They're seventeen again for a few minutes on screen
In a high school storage room in Pittsburgh, decades of football games were quietly dissolving into vinegar and shadow — until a maker named David Stein decided that memory was worth engineering for. Faced with the impossible economics of professional digitization, he married a 1928 film projector to a Raspberry Pi 5 and coaxed dying reels back into motion, cutting an eighteen-hour process down to four. His act of preservation is also a quiet argument: that the tools of community memory need not belong only to institutions with budgets, but to anyone willing to build the machine that time forgot to provide.
- Vinegar syndrome is already consuming the forgotten reels — a chemical clock that waits for no budget cycle or institutional decision.
- At $200 per reel, professional digitization made saving the full collection mathematically impossible, forcing a choice between family history and financial reality.
- Stein engineered his way around the cost barrier, retrofitting a 98-year-old projector with a Raspberry Pi 5, a microscope lens, and an Arduino-controlled transport system that runs film at eight frames per second.
- Upgrading from Raspberry Pi 4 to Raspberry Pi 5 collapsed the scanning time from eighteen hours to four — the difference between a project that stalls and one that scales.
- The footage now plays: fathers watch themselves at seventeen, strangers become friends over shared games, and a storage room's worth of ghosts are given back their motion.
David Stein found them in a storage room at his father's high school — reels of 16mm film, decades old, already rotting. Among them were three games his father had played as a seventeen-year-old. The rest of the collection, spanning years and teams, was deteriorating just as fast. Vinegar syndrome, the chemical breakdown of acetate film, was already well underway.
Stein had managed to digitize some footage before — his grandfather's Super 8 reels, and games donated to the Heinz History Center in Pittsburgh. But these storage room reels had been forgotten. Professional digitization ran $200 per reel, making it impossible to save them all. So he built a machine instead.
The heart of his solution was a 1928 film projector, chosen precisely because its age made it gentle — designed for fragile stock, not speed. He modified it for accuracy, mounted a microscope lens at the film gate, and connected it to a Raspberry Pi 5 paired with a High Quality Camera module and an Arduino managing the transport reels. His earlier Raspberry Pi 4 build captured one frame at a time across eighteen grueling hours per reel. The Pi 5 changed everything: film now ran continuously at eight frames per second, with a soft ramp-up to protect the stock, while the camera recorded in sixty-second video segments. A Python and FFmpeg pipeline then extracted the sharpest frame from each group. Four hours per reel.
What came out of those four hours was something closer to resurrection. The players on screen are in their seventies now, but the footage returns them to seventeen — running plays under stadium lights long since demolished. Stein's father watched his own games for the first time in decades. One older reel led to an unexpected friendship with a man from that era who now calls Stein regularly, just to talk about what it felt like to be young and fast and part of something.
The project points beyond one family's archive. It demonstrates that a vintage projector, a modern single-board computer, and open-source software can accomplish what commercial services charge hundreds of dollars to do — and that other storage rooms, other forgotten collections, other families are still waiting for someone to build the machine that saves them.
David Stein was looking for ghosts in a storage room at his father's high school. What he found were reels of 16mm film, decades old, beginning to rot. Among the decay were three games his father had played in, back when he was seventeen and wore a football uniform. The rest of the collection—dozens of reels from other years, other teams—was deteriorating just as fast. Vinegar syndrome, the chemical breakdown of acetate film base, was already at work.
Stein had already paid to have some of his father's games digitized. His grandfather had shot Super 8 footage that got scanned. The high school itself had donated its collection to the Heinz History Center in Pittsburgh, and Stein had those transferred too. But the reels in the storage room had been forgotten, left to decay in the dark. At two hundred dollars per reel to have them professionally digitized, the math was impossible. He couldn't afford to save them all. So he decided to build a machine that could.
The centerpiece of his solution was a 1928 film projector. Ninety-eight years old, it had been designed to move fragile film gently through its gate without tearing it—a mechanical precision that modern equipment, built for speed, had largely abandoned. Stein modified the projector, made it more accurate, and taught it to talk to a Raspberry Pi camera. He mounted a microscope lens at the film gate, aimed at the frames as they passed through. The brain of the whole apparatus was a Raspberry Pi 5, paired with a High Quality Camera module and an Arduino controller that managed the transport and take-up reels.
His first version, built on a Raspberry Pi 4, captured one image per frame. A four-hundred-foot reel held sixteen thousand frames. The process took more than eighteen hours. When he upgraded to the Raspberry Pi 5, everything changed. The new processor let him run the film continuously instead of stopping for each shot. The Arduino and two stepper drivers pulled the film at eight frames per second, with a soft ten-second ramp at the start to keep the system from jerking and tearing the stock. The camera recorded continuous video at thirty frames per second in sixty-second segments, using a five-millisecond exposure to freeze each frame mid-pulldown. Afterwards, FFmpeg and a Python pipeline extracted the sharpest frame from each group. The time per reel dropped to four hours.
What emerged from those four hours was time travel of a kind. The players in these films are in their seventies now. But on screen, they're seventeen again, running plays under stadium lights that have long since been torn down. Stein's father watched his own games for the first time in decades. He shared them with his son. One of the older reels led to an unexpected friendship—a man from that era who now calls Stein periodically to talk about the games, to remember what it felt like to be young and fast and part of something.
The project matters beyond one family's archive. It shows what's possible when someone refuses to accept that deteriorating media must be lost to time and chemistry. A vintage projector, a modern single-board computer, open-source software, and patience can do what commercial services charge hundreds of dollars to accomplish. The reels in that storage room are being saved. And somewhere, there are other storage rooms, other forgotten collections, other families waiting to see their own ghosts move again.
Citações Notáveis
With this project, it's as though they're seventeen again for a few minutes— David Stein, describing the experience of watching digitized footage of his father and teammates
A Conversa do Hearth Outra perspectiva sobre a história
Why did you choose a 1928 projector instead of building something from scratch?
Because it was already perfect for the job. A projector from that era was engineered to handle fragile film without tearing it. Modern equipment is built for speed and volume, not gentleness. I needed gentleness.
The upgrade from Raspberry Pi 4 to 5 cut your scanning time in half. What changed?
Processing power. The Pi 4 had to stop and capture an image, then move to the next frame. The Pi 5 could let the film run continuously while the camera recorded video. It's the difference between taking snapshots and filming a movie, then extracting the best frame from each moment.
Two hundred dollars per reel is expensive. How much did your scanner cost to build?
I didn't track it precisely, but the projector was old and cheap. The Raspberry Pi, the camera module, the Arduino—those are all affordable. The real cost was time. Months of design and testing. But once it was working, I could save dozens of reels for less than the cost of sending one to a commercial service.
Your father watched his own games for the first time in decades. What did that feel like for him?
He was grateful. But there was something else—a kind of recognition. These weren't just memories anymore. They were real. He could see his teammates, the field, the way he moved. And he could share it with me. That mattered to him.
One of the people in the older reels reached out to you. How did that happen?
I shared the digitized footage. He saw himself on film and called me. Now we talk about the games, about what it was like to play back then. The film became a bridge between past and present. That wasn't something I expected, but it's maybe the most important thing the project did.