Official Trailer of "Dragonfly Eyes" at TIFF 2017, a feature film consists of more than 500 surveillance videos and online live-streaming images

The Chinese artist Xu Bing is internationally renowned for his controversial piece "A Book from the Sky", first publicly exhibited in 1988, during a turbulent moment both to China and to the whole world. Starting from there, his works have been developed with the changes of technologies and social structures of our era. In 1994-96, as China opened it's door wider, he developed a "Square Word Calligraphy" system to write English words in the forms of Chinese characters. In 2014, with the development of mobile internet, he wrote a narrative book, "Book from the Ground", with only emoji characters. In 2017, with a similar approach, he presented in the New York Film Festival with his first Feature Film, "Dragonfly Eyes", consisted only of surveillance footages and online live-streaming videos.

One of our interns Anthony Huang had a chance to chat with him about the film when he came for a Q&A during the film festival. Follow his instagram right here, read the full interview below.

Musée: Your film consists of more than 500 surveillance videos and online live-streaming images, could you please talk about how your combine your individual surveillance footages into a feature film with complete and mainstream narrative structure and coherent imaginary characters?

Xu Bing: In the beginning, almost everybody, especially filmmakers, told me the idea really wouldn’t work. Because you need to have some fundamental elements to support a narrative film, such as characters, certain circumstances and situations, the developments of the storylines, etc. But the materials of surveillance camera, and my ways of montage, obviously would violate these fundamental rules. Without an actual protagonist, how can you maintain the impulse to let the audiences follow the story? I was certainly worrying about that.

But I believed that I could pull it out through various techniques and tricks. Through some personal connections, I got my first videotape of the surveillance camera, it was a recording of a hospital parking lot. There was a woman ran into the hospital in a rush, and came back to the parking lot shortly after. There was a guy inside the car waiting for her. They whispered something to each other, and eventually drove away. As I was staring at the recording, I started to make up random stories.

But of course, to actually make it relatable, I still need some tricks from the narrative. For example---Plastic surgery is a crucial part in the story, with this plastic surgery, the problem about subject changing in different surveillance footages has been partially solved, since they changed their facial identification through the surgery. Another thing that we tried was sound design. For example, in the plot, we need to have a car in this corner, but obviously there wasn’t any car in the footage, then we can use sound to solve that problem.

M: What were your concerns and precautions taken to address about the ethical dispute that might be generated by this film: How did you get the consent from the subjects; and what was your concern about imposing your own narrative and interpretations over these real personages and incidents?

In the production phase of this film, we were observing day by day about these subjects’ daily lives, we diligently followed their timetables, for example, this guy will wear this shirt today, and that lady will put on that dress tomorrow. In this restaurant, she is closer to this colleague, etc. Then I started to think about the relationship between the subjects and us(me and my studio assistants). I got this impulse to search for these people(since all of these footages are uploaded to the cloud with their specific GPS positioning)---first reason, to gain their authorization for us to use their portraiture rights; Secondly, we want to know about their opinions. They are living a real life, while we were imposing some stories which were completely unrelated to them over that.

(M: Did the subjects upload these videos themselves?)

X: That’s more complicated...some of the footages are self-initiated, such as webcams, online livestream, etc, and others were uploaded by the institutions. While we were editing this film, we spent equivalent amount of energy to search for these subjects. In the end we made another documentary about us finding them.

(M: What were their reactions when they were told about becoming the characters in your story?)

X: Most of them agreed us to use the footage, on the other hand, their interpretations about Portrait rights are somehow different from the western or intellectual point-of-views. They are more open to the surveillance camera, they hope to get in touch with this world through this way, to share their lives with the world outside their own spaces.

M: Because of the limitations surveillance camera, Sound design is crucial in this film to let viewers emerging into the narrative; Please talk briefly about your ADR and sound-mixing process.

X: The sound design and mixing was very complex. We needed to foley most of the footages, since most of the CCTVs weren’t having the microphone; Even they have it, still don’t sound really good, or didn’t fit well with the emotions.

Another thing is dubbing the dialogues: We do want to emphasize the dialogues to throw audiences into the plot. On the other hand, we wanted them to be naturalistic, unlike mainstream TV series or movies. In the beginning we were using professional voice actors, but they all had this kind of “performing tone”, since they were mostly dubbing for TV series or feature films, which force them to sound performative. However, it really wouldn’t suit our film, since all of the actors in Dragonfly eyes, were not performing at all.

