Ghost Frequency

Chapter 1: The ROM

The oscilloscope pattern didn't make sense.

Elijah Reynolds leaned back in his chair, the springs protesting with a familiar creak, and rubbed his eyes with the heels of his palms. The clock in the corner of his monitor read 2:17 AM. His apartment—a converted loft above a closed barbershop on South Congress Avenue—was quiet except for the hum of his server rack and the distant thrum of Austin nightlife filtering through the windows.

He'd been at this for six hours. The ROM dump from the Starfield Corridor arcade board sat open across three monitors: raw hexadecimal on the left, a disassembly view in the center, and a waveform visualization on the right. The game itself was nothing special—a second-tier space shooter from 1983, produced by a small outfit called Parallax Interactive that had folded before the year was out. Elijah had acquired the original PCB from an estate sale in Waco three weeks ago, sandwiched between a stack of water-damaged Atari 2600 cartridges and a broken Pac-Man cocktail table.

He collected these things. Not just for nostalgia—though he had plenty of that, having grown up watching his father feed quarters into machines at the Putt-Putt on Burnet Road—but because the code itself fascinated him. Early arcade games were marvels of constraint. Developers in the late seventies and early eighties had worked with processors running at a few megahertz, with memory measured in kilobytes. Every byte mattered. Every instruction was deliberate. There was an elegance to it, a kind of digital poetry written under impossible limitations.

Elijah had made a modest name for himself in the retro computing community by reverse-engineering these ROMs, documenting their techniques, and occasionally discovering lost features—debug menus, hidden messages from developers, prototype stages that had been disabled but never removed. He'd found a complete unused level in a 1981 Galaga clone. He'd uncovered a developer's marriage proposal hidden in the sprite data of a Donkey Kong bootleg. Small discoveries, but they mattered to the people who cared about preserving this history.

Starfield Corridor was supposed to be routine. The game used a standard Z80 processor with a simple sound chip—a Yamaha YM2149, three channels of square wave synthesis. The gameplay was unremarkable: pilot a ship through a scrolling star field, shoot enemies, collect power-ups. The code was clean but unexceptional.

Except for what was hiding underneath the audio.

Elijah had noticed it almost by accident. He'd been mapping the sound data, tracing which memory addresses the game accessed during its attract mode music, when he found a block of data that the sound routine referenced but never actually played. It sat in a region of ROM that should have been empty—padding bytes, usually filled with zeros or repeating patterns. Instead, it contained structured data. Thousands of bytes of it.

At first he'd assumed it was leftover development data. That happened sometimes. Developers would compile their code and the linker would pull in objects that weren't actually needed, leaving orphaned data in the final ROM. But when Elijah extracted the block and visualized it, the pattern that emerged was unlike anything he'd encountered in thirty years of studying vintage code.

It was regular. Deeply, precisely regular. Not random noise, not leftover code, not a simple repeating pattern. The data contained nested structures—sequences within sequences, with mathematical relationships between the layers. When he'd converted it to audio and run it through a spectrum analyzer, the waveform showed characteristics of a modulated signal.

A signal embedded in a video game from 1983.

Elijah pulled up the waveform again and zoomed in on a section near the beginning. The pattern repeated every 4,096 bytes with slight variations—like verses of a song with the same melody but different lyrics. Within each repetition, he could identify at least three distinct layers of information, each occupying a different frequency band.

"What the hell are you?" he muttered.

He'd tried every standard analysis he could think of. It wasn't ASCII text. It wasn't a compressed image. It didn't match any known audio codec or data format from the era. He'd run frequency analysis, autocorrelation, entropy measurements. The data had characteristics of an encoded transmission—something deliberately structured to carry information—but the encoding scheme was nothing he recognized.

Elijah saved his work, pushed back from his desk, and walked to the kitchen. The apartment was a controlled disaster of vintage electronics. A half-restored Defender cabinet stood against one wall. Shelves held rows of EPROMs in anti-static tubes, labeled in his small, precise handwriting. A pegboard displayed his tools: logic analyzers, EPROM programmers, soldering stations. The kitchen counter was the one surface he kept clear, a concession to basic human functioning.

He poured himself a glass of water and stood at the window. South Congress was mostly dark now, the food trucks shuttered, the boutiques locked. A few figures moved along the sidewalk—night owls heading home, or heading out. Austin never fully slept.

