As I envision it, the idea for a film converter screams for the use of an Arduino or Raspberry Pi to automate the process, some mechanical engineering to figure out film transport, 3D printing to create the "sprockets" that precisely advance the film, and image capture capabilities. My idea is to create what is called a telecine converter that would automate the digital capture of each frame of the film as an image, and then allow those images to be strung together at the proper frame rate to create a digital copy of the film.
A little background:
I have a box full of 8 mm and Super 8 movie film that my parents shot of my brother and sisters when we were kids. Though I do have a projector, fiddling with it is always tricky, and you end up watching 3-5 minutes of film before you have to rewind it, and load the next roll. Needless to say, I never pull it out to watch the film.
There are commercially available systems, but the decent ones are prohibitively expensive, and the affordable ones are too low resolution to even bother spending your time doing the conversion. There are "devices" that allow you to project the film into one side, and then use a digital camera/video camera to capture the image from the other side in real time, but these systems suffer from incongruous display and capture rates, resulting in a "beat frequency" or "flicker" in the finished capture. They also suffer from a hot spot in the center of the film image (and relatively poor illumination in the corners of the image) caused by the filament in the lamp on the projector. And projectors can't be stopped to capture a single frame without danger of melting the film with the heat of the lamp.
Commercial sites (WalMart, Costco, and others) will also transfer film for you, but they are expensive, usually in the $10 range for a 50' (3-5 minute) reel captured at Standard Def, and more for HD. If you have a sizable number of rolls, this quickly adds up, if you are even willing to let precious old home movies out of your possession for 3-5 weeks to a lab located who knows where.
Of course digitization would allow for color correction, scratch reduction, and various other types of repairs to the digitally captured copy of the original film. Also, though I don't have any film with sound, but I understand there are 8 mm films out there that do have an audio track, and that could also be captured and synchronized with the film. Converting the images to a movie, and post-processing (color correction, scratch reduction, etc.) could be an interested second phase to the project. Or maybe there is an open source or Linux based solution to do it on the fly at the time of capture?
I imagine an Arduino (or preferably a Raspberry Pi) using a stepper motor to advance the film one frame, and trigger a digital camera (or phone) to capture an image of the frame and save it to an SD card or computer (via USB, WiFi or Bluetooth). Most cameras and phones today have sensors that are 10 MP or bigger, which means the resulting film created from these captures should be 4K or better (I think). Of course we're limited to the quality of the original film, but why capture it at less than 4K?!?! Another option would be to use the Raspberry Pi camera, which has a resolution of 5MP, I believe, enough for 1080P but not 4K. I was thinking something like a Canon Powershot with CHDK installed would do nicely. Of course whichever capture solution is used, the optics have to be worked out to enlarge the film frame and focus the camera (maybe one of those clip on smart phone lenses... or an old slide projector lens used in reverse???). I also figure the frame could be evenly illuminated with an LED or LED array to provide a cool, even light source.
There are homebrew hacks described on the web where projectors have been retrofitted with LEDS, instrumentation to advance a single frame, and microswitches to trigger a camera, but it seems a complete solution using modern maker tools would be very much in Ben's wheelhouse!