We were also trying to position the dialogues within the space, in order to create some distance. They should sound unclear or soft sometimes, since we need them to have a telepresence and distance that we can feel through surveillance footages.

M: The plot in your film is very dramatic and engaging, but it's consisted of the true events happened and happening in China. Could you please tell me how you resemble this mainstream dramatic love story with the Chinese social reality? And can you tell me why you were trying to arrange so many timely issues in 81 minutes?

X:This film is hyper-realistic, since it was extracted directly from daily lives. But at the same time, it’s an imaginary and symbolistic movie with a romantic tone. You can label this film because it mixes lots of genres. Our earliest struggle, was trying to make this film more cinematic than it supposed to be. So in the beginning, we used Panavision frame as our aspect ratio(2.35:1). But then we started to realize that this story flows too smoothly in the cinematic sense. These scenes are already so real and present, and then you use a realism way to narrate them, isn’t that a little bit boring and redundant? Then I realized that it’s harder to utilize the special features of the media to make the story unique, than telling a clear and mainstream story. So we changed the aspect ratio back to 16:9, closer to the surveillance footage, and we brought back all the timecodes and source informations, to let them feel exactly like CCTV. For instance, the earliest footage was taken during 1999, while the newest one was on April 2017. Through the montage, you can see the footages 4 years, 5 years or 3 years ago were edited all together in a linear narrative manner, which created an interesting relationship, which guides you to reflect about time, reality, future and the fluidity over distances. That’s the actual theme for this film.

First edition(more cinematic) of the trailer in 2016, uploaded by Xu Bing Studio

M: (Spoilers Ahead) The facial and object identification technology is very advanced in the film, but it stops functioning facing bunch of ladies who went through plastic surgery. It involved both issues about surveillance and identity confusion. Could you please talk about that?

X: Lots of the technology that this film has mentioned, including facial identification, Internet celebrity, and the live-streaming surveillance footage on the cloud, are all started to emerge after 2013, the year when we started film production. In some scenes, there were automatic zooms, tracking and stretching of the camera lens ,including the night mode infrared technologies, were all the camera’s own features that we did nothing about in the post-production.

And of course, facial identification is very important in this film, it involved with identity confusions, about what we can or cannot see, about how we can identify the essence. There are actually quite a few storylines in this film: firstly, it’s about how the machine handles highly complex issues and phenomenons in humanity, how to judge them or categorize them; secondly, it’s about how police uses the machine to track or identify the suspects; Lastly, is the romance between the female and male protagonists, in the beginning there was this female protagonist’s monologue, and it turned into the voice of her male counterpart in the end.

M: You have seldomly used camera in your previous works, but this time you have successfully challenged yourself again, using the actual footages, did you gain less control or more control over your arts?

X: It was a really big challenge since filmmaking is complicated. But throughout my art career, all of my series look different from each other, like pieces from different people, overthrowing the previous ones’ ideas, But deep inside, they all share inner connections and my own unique perspective. I like challenges, I will work as hard as I can for tackling the difficulties. Your creativity and energy will be instigated only when you are having a challenging task.

But filmmaking is extremely difficult, I probably won’t get the result that I wanted without the creation of “cloud”, I mean I could still find some footages, but It would be unimaginable to collect this amount. As a visual artist, I also have to reconsider my own patterns, and have to learn a lot from the editors and filmmakers, since each field has it’s own way to determine certain things. So I told my team that “we made something new from zero.” Before us, there was no film like this.

As I said before, we chose some more “cinematic” shots to make it more “blockbuster” like in the beginning, but then I found it uninteresting, and chose to return to and emphasize the surveillance nature. In fact, I found out later that, the images grabbed by the surveillance, are vivid and peculiar, they transcended our traditional understandings over photography aesthetics. Just like when we were drawing still objects in art classes, it’s very formulaic and repetitive: Put a bottle on the left, put an apple on the right, add a banana on top, and hang a drapery from the table... Same for the surveillance, all of it’s automatic movements also contain elements of traditional aesthetic practices. But the purpose of surveillance camera is not for aesthetics nor compositions, but for obtaining the information in this area, so it needs to have as many perspective as possible, which brought unexpected abundance for our film.

Oct 17 Fruits of surveillance: Xu Bing Interview about "Dragonfly Eyes"

Oct 19 Women Crush Wednesday: Sarah Sudhoff

Oct 16 CURRENT FEATURE: Federico Solmi

Related Posts