Building your own Super 8 film scanner

Nowadays, you don’t have to be an engineer or electrician to make your own tools. Thanks to the technology available today and the wealth of information available, many problems can be solved. Building your own equipment is particularly appealing. Not only does it allow you to customise it to your own needs, but it also creates a deeper connection and appreciation for what you create. In this article, I want to show that everyone has the opportunity to make their own tools. To do this, I will present a project that is very close to my heart. I will explain a problem and the steps to solve it, with the confidence that many people can find approaches to overcome their own creative challenges.

by Robert Krupa

The construction shown here is a DIY film scanner for Super 8 films. This device makes it possible to copy analogue Super 8 films into a digital file format in very good quality. Each film frame is scanned individually to avoid image interference such as ghosting or flickering. A conventional camera is used for digitisation, giving the user a high degree of control over the image quality.

• Links!
  • My website: RO-KR.net
  • LinkedIn: linkedin.com/in/robert-krupa-3d
  • Website of the Hamm-Lippstadt University of Applied Sciences: www.hshl.de
    (Computer Visualistics and Design programme, contact person Prof. Rainer Baum)
  • Arduino: www.arduino.cc
  • Electronic camera control: www.doc-diy.net/photo/remote_pinout
  • Adapter for microscope lenses: bit.ly/youtube_makrolinse
    

Where did the idea come from?

It all started when my aunt gave me a large collection of old Super 8 films. First I got myself a film projector and learnt how to handle the films properly. After a short time, I had successfully played the films. Of course, I wanted to show them to my aunt, so I filmed the projected image from the screen. But I soon realised that there was a problem. The Super 8 film runs at 18 frames per second, while my mobile phone films at 30 frames per second. Due to these different frame rates, the image flickered a lot, double exposures and ghosting artefacts appeared. An initial solution was to equalise the two frame rates. The film projector had the option to increase the film speed to 24 frames per second, which was also the lowest frame rate my DSLR could record. I therefore filmed the projected image again at this adjusted speed and later adjusted the speed back to 18 FPS on the PC. This led to a considerable increase in quality, as the flickering was now greatly reduced. However, there were still ghosting effects and slight fluctuations in brightness. The reason for this was obvious: the film projector was an old device that was relatively simple in design and, above all, mechanically operated. There was no digital computer to control the timing down to the nanosecond. This did not play a major role in the viewing of the films
Role. As long as the image reproduction was around 18 frames per second, minor timing differences were not particularly noticeable. For the digitisation of the films, however, this was a decisive factor.

However, there were other problems such as vignetting and blurring. So I looked for other options to digitise the films. One option was to send the films to a digitisation service, although there is always the risk that they could be lost. Also, film digitisation is expensive. With the amount of films I have, digitisation by a service was not financially viable for me. Another option was to buy or rent a film scanner. There are film scanners for home use that allow users to digitise their films themselves. However, there are problems here too. There are examples of the quality of such scanners on YouTube. They have a low dynamic range, fluctuations in brightness and oversharpening artefacts. In addition, the colour reproduction is very poor and colour fringing occurs. I therefore wanted to find a cost-effective solution that would give the user full control over the digitisation process and deliver good quality digitised films.

Structure of the device

The design of the device is very similar to a portable analogue tape recorder. A stepper motor rotates a reel that moves the film one frame at a time. The film is fed through a device that illuminates it from behind with an LED so that the respective film image can be photographed from the front with a camera. Between each rotation of the reel, the film stops for a moment and an electrical signal is sent to the camera to trigger it. An Arduino microcontroller is used to control the electronic components.

For automatic film unwinding, the machine has two turntables on which the film spool and an empty spool are placed. These are each driven by a belt with a motor. The films lie on a flat disc that rests loosely on the turntable. This keeps the film taut and reduces the risk of tearing due to excessive tension.