He thought about the data. The thing that bothered him most wasn't its existence—ROM dumps turned up surprises all the time. It was the complexity. This wasn't an Easter egg. It wasn't a developer's signature or a hidden message. The mathematical structure was too dense, too layered, too deliberate. Someone had put serious effort into encoding this, using techniques that seemed far beyond what a small game studio in 1983 would have had reason or ability to employ.

He went back to his desk and opened a browser. The forum he wanted was RetroCircuit, a private community of about two hundred serious vintage computing researchers. No casual collectors here—these were people who could read Z80 assembly in their sleep, who spent weekends decapping chips under microscopes to document undiscovered silicon revisions.

Elijah composed a new post.

---

**Subject: Anomalous data pattern in Starfield Corridor (Parallax Interactive, 1983)**

I've been doing a full disassembly of the Starfield Corridor ROM (Z80/YM2149, PCB rev 2) and found something I can't explain. There's a block of structured data embedded in the audio ROM region that the sound driver references but never plays. Approximately 32KB of highly organized, non-random data with nested repeating structures.

This isn't orphaned code or development artifacts. The entropy profile and autocorrelation suggest a deliberately encoded signal. I've attached the extracted data block, spectrum analysis, and my preliminary notes.

Has anyone encountered anything similar in other ROMs from this era? Specifically interested in Parallax Interactive titles, but open to anything.

—Elijah

---

He attached his files, reviewed the post twice, and hit submit. Then he closed his laptop and went to bed, though sleep came slowly. His mind kept turning over the waveform, trying to find the edge of the pattern, the seam where it might unravel into something recognizable.

He dreamed of static.

The responses started coming in the next morning. Most were from regulars offering standard explanations—development leftovers, copy protection schemes, ROM banking artifacts. Elijah had considered and ruled out all of these, and he said so politely in his replies. A few people downloaded the data and ran their own analyses, confirming that the structure was genuinely unusual.

By the second day, the thread had fifty-three replies and had been pinned by the forum moderators. Someone had cross-posted a summary to a larger retro computing subreddit, which Elijah wasn't thrilled about—he preferred the controlled environment of RetroCircuit—but the genie was out of the bottle.

On the third day, he received a private message.

---

**From: NCorrea_Investigates**

Mr. Reynolds,

My name is Nadia Correa. I'm a journalist based in Washington, D.C., currently working on a long-form piece about unexplained anomalies in Cold War-era technology. Your Starfield Corridor finding caught my attention because it matches a pattern I've been tracking for the past eight months.

I believe the signal you found isn't unique to that game. I have evidence suggesting similar embedded data exists in at least three other arcade titles from 1982-1984, all from different developers.

I'd like to talk. I can be in Austin by Thursday.

Nadia Correa
Freelance Investigative Journalist
Formerly: The Washington Standard, Wired

---

Elijah read the message three times. He was naturally cautious—years of working in software development had taught him to be skeptical of unsolicited contacts—but the specificity of her claim intrigued him. If the same signal existed in other games from the same period, that changed everything. One anomaly was a curiosity. Multiple anomalies across different developers suggested something systematic.

He replied: *Thursday works. I'll send you my address.*

Then he went back to the data and kept working.

Nadia Correa arrived at his apartment at 10 AM on Thursday, carrying a leather messenger bag and looking like she hadn't slept much on the flight. She was in her mid-thirties, with dark hair pulled back in a practical knot and sharp brown eyes that swept his apartment with undisguised curiosity.

"You actually live with all this?" she said, taking in the arcade cabinets, the shelves of components, the server rack humming in the corner.

"It's a lifestyle," Elijah said. "Coffee?"

"God, yes."

He made pour-overs while she unpacked her laptop on his cleared kitchen counter. She worked efficiently, pulling up files and arranging them in a sequence that suggested she'd rehearsed this presentation.

"Here's what I have," she said when he handed her a mug. "Eight months ago, I started researching a story about anomalous signals detected by NATO monitoring stations during the Cold War. Most of it was mundane—atmospheric interference misidentified as Soviet transmissions, that kind of thing. But I found references in declassified NSA documents to a specific signal detected in late 1982. The documents called it 'Pattern Sigma.' The signal was received at multiple monitoring stations simultaneously, which should have been impossible given the frequencies involved. The NSA couldn't identify its origin or decode its content. They classified it and moved on."

