An obsession with language, symbols, words – how they structure meaning, value, and communication – has long been a central focus for the Beijing-based visual artist Xu Bing. When it comes to cinema, however, Xu is a relative newcomer. Collecting CCTV, surveillance and streaming footage, Xu weaves it into a complex, paranoid, voyeuristic and brilliant entry into filmmaking with Dragonfly Eyes: an indication of his impressive malleability as an artist, both conceptually and formally. Bypassing the usual growing pains faced by many interdisciplinary artists, Xu and his crew have pieced together a work that couldn’t have been made a decade ago. Until recently, the footage used in Xu’s film hasn’t been publicly available, yet despite the haste he has assembled it (into a broad, intricate, and fascinating sociopolitical foray into China’s collision with digital life), there’s a genuinely impressive quality and depth to the work.
We spoke to Xu Bing about working with film for the first time, despite his lengthy career in art, the process of assembling footage, and how he initially conceptualised the narrative that eventually developed into Dragonfly Eyes. At the end of the interview, we spoke briefly to the editor of the film, Matthieu Laclau, who also edited Jia Zhangke’s Mountains May Depart.
*This interview was conducted with the assistance of a translator.
There’s an atmosphere throughout the film that is built on a very particular process: the CCTV footage from which it’s assembled. It comes together with a coherency I wasn’t expecting. I wanted to know where the initial idea came to make a film out of this footage?
The idea came to me back in 2013 and I was focused on this project from that time. I was watching TV and I saw some programmes talking about surveillance videos. My interest began there. This kind of surveillance video image was very special for me, because I had never experienced this before. This is a kind of condition of the person that they did not have awareness of, that were being filmed. The motion pictures we watch today, including documentaries, I think most include these kind of images.
How did you go about the process of locating the footage, and then choosing how to collect and assemble it? Did you have any preconception, or did it develop as you worked on the film?
Firstly, I was surprised by this image. I was interested in doing something with it. After that, I began to figure out what I could take from the image, and where I could find surveillance footage to work with. At the beginning I asked my friends from a TV station, then the guardians of the apartment we see. I went everywhere to find images, but at that time there was not that much. In 2015, a few years later, there were many images uploaded to iCloud, so I began to work on this project.
Did you still have to go through a process of contacting these friends in different agencies, or were you then able to collect these through your team?
I stopped using this kind of friend’s help through the company because it is not the best way to get this image. All the images that are used in my film are downloaded from the internet, and that means they are already public.
I noticed a photo in the press booklet where you’re standing with your crew, and you’re all looking at different computers with various footage on the screens. Did you have people collecting certain things or looking at certain channels?
In our team we had four to five people. Their work was to download the files from the internet every day. In the beginning we did not really know how to do this – how to work like this – so we started by classifying all the images we downloaded. It means we have different to move through effects and templates, for us to be able to have people classify these images.
What happens in that stage of bringing it all into these different sections, and working through it?
After that, we have the stories. We are trying to find a way to find images to fit to our story. For some of the images, we just do not have what we want to find, they’re classified. For example, if we need a scene with a car passing by, we look at the broadcast and find a camera that works. We would wait for some days to see if a car passes by the camera for our video, sometimes, but we would use that.
There’s a lot in this movie that couldn’t have been made ten years ago, as you said. I feel like it’s a movie where the footage you’re taking from is trending towards becoming more available. In using a lot of surveillance footage and streaming footage, what did you hope the film would be able to say in taking that on, I guess privacy and the internet is kind of growing prevalence?
What we really want to talk about in my film is the relationship between the surveillance image and the real human being. The relationship between these two. If I have a story, for example, a love story… the boy falls in love with a girl in my movie but the images I use are taken from surveillance footage. If I am successful in using surveillance footage to create a film, I think that proves that there’s a really important relationship between the human being and the surveillance image.
The structure of the film utilises all of these elements of surveillance footage and streaming footage – did you have any considerations when it came to assembling it all?
I think if you have this question, that maybe it was difficult for you to imagine using surveillance footage to tell a story. Did you think it was possible before watching this movie, that you could tell a story with these images?
Yeah, I think it has a certain intimacy that I wouldn’t have expected beforehand. In that kind of film and I think in the final few scenes it’s interesting to watch how that –
How that finished, how the story finished?
I really want it to have a story. I want to make it look fake. Fake like a feature film, where I want to have an ending that looks like it’s from a feature film. That’s why I use this kind of storytelling. The more that I can finish, the more it’s completed. The more it’s completed, the more it affirms my conception that the relationship between the surveillance image and the human beings is strong.
I think it works very brilliantly, and it approaches cinema froma unique angle in terms of expression. I haven’t seen works using found footage in such a focused and coherent way, to create that sense of voyeurism, that eerie atmosphere, while still maintaining a clear narrative. I’ve seen movies that will take landscapes or make experimental compilations, but I haven’t seen many works that use human beings in found footage in such an interesting kind of way. I’m curious as to whether this is a style that you want to continue kind of working in, or whether this is kind of a standalone work?
I don’t know if I will continue to do this, maybe. For me, art doesn’t come with a plan. As in, I can’t plan what I work I will be doing next. With the change of society, the conception we get, I think I will seek a conclusion to these ideas that this film explores.
I think that kind of spontaneity is what has made a lot of your work in the past as fascinating as it’s been.I think that’s probably all I need to ask of that. Thank you so much. I hope the film goes really well in the competition.
Yeah, thank you for talking.
Thank you. You have some questions to ask Matthieu [Matthieu Lachlau, editor of the film]?
[To Matthieu] It’s a work that rests quite a lot on the editing, how exactly did you approach that role?
How did we make it, you mean? How did we process the editing?
Yeah, I guess just how you worked as an editor. Were you the editor for Mountains May Depart as well?
Right, I guess just comparing two kinds of works where the editing seems a lot cleaner on one film than the other–
Yes, it’s a very different work.
For this film it was complicated, because we have so much footage, and at the same time it was very hard to find good footage to tell the story. It was a constant back-and-forth between editing and recording more footage, and recording dialogues and rewriting script.
It was quite complicated.
Do you think it’s worked out well?
Yes, I’m happy with the film. We all are happy with this version. Normally, you write the script first, after you shoot it and when you edit, you cannot change the dialogue to match, but here you can rewrite the script and you can recall more shots, and you can do everything at the same time so there are more possibilities.
There’s a lot behind how this film exists, it hasn’t got into that state of taking on a very normal kind of approach to scriptwriting and shooting. It’s interesting.
No, it’s not normal. It’s a very experimental way of making it, yes, because there is not a real… Of course there was a rough script before we started editing, but then you cannot find lots of scenes that are in the script. You cannot find matching footage to that scene, so you need to find some similar footage. Sometimes you cannot find it at all, so then you need to change it completely. Sometimes you find something which is similar but a bit different, so you need to adapt it, adapt the script and the dialogue. When you record the dialogue, you put the dialogue in the shot to see if you can match the lips, but most of time it’s not the same, so you have to rewrite it again and re-record and go back and forth like this for a year.
I think it’s worked out really well!
Oh thanks, nice of you to say that.
Thanks so much for the interview.