Close-up of a film reel showing multiple frames of a scene in which three people talk to each other outside a building The images are very slanted as the film reel is lying horizontally.

A Brief Introduction to Film Restoration in Taiwan

For the past four years, the Taiwan Film Institute has been undergoing a film restoration project to restore old Taiwanese films. M+ Screenings: Restored Images from Taiwan (4–6 October), co-presented by M+ and Kwang Hwa Information and Culture Center, brings together moving image works that are part of the Taiwan Film Institute’s undertaking.

Below, we invite Dr Wang Chun-Chi, director of the Taiwan Film Institute, to explain how film restoration works, why it happens, and the debates that occur within the field.

What is film restoration?

Dr Wang: Film restoration means rescuing and preserving old films with deteriorating stock. For us, this means both preserving and conserving the film reels, cleaning and repairing them, and then converting them to a digital format. Each year we have a group discussion about what to digitally restore out of the films we have already preserved.

Film restoration allows us to discover films we may never have seen before, or have only read about in books. For example, many people think that we have lost all Taiwanese-language films, but they do exist—they just need to be restored.

How does film restoration work?

Close-up of two hands in white gloves holding down an orange film reel on a table under a light. The right hand holds a cotton swab and is using it to clean the reel.

Scratching away sticky adhesive residue during the physical repair process. Photo courtesy of the Taiwan Film Institute

Dr Wang: After we have decided which films to restore, the process of restoration involves several different stages.

The first is the technical selection of the film elements. This means finding sources for the film, possibly from numerous reels, and deciding on which one to use. Negatives can be filthy and riddled with scratches, rips, and tears. Sometimes, parts of the image have been damaged. We reduce scratches, remove dirt, and scratch away sticky adhesive residue. In some cases, we have to use new tape to remake a splice, meaning a spot where different sections of film have been combined.

Once the cleaning process is complete, we assemble all of the master elements. Then, this master version of the film is ready for digital duplication by scanner. When this is done, the images are ready to be digitally restored.

There’s also sound restoration, which is a whole other project. For sound restoration, you start by figuring out how the audio was recorded and stored: optically or magnetically. A common practice to restore soundtracks is to digitise the image of the soundtrack, which creates a pictorial representation of it, essentially turning the sound into a waveform. Although you can look at the picture to see which parts need improvement, a sensitive ear is still a must.

Close-up of a film reel covered in oily stains, obscuring the image on the reel.

Visible oil stains before ultrasonic cleaning. Some projectionists might have added lubricating oil to facilitate screenings. Photo courtesy of the Taiwan Film Institute

If you watch a decades-old film that hasn’t been restored, you will see a lot of vertical scratches, created by it having been run through a projector hundreds of times over the course of many decades. You will also notice a flickering effect that makes it seem as though the film is flashing. This flickering effect is a result of vinegar syndrome, a chemical reaction that causes parts of the print to change colour. Finally, dirt can cause a dot to appear within a single frame and then disappear in the following frame, which is known as a one-frame defect. Sometimes you will also see the crystallisation of the chemical materials. Through restoration, all of that will be repaired and cleaned, so you see a clear picture.

Are there any interesting restoration stories amongst the works displayed at the M+ Screenings programme?

Film still of a scene in which three people in ancient Chinese robes stand or kneel in front of a yellow table in a courtyard. A sparse assortment of neatly laid out gold-coloured tableware, including two candlesticks and two bowls, lie on the table. A crowd of people sit in rows in the courtyard watching the people at the table.

King Hu, Raining in the Mountain (film still), 1979. Courtesy of the Taiwan Film Institute

Dr Wang: King Hu’s Raining in the Mountain (1979) is an especially interesting case, as it’s the first film that the Taiwan Film Institute restored entirely on our own. For this film, we had to integrate different sources from different countries—such as the camera original and different release prints— to make one single-reel master.

When we realised that we had to put together so many different sources, a debate occurred amongst my colleagues: When should we use the camera original—with shorter length but clearer images—and when should we use the release print—with longer length but worse image quality? For example, what should the colour tone be? Should it be brighter, or redder? When you have so many sources, it’s challenging to decide which one will be the standard.

