WUSHU – Behind the Scenes

By June 4, 2013Music, Project, Travel, UVa


First, an infinite amount of thanks goes out to my dear friend EFuzzy. Without his help and patience this project would never have completed. He endured much bickering and criticism during a personal time of stress, and I am forever grateful for the time he gave me for this project, despite how much he came to hate filming and my lack of communication upon the filming’s completion. A huge amount of work, time, and sweat went into this project, and unlike most of my assignments that I’ve worked on, this was definitely a two-man effort. Previous projects like Together Alone were done with multiple people on board, but both of us had knowledge and expertise that overlapped greatly, but here, I worked with a friend that had a completely different skillset and knowledge pool than I did. Attempting to convey each others ideas and communicate properly quickly became a project challenge but it provided a wonderful learning experience. I traveled up to Cambridge/Boston to visit my friend, EFuzzy from March 22-24. My initial purpose in fling to Massachusetts was actually mainly to attend a the ACGA Go (Weiqi) Expo that weekend. I was a semi-VIP member as I had helped design promotional material for the event, as well as helped the closely related Go Documentary develop their website and brand identity. My TRUE goal, however, was to squeeze in time to film a wushu project for one of my classes. An assortment of photos during my limited time of sightseeing/travelling/walking around can be found below the twirl tag:

Travel Photos

1 2 3 5 4 8 6 7 10 9 11 12 13

I. Concept Planning

The project began from an assignment titled “What Time Is This Place” in my ARCH5424 (Direct Cinema Media Fabrics) course. Within this exercise, we were asked to present something that depicted “time,” whether it be accelerated, decelerated,or a sequence of different times. The overall interpretation of the assignment was quite lax and flexible, so I took it upon myself to loosely generalize the assignment as either a time-lapse or slow-motion project. Given that, I wanted a project theme that would exude a certain beauty and perhaps even inspire those watching. I initially drafted up an idea to film Charlottesville and its neighboring nature parks with extended time lapse sequences, showing evolving landscapes with the sun and moon dancing through the sky in removed time space. Numerous difficulties, however, arose from such a proposal: 1. obtaining a camera that was cheap and durable enough to withstand extended outdoor shots, 2. finding the time to shoot enough footage for extended time lapse, and 3. and travelling to the particular location to shoot (would I have to drive many hours and camp many days?). After further reflection it quickly became apparent that a time lapse video without the proper equipment and time would prove to be time consuming and ultimately too difficult. Thus, the next logical thing to pursue, if the speeding up of time was impractical, was the slowing down of time. This then begged the question of what the subject material would be. YouTubers like the Slow-Mo Guys film slow motion sequences of every day objects colliding, falling, and exploding. The largest issue with their case is that their equipment is expensive. Using phantom cameras of super high frame speeds cost upwards to tens of thousands of dollars. Given that my personal budget for this assignment hovered around zero, obtaining such a camera was unrealistic.

II. Pre-Production

Obtaining the materials needed for this project turned out to be a pretty large headache. For nearly three weeks, I hadn’t found a single person I knew that carried a camera that could film at 60p (progressive – 59.94fps)/had one available. I got increasingly anxious when two days before shooting I still hadn’t yet safely found a camera to film with. Fortunately, with extended discussion with managers in the University Digital Media Lab, I was able to borrow a Canon EOS 7D. The complementing stress that came with setting up for shooting was finding lighting equipment as well as a location to shoot at. Most of this fell onto EFuzzy’s shoulders and opened the beginnings of confusion, different expectations, and numerous periods of exasperation. I’ll spare you all the details, but eventually, EFuzzy fortuitously found RULE Boston Camera, a suprisingly legit film production rental company. Here, we rented a set of four lights (two 650w omni and two 700w tota); it came with umbrellas, reflectors, and stands. This was super crazy to look at when we first picked the equipment up; we felt like real filmmakers! What sucked, though, was having to lug the entire case around Cambridge and Boston for several days. The case was poorly designed for distance travel not moderated by vehicle. A large facet of this project depended on what kind of music was ultimately chosen. With slow-motion in mind, I decided that music with drums and high energy/tempo would be most adequate in juxtaposing the freeze-slow portion with the normal/fast cuts as this adds greater tension and anticipation. Ultimately, I chose Foreign Identity by Two Steps from Hell for its very percussive nature and natural build up towards the end. The piece gave me a frame of reference and allowed me to create a rough outline of how I wanted the shots to be organized with a slower preparation segment in the beginning and a subsequent growth towards the final climax. Discussing with EFuzzy during the Pre-Production phase was probably the biggest area that could have improved. By the time of filming, I had never really solidified specific sequences or shots I was truly aiming for, which ultimately led to confusion and lack of focus. Combined with my limited exposure and knowledge of wushu, there seemed to be a slight barrier in communicating with EFuzzy on a level where we both could carry a unified front and effort in the project’s vision. Fortunately, EFuzzy was empathetic enough to stick with me even though most of the difficult work rested on his shoulders. Thanks, again, EFuzzy.