The film gate

The film is guided through a 1 mm wide gap on a 3D-printed device, the “film gate”. There is an area where the film slides freely through the gap, i.e. without touching the device. At this point, an LED with a satinised plastic is mounted behind the film. This illuminates the film with even light so that the film image can be optimally photographed. Further along the film guide, a light barrier is mounted exactly at the height of the film’s perforation hole. This ensures that the film remains in position. The camera with the macro lens is placed in front of the film gate so that the film image is in focus.

Operating the device

The device is equipped with a control panel that is connected via a cable connection. The control panel consists of 9 push buttons that can be used to perform various operations such as forward and rewind, manual release of the camera or precise positioning of the film image. In addition, user inputs are shown on a small LCD display to give the user full control over the scanner.
To connect the control unit to the Arduino, it was important to reduce the number of tracks. As the control unit has 9 pushbuttons, at least 10 tracks would otherwise be necessary. Instead, the buttons are connected in series with electrical resistors. Depending on which button is pressed, a certain voltage drop occurs, which can be measured via the analogue inputs of the Arduino. In the scanner software, an if query can be used to define a specific voltage range in which the respective commands are executed. See the pictures on the next page.

Controlling the camera

In order to capture the individual images, it was important that the camera was triggered automatically at the exact moment when the film image was in the correct position and the film reels were stationary. I bought a remote shutter release for this purpose. My first idea was to bridge the mechanical push button with a transistor to send the release signal to the camera. However, when I unscrewed the remote shutter release, I realised that only 3 contacts are short-circuited inside, which means that the remote shutter release is ultimately not necessary.

From this I was able to derive the circuit diagram above: Depending on which switch is pressed here, either the autofocus or the shutter is triggered. The majority of all camera models work according to this principle, which makes it easy to use different cameras. As the lens I am using does not have autofocus, only the switch for the shutter is important here. Instead of a mechanical pressure switch to close the circuit, a switch that can be controlled via an electrical signal had to be used at this point. In my opinion, the best solution for this is to use an optocoupler. This works in a similar way to a transistor, with the advantage that, unlike a transistor, the respective circuits are separated from each other. This means that there is no risk of damaging the camera by applying voltage.

In my case, I constantly connected the focus connection to GND. In hindsight, it would have been better to control the connection with an optocoupler as well. Otherwise, the camera can no longer be operated, as the permanent connection of the focus contact to GND simulates a permanently half-pressed shutter release and thus blocks the operation of the camera.

Correct colour reproduction

To ensure that the colours of the film are reproduced as faithfully as possible, a neutral white LED with a high colour rendering index (CRI) of over 90 was used. This means that the colour information is reproduced with a very broad colour spectrum during digitisation. In addition, the white balance of the camera is set to the Kelvin value of the LED in order to avoid unwanted colour casts.

Using the scanner

After the film has been fed through the film gate and attached to the take-up reel, a camera with a macro lens is positioned on a tripod so that the individual film image is focussed and can be seen with an overscan. The scanner has a 2.5 mm jack connection at the front, via which the camera receives the trigger signal. The camera is then connected to the remote trigger input and the jack socket via a cable.

At this point, the control unit is rewound to the position at which the first image in front of the
Image appears in front of the LED and on the camera image. The camera must be set to manual mode so that every picture is taken with the same settings. The automated digitisation process is started by pressing the red button on the control panel. Once the film has run through completely, the film can be rewound.
Practically any camera can be used to take the pictures as long as it has an electrical input for a remote shutter release. In my case, I opted for the Sony Alpha 6000 as it is relatively inexpensive and can also take pictures in RAW mode. This allows more freedom in terms of colour and brightness adjustment in post-processing.

I also used the 25mm ultra-macro lens from Laowa, as I think it is perfect for this application. A cheaper alternative, which I have not personally tested, is to use a microscope lens with a 3D printed camera mount. There are instructions for constructing this on YouTube.

Post processing

After the film has been completely scanned by the scanner, the images are now available on one or more SD memory cards. These must now be copied to a computer and combined into a video. To do this, I made a rough histogram adjustment to the RAW images. Essentially, it’s all about ensuring that no clipping occurs and that the image matches the original as closely as possible. The individual images were then exported in lossless TIFF format. The advantage of the images is that the files are already numbered sequentially. This makes it relatively easy to convert them into a video as an image sequence. To simplify further processing, a proxy file was created from the TIFF sequence. This is a file of lower quality and size, which requires less computing power
and is therefore better suited for further processing. As the film image is photographed directly from the film, it is initially inverted, which is why it has to be mirrored back on the computer.

