Advanced Listening Practice

I've worked the overnight shift at WKRN for eleven years. Public radio, out of a transmitter in the Berkshires, which is a polite way of saying we cover a listening area of roughly forty miles in every direction and our annual fundraising drive barely pays for the coffee. I'm an engineer, not a presenter. My job is to make sure the signal goes out clean, to monitor levels, to switch between the network feed and our local programming at the correct times, and to sit in a room full of equipment from midnight to six in the morning making sure nothing breaks. It's not glamorous work. Most nights, nothing happens at all, and that's a good night.

It's a good job if you're the kind of person who doesn't mind silence, and I am that kind of person. Always have been. Even as a child, I preferred the hours after everyone else had gone to bed, the particular quality of stillness that a house has when you're the only one awake in it. My mother used to find me in the kitchen at two in the morning, reading at the table, and she'd say, "Sandra, go to bed," and I'd say, "I'm not tired," which was true. I was never tired at two in the morning. I was alert, focused, completely present in a way that the daytime, with its noise and interruptions and obligations, never quite allowed. The station has that same quality, amplified. After midnight, the building empties out completely. The last of the evening staff leaves around eleven-thirty, and then it's just me and the equipment and the faint hum of the transmitter, which is a sound I find, and I realize this will seem strange, genuinely comforting. It's a constant. It doesn't change pitch or volume. It just persists, and there is something reassuring about a sound that simply persists without asking anything of you.

What I want to tell you about is the dead air.

In broadcast terminology, dead air is any unplanned silence in the signal. It's considered a failure, a gap, evidence that something has gone wrong, that someone wasn't paying attention. My job, among other things, is to prevent it. And I'm good at my job. In eleven years, I have allowed exactly two instances of dead air longer than three seconds, both caused by equipment malfunctions that were beyond my control, and both corrected within fifteen seconds. I take this personally. I understand that most of our listeners at two in the morning are either asleep or not listening closely, but that's not the point. The point is that the signal is a promise. You put something on a frequency, you're telling anyone who tunes in that you'll be there. Silence on a broadcast frequency isn't just absence. It's a kind of betrayal of the listener who is depending on you to be there, even if they don't know they're depending on you, even if the radio is just background noise in an empty room.

But dead air is never truly silent. This is something that most people don't know and that engineers learn early and take for granted. Between segments, between programs, in the fraction of a second between one piece of audio ending and the next one beginning, the frequency is still active. It carries noise. Electrical interference, atmospheric disturbance, the faint thermal noise of the equipment itself. If you amplify dead air, you hear a wash of static, like the sound of rain on a distant roof, textured and continuous and utterly without meaning. It's the sound of the medium itself. The sound of the air between the sounds.

Or so I believed. For ten years, I believed that.

About fourteen months ago, I was running a routine check on our archival recordings. We record the full broadcast signal continuously, twenty-four hours a day, onto a digital server. The recordings are kept for thirty days and then overwritten, a rolling archive that exists primarily for regulatory compliance. Nobody listens to them unless there's a complaint or a technical issue. But I make a habit of spot-checking, particularly the overnight hours, because that's when problems are least likely to be noticed by anyone else and most likely to accumulate quietly until they become serious.

I was listening to a recording from the previous Tuesday, specifically a transition between the end of a BBC World Service relay and the start of our local classical music block, which airs at 0200. There's a four-second gap between these two programs. I'd flagged it because four seconds is longer than it should be, the automation should produce a gap of no more than one and a half seconds, and I wanted to understand why.

I amplified the gap. Standard procedure. Boost the gain, apply a noise filter, listen for anomalies. And underneath the static, faintly but distinctly, there was a voice.

Not a bleed from an adjacent frequency. Not crosstalk from the phone lines. Not a fragment of audio that had leaked from the previous segment. I checked all of those possibilities immediately, because that's what you do, that's the training, and I eliminated each one systematically. The voice was present only in the dead air. Only in the silence between programs. And it was reading, as far as I could make out through the noise floor, what sounded like a newscast.

I spent the next three weeks going through archived recordings, every night after my shift ended, sitting in the control room with headphones on while the dawn came up over the Berkshires outside the window. I examined every transition, every gap, every instance of dead air longer than two seconds. The voice was not always there. Some nights, nothing. Some nights, a few words, fragmentary and indistinct, like someone speaking in the next room with the door closed. But on certain nights, usually Tuesdays and Thursdays for reasons I cannot begin to explain, the voice was sustained and clear enough to partially transcribe.

It was reading news stories. They were structured like news stories, with the cadence and phrasing of a broadcast journalist: location, event, context, consequence. Short, declarative sentences. Neutral delivery. No editorializing, no commentary. The kind of voice you'd hear on any news program anywhere in the country, reading copy that someone had written for them. Professional, practiced, and completely without affect.

The stories hadn't happened yet.

I should qualify that. Most of them were mundane. A water main break on Route 7. A school board vote in Pittsfield. A minor earthquake in the Connecticut River valley, magnitude 2.3, no damage reported. Events that, in the days following the recording, occurred. Not all of them. Perhaps sixty percent. The rest either didn't happen or hadn't happened yet at the time I stopped tracking them, and I acknowledge that I may have stopped too soon.

I recorded everything. I have timestamps. I have transcriptions, partial ones, of thirty-one separate broadcasts, if that's the right word for them, embedded in the dead air of a public radio station in the Berkshires. I have not told anyone at the station about this. I've thought about how to frame it, rehearsed conversations in my head, and every version sounds like the opening of a story about a woman who has been working alone in a dark room for too long. Which I have. But that's not the point, and I resent the implication, even when I'm the one making it.

The broadcasts that trouble me are three in number. Three stories, read in that same neutral, professional voice, that describe events I very much hope will not occur. I'm not going to share the details. I've thought about this carefully, more carefully than I've thought about most things in my life, and I've concluded that sharing them would be irresponsible, because either they're going to happen regardless of what I say, in which case publicizing them serves no purpose except to cause fear, or they're not going to happen, in which case I would have caused fear for nothing. There is a third possibility, which is that the act of publicizing them might somehow prevent them, but that presupposes a mechanism I don't understand and can't rely on, and I am not willing to gamble on a mechanism I don't understand.

So I sit with them. I come to work at midnight. I monitor levels. I switch between programs. And in the gaps, in the silences, in the spaces between one piece of audio and the next, I listen. The voice is still there. It was there last Thursday. It read a story about a fishing boat out of Gloucester that would lose its engine in heavy weather and drift for nine hours before the Coast Guard reached it. I wrote it down. I'll check the marine reports next week. I always check.

I want to ask you something, and I know you can't answer, but I want to ask it anyway because asking it out loud makes it feel less like something I'm carrying alone in a room where nobody else can hear it. If you could hear the future, but only in fragments, only in the spaces between other things, and if most of what you heard was harmless, ordinary, the texture of daily life continuing as daily life does, but some of it wasn't, what would you do? Who would you tell? And would they believe you, or would they hear only the static?

I think you know the answer to that. Good night.