Building a frankenstein camera - Part 1: Finding the Shell

Missing the Joy of the Single Shot

While working a lot with cameras and new technology, always being kind of on the edge of the newest tech like Gaussian Splatting or Artificial Intelligence - I've been missing something for a long time. Something specific: the fun of making new images. You know, going out and taking out my camera not because I want to shoot the perfect image. Not because I need 1000 images to make a good photogrammetry scan or Gaussian splat. I miss the feeling of taking out the camera and just having fun with it. No editing, just live the moment and get out whatever you get.

Don't get me wrong, I have a lot of fun doing all that other stuff. But just taking out your camera without thinking about the tech behind it - just shooting - is something I've missed for a long time. Not thinking about whether the image was good or bad. Not checking the colors, not looking back to see if the brightness was correct. Just taking the shot and having fun doing so. I'm a person who loves going deep, who loves understanding every technique possible, who loves knowing every little piece I can tweak in cameras or code - but that never really lets me be free from thinking.

One day I was looking back at an old camera from my grandfather, a Canon A-1 film camera, and that brought up an idea. Limiting myself to get the most fun out of shooting. What was great about film cameras? No screen to check if the shot you took was good. Not that many options - choose your film, your lens, and just shoot. Only later did you see the results. Shoot without thinking about which settings you could tweak. Just shoot.

And what I really love about film is all the imperfections. You don't need to tweak images later in Lightroom to get a feeling - every image is different. Different noise patterns, different light breaks, no single image feels like another. Its own unique style. So I started thinking about combining old film cameras with new technology, while still limiting myself as much as possible to stay focused in the moment, without thinking about the science behind it.

What's Out There: Scouting the Market

With that in mind, I started researching - what's on the market, what options do I have, what am I actually looking for. Ultimately I wanted a camera that's as cheap as possible, one I don't need to worry about. A camera that can break and I'll just shrug and say "that's okay, it happens" - not one I have to babysit because it's worth as much as a car. My ideal range for something like that is around 100-200 €.

That alone limits the market a lot. Screenless, limited cameras that try to do something similar - like the Leica M-D - go way past that price range. Even older point-and-shoot cameras with that kind of vibe have mostly priced themselves out too, since the market for them has grown a lot in recent years. So most cameras that would fit "out of the box" are already out. (Which, honestly, I don't mind - I really love building stuff on my own.)

Looking further, I found a handful of crowdfunded "retro digital" cameras that all try to capture roughly the same idea: 90's point-and-shoot vibes, two or three filters, and just shoot.

I was really close to buying one of these, either the Camp Snap Pro/2 or the Rewindpix. But after digging through reviews, I felt none of them were the perfect fit. Cheap 'plastic' feel in hand, noticeable shutter delay, software limitations baked into the design - that's not what I was looking for. I really love a device that feels like it's worth its price. Plastic shells are very "2000s," sure, but a camera that tries to imitate a 1960s Leica with a metal design and then feels like plastic in your hand just doesn't sit right with me. I get that lighter materials make a camera easier to carry everywhere, but it's just not what I wanted.

Not being able to swap lenses was another dealbreaker. Yes, I wanted to be limited - but limited in the most flexible way possible. Not limited for good, but able to switch things around when something breaks, or if one day I just want to shoot wide-angle only. Still, these cameras have some interesting concepts I'll come back to later.

At that point I drifted further toward old film cameras and found I Am Back Roll, another crowdfunded project - a digital "film roll" you load into your own 35mm camera. Sounds like the perfect fit, right? The shell of an old camera, but digital inside. Then I saw the price tag: around 500 €. Out of view, again, very fast.

Building My Own Camera

With every option on the market checked off, none of them fully matched what I wanted. So - why not build it myself? I started looking into it more seriously. The basic idea:

• Raspberry Pi Zero 2 W
• Raspberry Pi 12 MP HQ Camera Module
• An old camera, used purely as the shell

Plan A: The Raspberry Pi HQ Module

Shooting on a Raspberry Pi with the HQ module has some real advantages. I can add my own lenses and filters - not just color filters, but noise and whatever else I want. (Side note: at the end of last year I was already working on a small Python project to imitate old film as closely as possible.) A mini computer in the shell means I can write my own logic behind everything, fully flexible.

The HQ module also already has a C-mount built in, so I could strip the old camera down, mount the HQ module directly to the lens mount via an adapter ring (C-mount to 39mm), and get the correct sensor-to-lens distance for sharp images.