"Okay," Elijah said, sitting across from her. "What does that have to do with arcade games?"

"I'm getting there." She pulled up a spectrogram. "I obtained a fragment of the Pattern Sigma recording from a source I can't name. When I analyzed it, I found the same nested mathematical structure you described in your forum post. Identical autocorrelation profile. Identical entropy distribution."

Elijah felt something cold move through his chest. "You're saying the signal in the game matches a classified government recording."

"Yes. But here's what really got my attention." She opened another file. "After I published a preliminary piece about Pattern Sigma—no classified details, just the broad strokes—I was contacted by a collector in Japan who'd found similar data in a 1982 shoot-'em-up called Void Runner. Different developer. Different hardware platform. Same underlying signal."

"Same signal," Elijah repeated.

"Same signal. Since then, I've identified two more games with the same embedded data. Galaxy Drift from 1983, and Photon Lance from 1984. All different studios. All different hardware. All containing the same core pattern, encoded differently to fit each platform's audio architecture but mathematically identical underneath."

Elijah was quiet for a long moment. He stared at the spectrogram on her screen, recognizing the shapes he'd spent the past week studying.

"That's four games," he said. "Plus mine makes five. You said in your message you had evidence of at least three."

"I've since found a sixth. A game called Nebula Circuit, 1982. The earliest one yet."

Six games. Six different developers. The same signal.

"Someone distributed this," Elijah said slowly. "Someone gave the same data to six different game studios and had them encode it into their products."

"That's one possibility," Nadia said. "But I've interviewed former employees at three of the studios. None of them remember receiving any external data or being asked to embed anything unusual. Two of the studios were in Japan, two in the United States, one in the UK, and one in West Germany. They had no business relationships with each other."

"Then how—"

"That's the question. How did the same signal end up in six games made by six unconnected studios across four countries?"

Elijah looked at the waveform on his own monitor. The nested patterns, the mathematical precision, the layers within layers.

"We need to find the developers," he said. "The actual people who wrote the audio code for these games. Someone had to put this data in. Code doesn't write itself."

"Agreed. Which is why I've been trying to track them down. Most of the studios are long gone. The developers are scattered, retired, some deceased. But there's one I haven't been able to reach." She paused. "Jack Kramer. The lead designer of Starfield Corridor. Your game."

Elijah knew the name. It appeared in the ROM's header data and in a few old magazine interviews from the period. Jack Kramer had been the founder and sole programmer of Parallax Interactive, a one-man operation working out of a rented office in Albuquerque, New Mexico.

"I've found an address for him," Nadia continued. "He's still in New Mexico. Rural area outside Santa Fe. But he doesn't answer letters, doesn't have a listed phone number, and his property has a gate with a no-trespassing sign."

"So he's a recluse."

"Very much so. I was hoping you might have better luck reaching him through the retro gaming community. Collectors sometimes have connections to old developers."

Elijah thought about it. He didn't know Jack Kramer personally, but he knew people who might. The vintage gaming world was small enough that connections existed if you knew where to look.

"I can try," he said. "But first, let's publish what we have. A joint analysis of the six games, the signal comparison, the whole picture. Put it on RetroCircuit and let the community help."

Nadia hesitated. "I'm used to controlling the release of information. Journalist instinct."

"I understand. But this is bigger than a story. If there's a pattern here, we need more eyes on it. The people on RetroCircuit are the best in the world at analyzing vintage code. They might find things we'd miss."

She considered this, then nodded. "Okay. But I want to write the contextual framing. The connection to Pattern Sigma, the cross-studio analysis. You handle the technical deep dive."

"Deal."

They spent the rest of the day working side by side, Elijah at his desk and Nadia at the kitchen counter. By evening, they had a comprehensive document: technical analysis of all six ROM dumps, spectral comparisons, mathematical proofs of the signal's consistency across platforms, and Nadia's narrative connecting the games to the 1982 NSA detection.

They published it at 9 PM Central Time.

By midnight, it was the most-discussed post in RetroCircuit's fifteen-year history.

By the following morning, Elijah's inbox contained 347 messages.

One of them was from Jack Kramer.