Position errors must also be corrected, which can occur if the film images are not always in exactly the same position during the digitisation process. To correct this, the film’s perforation hole is best suited as a tracking point. Once the image stabilisation has been carried out, the image section can be defined. My personal preference is to fit the film image, including the perforation hole, into a 16:9 image. This means that no image information is lost and it is immediately recognisable that it is a Super 8 film, which gives the film its own charm. Alternatively, part of the edge of the film could also be cropped to create a clean image edge. Finally, the proxy file is exchanged with the TIFF sequence again before a new TIFF sequence or your video file can be rendered.

Digitisation of the soundtrack

The sound is not digitised directly from the scanner, but in a separate and simpler process. The advantage of the soundtrack is that the audio signal can be output in electronic form for speakers or headphones on the projector. On my projector there was an electronic socket for a “DIN” connection. It was easy to find an adapter to 3.5 mm jack. The output audio signal only needs to be recorded with a digital device during film playback. I used a field recorder for this, but it should also be possible to record the signal directly via a line-in input on a computer. The resulting digital file can then be synchronised with the digital image on the computer. The timing often needs to be adjusted. It is important to note that the pitch also changes when the timing is adjusted, i.e. the sound becomes lower when the audio track is extended. This has the same effect as a slow running tape, where the voice sounds lower.
The quality of the sound is relatively good, but there are occasional noises such as crackling and hissing. This is not surprising with old analogue tapes. Such problems can be eliminated very easily with the help of digital processing.

Replica of the scanner

The implementation of the project required knowledge in various fields, including electrical engineering, programming, CAD modelling and mechanical engineering. Nevertheless, everything was kept at a simple level, which meant that a replica should not be too difficult. Through my degree course in Computer Visualistics and Design, I was able to successfully apply some of the skills I had acquired, such as CAD modelling and programming. However, I also had to acquire new knowledge, especially in the field of electrical engineering. In addition, I was able to significantly improve my manual skills, such as soldering, through the project. I received valuable support from the competent assistants at our university. The university has a project workshop that I was able to use. I had access to a stationary drilling machine there and also received valuable tips.

My plan is to provide comprehensive design instructions as well as all the necessary parts and software to enable easy replication. When this is ready, we will put a link on Digitalproduction.com!

One of my aims with this project was to use only components that are open and readily available, rather than, for example, converting a film projector. As this is old technology, not all parts are available for purchase. For this reason, I also made my own components using 3D printing. For example, I made spools with teeth that fit exactly into the perforation of the film to be able to move it. To develop the components, I took measurements from the existing film spools with a caliper gauge and created the 3D files using CAD. The components were also matched to different motors. I used two small N20 DC motors (12V, 300 rpm) and a NEMA 17 stepper motor. In addition, I used a laser diode with a corresponding sensor to detect the perforation hole of the film. However, this function is optional and does not work for all types of film.

The electrical components are controlled using an Arduino microcontroller and a motor control board from Adafruit. In addition, a small LCD display shows how many pictures have already been taken. First of all, holes and cut-outs have to be drilled and filed in a tool case. The case serves as a housing for the device and makes it easy to transport or store the apparatus. The cut-outs are necessary to mount the LCD display, the power connection and the connection sockets for the camera, control unit and interface for the microcontroller. Brackets are also attached to the inside of the case. A Plexiglas disc (or alternatively a metal sheet) in the same format as the case is then produced with holes and cut-outs for the various components of the scanner and mounted on the case with hinges.

The 3D-printed components also have to be assembled, for example mounting devices are attached to the motors and pulleys are attached to the motor axles. The components are then attached to the mounting surface. This is followed by the wiring of the electrical components. For this purpose, breadboards were used to which the components were soldered.
A major advantage of the self-build is that the project can be easily adapted to other requirements.

