Angela: Ulrike, would you briefly explain to our readers what is your professional field?

Ulrike: As a self-employed person, I am mainly involved in digitizing of historical film material, including city archives and film archives, universities, etc. From two years, I have been doing digital restorations for archives. I am also involved in the development and distribution of technology for film transfer. In the past, I worked as a graphic designer for a long time and studied fine arts. So although I have a lot to do with technology, I also have access to the other, creative side of digital film restoration.

Angela: What is your view on film archives and restoration practices? Why is it important and how could it help shape the historical landscape of cinema studies?

Ulrike: The analog archiving of historical film material is of course extremely important for cinema studies. However, archiving and restoration are completely separate areas, just as analogue and digital restoration are. Conventional restoration is important for cinema studies as it can help to restore film shot in analog. Digital restoration goes a big step further, however, because it attempts to reconstruct historical reality. Smooth movements are created from jerky ones, film grain is retouched, and colors are restored …

Angela: What are the benefits of integrating AI into your workflow?

Ulrike: AI can be used in film restoration to create moving visualizations of the past that were almost unimaginable until a few years ago. The advantage lies in giving the viewer easier access to times long past and thus awakening more people’s interest in old film material. The peculiarities of analogue films, such as streaks, grain, blurring, non-fluid movements, and non-existent or unrealistic colors, often stand like a frosted glass screen between the film content and the viewer.
The possibilities of digital restoration, however, make this glass screen, this separation, much thinner. The viewer can be drawn into the action much more quickly and often be transported back in time much more easily.

 The biggest challenge is historical accuracy. It can never be achieved 100%, which is why it is always a balancing act, especially when using restoration methods for historical archives. A lot of trial and error, research and manual work can improve the results. But at the current stage of development, it will always remain an approximation. The second challenge, in my opinion, is the danger of over-restoration. Even if you try to remove the obstacles for the viewer, for example, film grain, monochrome, scratches, and film damage, and thus reduce the distance between the film and the viewer, if you overdo the restoration, you could create a new distance. If everything looks too smooth and perfect, it goes against the human experience of viewing reality and even more so of viewing older films. This challenge is also a real balancing act.

Angela: Do you think that someday AI will be able to replace manual work?

Ulrike: When it comes to AI, there is a fear in many areas that it will replace human labor. But in digital film restoration, it is only used to carry out work that could hardly or not at all be done manually or without AI.
It is, therefore, an extension of the possibilities, which would also be almost unimaginable without a very complex workflow in the periphery: a lot has to be taken into account when scanning the film because mediocre digitization prevents good digital restoration. Detailed research and precise descriptions for the AI are a necessity. No good restoration can be carried out without extensive background knowledge.

Angela: What are your views on the aesthetics of film restoration? Is there any guidance on how the film should look after digitization and colorization?

Ulrike: I think every digital restorer should preserve authenticity as far as possible, which of course often greatly increases the amount of work involved. Research is often necessary:
– What color was this company’s advertising poster in 1930?
– Does the mayor’s speech still exist as an audio document, or should we hire lip readers?
– Was this device made of brass or steel?
– What time of year was this event?

Apart from authenticity, there are of course always questions of design, such as the lighting mood, to be considered in the aesthetics of the restoration.
This is where the restorer also becomes an artist. This is precisely the reason why many people tend to be skeptical about digital film restoration.
Even if there are no binding guidelines etc as yet, in my opinion, the most important thing is always to identify a digital restoration as such. It is essential to communicate any digitally restored film, no matter how short the snippet is, for example using a note like: “We have tried to reconstruct reality as accurately as possible, but there may be errors. Not every detail can be represented authentically. It’s a mixture of reconstruction and artistic representation.”

Ulrike Schmidt’s work can be found at:
https://www.film-digital.com
https://www.film-overscan.de/
https://www.youtube.com/@film-digital2129