In the end, my colleagues actually visited Chin Pao San cemetery at New Taipei City where King Hu's commemorative statute is located. They had a drink with him, hoping that he would give us his blessing so that the film could be completed perfectly. I really respect how they tried to communicate spiritually with the filmmaker, and, with the final result, I do think we got his blessing.

This case brings up some hotly debated issues in the field of film restoration. For one, what is your definition of the ‘original’ film? Sometimes you have the camera original, but sometimes you only have a release print, which may also come in different versions. You also have to question the extent to which you ‘improve’ the quality. For example, many people don’t like the restored version of Gone With The Wind because the quality is so crisp that it no longer looks like a film from 1939. So is that, in fact, an improvement? It depends on your objective and on the condition of the so-called original.

Every time you restore a film, you’re technically creating a new version of it. We have to be aware of the fact that, no matter what we have done, we have always done something. It’s not going back to the original.

How does the Taiwan Film Institute project team decide on the film list to be restored every year?

Film still depicting a scene in which two people use farm tools to work on the ground in an empty desert environment. The person in the foreground wears a blue suit and a white brimmed hat, while the person in the background only wears a pair of white shorts.

Richard Chen Yao-Chi, Liu Pi-Chia (film still), 1967. Courtesy of the Taiwan Film Institute

Dr Wang: We started our restoration project with a few criteria. The first one was the canon: the films that are most commonly considered ‘important’, either critically or commercially. This perceived status can help persuade people that a particular film is important to restore.

The second criteria was auteurs: that is, film-makers with a highly recognisable style. King Hu would be one of the best examples, as we have pretty much restored all of his films. We also restored one of Hou Hsiao-Hsien’s first films, from the early 1980s, as well as a film by Richard Chen Yao-Chi, an avant-garde director in the 1960s.

Each division in our institution—research, promotion, education, preservation, and digital restoration—has different ideas of which films to restore. We ask each division to propose their list, and then we have a discussion. The research division now leads that discussion, because we realised that researching the films and what happened to them helps inform our final selection. Film restoration is a highly ideologically charged process. It’s about the politics of selection. We need to have a clear idea about what we restore and what impact it will have. Research greatly aids in the attainment of that understanding.

Personally, I would like to turn our focus to films that may not receive as much attention, such as those made by female directors. Because of the conditions they faced during the era when they were first released, these films may not be considered part of the canon, but they are important works whose value deserves recognition today. I would also be interested in restoring images from old films and newsreels that constructed depictions and stereotypes of indigenous people.

I hope we can turn our attention to whatever’s being ignored in film history. Perhaps if those previously ignored works can be seen, we will come to view film history differently.

How does the situation of film restoration in Taiwan compare to Hong Kong?

Close-up of a person using a tool to scratch tape off a film reel. One gloved hand holds the film reel down on a flat illuminated surface to backlight the reel, while the other holds a small tool and is in the middle of scratching the surface of the reel.

Peeling off residual tape during the physical repair process. Photo courtesy of the Taiwan Film Institute

Dr Wang: I would say that the Hong Kong Film Archive has a better preservation project than we do. If I go to Hong Kong Film Archive, it’s very easy to ask the librarian and get what I need for my research. They have also been repairing and cleaning the films. However, because of funding, they have not been able to start the large scale digital restoration project yet. And because film is prone to deterioration under high humidity, the climate in Hong Kong is also challenging to the restoration process, similar to how it is in Taiwan.

We have been doing digital film restoration for a few years now, because in 2012 the government and our institute co-initiated the digital restoration project. We started by visiting foreign institutions and learning from them. It took us three years to finally finish our own complete restoration, of Raining in the Mountain. We now have a restoration crew so that we can do the work completely on our own.

We hope that the restoration work we have carried out in Taiwan can help prompt discussion in both Hong Kong and Taiwan on the way forward in preserving film art and culture.

Learn more about M+ Screenings: Restored Images from Taiwan, happening from 4–6 October 2019.

The above interview has been edited for clarity.


Image at top of post: A film reel during the film restoration process, after ultrasonic cleaning. Photo courtesy of the Taiwan Film Institute

Return to Blog