Here are some ideas or possible improvements

• Remote triggering of the camera via wireless methods: Triggering by cable is not possible with every camera and is also somewhat cumbersome, as the cable connection makes it difficult to position the camera. One solution could be to control the camera wirelessly via Bluetooth, WLAN or infrared signals, provided the camera is suitable for this.
• Wireless control: In the same way, the control panel could also be converted into a remote control. For this purpose, the control panel could be equipped with its own microcontroller, which processes the signals and outputs a signal via an infrared LED. The signal could then be received and executed using an IR sensor on the scanner.
• Recording the images with a smartphone or a WLAN-capable camera: One disadvantage is the large amount of memory required. In my case, a 128 gigabyte memory card was enough for around 5000 images, which requires several memory cards and copying processes for longer films. It would therefore be conceivable to transfer the individual images directly to a cloud or private server during the digitisation process. This can save time that would otherwise have to be spent copying the images and replacing memory cards.
• Adaptation of the scanner to other film formats: Depending on your needs, the scanner could easily be adapted to different film formats. For example, if standard 8mm films are to be digitised, spools could be produced that are adapted to the perforation of the standard 8mm film.
• Development of security mechanisms: At present, the digitisation process requires constant monitoring. There are no mechanisms that automatically switch off the scanner if, for example, a belt breaks and the film is therefore no longer wound up. This could cause the film to get caught elsewhere and be damaged. Sensors could therefore be attached to the winding spools to interrupt the digitisation process if the film is no longer being wound or unwound correctly or if the film has been completely scanned. It should be noted that no damage to the film has occurred in my tests to date.

Advantages and disadvantages

The biggest advantage is the cost saving when using large quantities of film and the high quality results. This also makes a replica of the scanner interesting for smaller video productions. For example, it is quite conceivable that the scanner could be used for the production of wedding videos,
as Super8 film is popular due to its characteristically unmistakable image. In addition, in contrast to commercial film scanners and thanks to the simple design, it is relatively easy to carry out repairs or modifications. Disadvantages, however, are factors such as the slow scanning time and the time required for post-processing. In addition, the scanner has no safety mechanisms, which is why the digitisation process should not be left unattended.

Possible improvements

The scanner still has room for improvement in many areas. For example, at the film gate, where the film is currently only fed through a gap about 1 mm wide. This prevents abrasion of the film, but sometimes the film is not exactly in focus. Therefore, for the next iteration, I would like to develop a film gate that guides the film over more spools to keep it a little tighter without the film being subjected to too much friction. The backlight should also be better shielded. The current film gate lets too much light past the sides of the film, which leads to reflections in the lens. I solved this problem by attaching some insulating tape to the film gate to block out the light. However, the light is also dazzling during use. In addition, the wiring is still a bit confusing at the moment. This would be remedied by a separate circuit board that makes the electrical connections and provides connections for the camera and control unit. This board should be made in the form factor of an Arduino shield so that it can simply be plugged onto the Arduino.

Looking back, the most complicated part was the conceptualisation. The actual manual assembly is very simple compared to this. This mainly involves drilling holes, filing some of them out and then fastening the components with screws. On the electrotechnical side, parts have to be soldered or plugged together. Therefore, a replica should not be very complicated. For me personally, this project has above all taken away the fear of facing new, unknown challenges. Subject areas such as electrical engineering were previously difficult for me to fathom. By building the scanner, I was able to gain a basic understanding of it.

I can advise anyone who wants to tackle this or a similar project: Just get started, don’t be afraid of making mistakes. There will certainly be points where you get stuck. At such times, it makes sense to continue elsewhere until you find a solution. There are enough people who are willing to help, and it doesn’t cost anything to ask. My plan is to publish the construction plan, the circuit diagram, the software and all the 3D data. I hope that this will allow the project to grow and improve. I have already worked on a first version of the construction manual as part of my bachelor’s thesis. This will be available online and will be easy to understand, even for non-experts, as each work step is shown in an interactive 3D visualisation.