But there's one big tradeoff: a Raspberry Pi is a mini computer, not an image processing unit. Taking a photo, applying filters, and more would take seconds rather than a split second. And the bigger tradeoff - the sensor is small enough to give a crop factor of 5. A 50mm lens suddenly becomes a 250mm tele, which isn't great for a daily camera I want to use everywhere. Even a 25mm becomes 125mm - way too tight. And since I really want to use my vintage lenses and filters, that ruled the HQ module out. Such a shame, it would've been a clean setup: strip the old camera, mount everything, done.

Choosing the Shell: Why an Old Soviet Rangefinder

You'd think removing the insides of a film camera is easy - I mean, what's even in there? A place for the film and a piece of plastic that winds it. Turns out, not so simple. Ideally I wanted to remove a few screws and pull everything out without extra tools. Browsing eBay and digging through old camera listings, I realized only a handful of cameras actually fit that profile - in most cameras, the internals are directly integrated into the shell, meaning I'd have to saw stuff out.

That's where Leica's bottom-loading principle comes in. Starting in 1925, Leica cameras loaded film from the bottom - which meant you could pull the whole film mechanism out in one go, leaving you with an empty metal shell. Just a housing. And in 1934, the Soviet Union copied that exact mechanism for their own cameras. (Contrary to popular belief, this wasn't because Leica lost its patents after the war - the Soviets just decided to copy it outright.) That's how FED started building what was essentially a clone of the Leica II, and Zorki followed with the same principle later.

So the best case for me: a camera I can gut without much effort. An alternative would've been something like the Olympus Trip 35, where all the gear sits inside the lens - but that brings back the lens-swapping limitation. Ultimately I needed an old Leica from before 1954 (the system changed with the Leica M3) - but an original Leica is anything but cheap, and well outside my ~50 € budget for a shell. Luckily, since the Soviets copied the mechanism, I had cheaper options:

The FED 2 was the clear winner - being able to open it from the back means easy access to mount the new tech, without fighting the shell. I found one on eBay for 25 €.

Frankenstein: Donor Camera + Raspberry Pi

With the shell decided, the next question was the main logic. Since the Raspberry Pi HQ module is off the table because of the crop factor, the only option left is going full Frankenstein: using a newer digital camera as a donor and fitting it inside the FED 2.

You might be thinking - wait, doesn't a modern sensor inside an old shell just defeat the whole "vibe" of the project? Not quite. My idea: pull images off the donor camera's SD card via a Raspberry Pi, then apply filters afterward rather than live. Space is the constraint - a Zero 2 W is pretty limited in processing power, and processing a 16 Megapixel image with procedural noise on top would take seconds, not be instant or reliable in real time.

So instead of live processing, I'm moving it off-device entirely - which works out nicely, since the camera has no screen anyway. Why process on the go if nobody's looking at a screen in the moment? Instead, I'll write an app that connects to the camera and has a single "develop" button - a bit like the darkroom, but digital. One press, and it processes the whole batch. No live preview to fuss over, but also no manual editing of every photo afterward in Lightroom.

Why Still Use a Raspberry Pi at All?

Sounds like the Pi might be unnecessary at this point, right? Not quite - I still need a way to adjust ISO and shutter speed without a screen. The plan: use the Raspberry Pi together with the PTP protocol to remotely control the donor camera's settings. Physical rotary switches and buttons connect to the Zero 2 W, which then pushes those settings to the camera whenever they change.

That requirement narrows the donor camera search to one brand with reliable PTP support that still fits the housing and budget: Sony. Early-2010s Canon and Samsung compacts had limited PTP support, and anything newer is outside budget. That leaves the Sony NEX series:

• Sony NEX-3 / NEX-5 - around 100 € used, good PTP support, fits the budget
• Sony A5000 - would be ideal, but priced out of range

The Next Steps

I'll post updates as I go. Up next:

• Strip all the internals out of the FED 2 and measure how much space is actually left
• Mount the donor camera and check remaining space for batteries - one pack for the camera, ideally a second for the Pi (plus a 7V → 5V step-down)
• Install the Pi Zero 2 W as the controller, including Wi-Fi file transfer

One more idea for the controls: rotary switches (not stepless) for shutter speed and ISO, a shutter button, a power button, and - borrowing from the new point-and-shoot cameras systems like the camp snap pro - a physical filter switch. Before shooting, I want to dial in a "film preset" (1-3, kept deliberately limited). Whichever preset is selected gets written into the image metadata, so the processing app automatically applies it later - like the image was already developed in the darkroom.

Ending this part of the series with one more thought: my goal is to limit myself in the moment, to not think about anything and just shoot and have fun - but still be free to tweak whatever I want, later.