The idea for this project was to buy an old analog super8 camera, and have some fun experimenting with it, filming with both new stock and expired stock (more about that in a minute). So in the spring of 2016 I started the hunt for a suitable camera. I ended up landing a great 1968 Minolta super8 camera with just the right tactile qualities for my tastes. The camera set me back a cool $20 from ebay. Not bad! 

The camera is just right for that kind of old-school leather and metal feel. Its devilishly simple, with two film speeds of 18fps and 32fps, as well as a "single frame" option. Manual zoom and focus, and not much else.

With the camera secured, I next set out to purchase a few different types of film to shoot with. The thing about super8, and there really aren't too many ways around it these days, is that it is EXPENSIVE. I mean, similar to analog 35mm photography, the up-front investment into the camera and film still put you pretty far in the black compared to buying a digital DSLR or similar. But with super8, the price and hassle of buying, developing, and digitizing the film is much more noticeable. So one of the key aspects of this project was to find some sneaky solutions around this hurdle. That brings me to the film itself. Buying a 50ft roll of new super8 stock sets you back around $25, which isn't in itself horrible, but combined with the processing fee of another $25 per roll, and then Telecine professional digitizing...It can add up fast. So for this project I went with 2 rolls of new stock, and on top of that one roll of "old stock". 

Old stock super8 film is basically old expired Kodachrome film that was never used new. Normally the film has expired between 15 and 50 years ago. Its a bit of a niche market for this stuff, and obviously there isn't an infinite supply, but normally you can pick a roll up for between $5-$15. Why so cheap? Two reasons: First its expired, and the chemical treatment on the film can be degraded and less that stellar quality. The second reason is the kicker. Kodak decided in 2010 to stop producing the developer chemicals for Kodachrome super8, so anyone with the film is SOL when it comes to doing anything with the exposed film. That is...unless you think a little bit outside the box, and can re-awaken your inner mad scientist. This old stock super8 requires us to get a little 'Breaking Bad' to develop it. Just my kind of thing!

Film bought, I decided to lug the camera around with me through Eastern Europe, filming the two new stock rolls in Germany, Poland, and Ukraine. The experimental old stock was exposed through Ukraine, France, and Spain. 

Before the film has even come out, I must say that shooting super8 is a real treat! For one, the mechanical, analog feel of the camera is just lovely. The weight of the camera, and the sound are so pleasing. And there is something...nostalgic feeling about looking through that old viewfinder, with its glass split focusing, and bold black markings. The other unexpected joy of shooting with the camera is the incredible interest and curiosity of strangers when they see me shooting film. I had a lovely conversation with a young Ukrainian about how the process works, and let her shoot for a little bit with the camera. Seeing her face light up when she played with the dials and heard the sound of the mechanism whirring makes it all worth it in my mind! I'd like to think that its the first time a lot of (young) people have seen analog film in action, and that is a nice feeling, to teach a little bit about how the old days of film worked.

After shooting all three film stocks, it was time to focus on home-developing the Kodachrome. The first step was to set up a temporary dark-room in my workshop, and get together the chemicals and materials needed for the development. The make up of the developer is, at fist glance, completely crazy. It is a concoction of instant coffee, washing soda crystals, vitamin C powder, and black-and-white photo fixer. Personally I was extremely suspicious, but many hours spent on film forums, blogs, and youtube videos gave me at least a general courage to try it out. 

The general instructions that I followed for the mix was sourced from Daniel Berrangé from the fstop forums. His writeup can be found in the link below.

I wont get into too much detail about the intricacies of the chemicals, but safe to say it was a very touchy thing to mix and use. Having little faith in my own abilities, I called on the help of a friend to aide with timekeeping, measuring, and for an extra brain and set of hands. We decided to test out 6 inch strips of the film in the chemicals, to do some timing and temperature tests. First test was bunk, but we tweaked some things here and there, and got some promising results the second time through. 

Next came the big one. Developing the whole film. My heart was beating fast, and my hands shook. Its remarkably stressful to realize that everything rests on the next 9 minutes, whether I will see the film of my travels or not. 

Well...

Unfortunately the results were not ideal. I think around 60% of the film came out well, with the rest being blank or splotchy. The parts that came out looked lovely, but the film was overall not well developed. I have a suspicion that I messed up the temperature of the developer. I think it may have been too cold for a successful development with the time that I had it in. Below you can see some of the areas of the film that came out well. 

Below is the much better results from the two new-stock films that I developed afterwards! I am super happy with them, and the consistency with new stock can really be seen in the films

From these great results, I wanted to further experiment with some new AI that can colour films.

I used Jason Antic’s ‘DeOldify’ algorithm through a virtual machine with google colab to colourize my films. The link below takes you to the github repo for the DeOldify algorithm

Below are the results!

My future plans for this project are to see what other algorithms I might be able to use for interpolation and upscaling of the films. I recently watched Peter Jackson’s great film on the first world war where they did an incredible job bringing old film to life using these algorithms. There is also Denis Shiryaev’s work on the same subject which has inspired my continued experimentation.