick scroll, logged. It’s identity theft, sure, but more interesting. It’s a signal processor of you. There’s nobody in a cube in Cupertino, Redmond, Mumbai or Guangzhou looking at your data. Signal processors filter it, amplify it, aggregate it, trend it. They know I’m an old white man, and they know what that means probably better than I do. They don’t need my social security number. No coincidence that whether they’re selling solar panels or reverse mortgages they include an image of an attractive woman, at least what I will think of as such. I wonder what the ads show on my wife’s phone.

We’re pretty good signal processors ourselves, but a lot is coming at us these days, and the old wisdom is wearing thin. Trust is earned, and familiarity breeds contempt, but not enough to help me get through all this noise and find the truth. We make choices all the time about our time, how we invest it, and with whom. And that means we avoid a lot. Our experience is filtered, our reactions amplified, our days aggregated. The world we know is not objective, not even fair usually. We know what we like, and we like what we know, and our lives are heavily processed, like a hot dog from a pig.

Most of what we try to teach our children are signal-processing crafts and strategies; how to read, what to read, what to believe, who to love. I hope you’re good at it. Hope I’m good at it. I’ve been at it a long time, and I promise I’m still learning.

Maybe the most resonant teaching I can offer from my professional and personal years at the task is the importance of deciding what signal you want. Of course, it has to be real, not some yammering bot or devious marketeer. But beyond that, what matters? What matters to you and your friends, your country, our planet? How many times will I get worked up over something and only later remember that this is not important, those signals are noise and oh boy I could have handled that better.

There’s a balance between signals you want and signals you need. A pleasant life is a fair goal. As I age, I am unapologetically slipping toward the beautiful signals. Honestly, I can’t handle noise like I used to. My memory, my hearing, my vision, my strength, my open-mindedness, all less. I pulled my neck the other day when I changed my mind. To learn anything now is more work, so it’s got to count

If I can help someone I hope I will, but I need to get some ugly signals — information about things that are not right — so I can make choices based on some kind of truth, and act on it intelligently when I choose. I hope you can find signals that matter, that your circuits are not overheating, that your software is bug-free, and that when you wake to face the bright morning you can comfortably smile and embrace all that you’ll learn today.

Jack Dunne lives gratefully in Lakebay.