III. Filming

The entire project was filmed in one day, with a pilot shooting two days prior. The final shoot took roughly 4.5 hours and yielded about 30 minutes  (1767 seconds to be precise) of actual footage for a 1 minute project. I’ll say that I probably could have focused a lot more on engaging with EFuzzy and planning out what kinds of moves, sequences, and styles I wanted to film. I had never developed anything super concrete besides basic shapes synchronized to music while laying in bed, so it was difficult to get EFuzzy on the same boat and go along with my repetitive stop n’ shoot style. After several hours, EFuzzy’s energy was draining and his patience was wearing thin. Not to mention I had broken one of our rental lights due to clumsiness and EFuzzy had a thesis to write, so tension was definitely in the air. After the shoot, however, we left with some really good footage, only marred by the fact that we had filmed indoors as opposed to outdoors where bright light would help with higher shutter speeds significantly. When shooting, it was important to always keep in mind where the lights were relative to the subject. Every time EFuzzy moved out of the center of light, any footage from that sequence would have to be cut out as the lighting conditions were already poor enough. Keeping the subject under focused light also allowed for a higher shutter speed, which ultimately reduces motion blur; this was hardly flexible, though, since omni lights have nowhere near as much power as the natural sun and the DSLR camera was stubborn and never went about 1/500 (1/1000 is ideal). Also, attempting to orient EFuzzy from different perspectives and locations was paramount in getting angled shots as the room we shot in had only 1-2 walls that we could naturally film against. Many wushu videos focus on filming forms from far away, capturing the entire body and displaying the sequence in the most wholesome manner possible. Even martial arts films tend to have many shots from far away, framing the entire fight rather than just one person. Of course, this is completely a stylistic choice, but it’s an understandable one as if we had a face-shot of both fighters the entire time, there would be no understanding of what was going on. An additional reason why martial arts, wushu in particular, are filmed from a macro stance is that the movement are so fast and varied that it’s hard to capture them

IV. Post-Production

Here is where the real grunt work started; importing, labeling, and organizing all the clips so they were easily accessible took a handful of time. Even after painstakingly trying to figure out which clip corresponded to what wushu form/sequence, I still feel that the final result could have had some better clip choices. The organizing, however, wasn’t the time-sucker; it was ultimately the editing together of clips, adding effects, and waiting for each scene to render/compile after every small edit. I personally am not too familiar with film production workflow and what’s the best way to set up a sequence and when to render the workspaces, so I usually end up rendering for several minutes/hours and doing homework or playing starcraft in between. I used Twixtor from RE:Vision Effects in After Effects to render my 1% slow motion sequences and Premiere Pro to compile everything. Unfortunately, I currently use CS4 AE and CS5 PP, so I had no dynamic linking available, which would’ve been wonderfully helpful. When using Twixtor, it’s important to use scenes with good contrast in lighting (which can be slightly remedied with color correction techniques prior to the application of Twixtor, as well as a good subject against a flat background. Twixtor models extreme slow motion by interpolating between two frames and ‘smearing’ the first image to turn into the second. Of course, complex algorithms and one can achieve a surprisingly good result if the clip and settings being used are most optimal. In the case where the background has textures, viewers normally can see warping around the subject that is being ‘smeared’ as Twixtor is unable to intelligently decipher what is the main object and what is the background. Additionally, Twixtor is normally not very useful with footage that has movement of extreme speed where even 59.94 frames per second cannot pick up. If the subject move further than its own width (or even half) in less than 1/60th of a second, then Twixtor will normally not understand that these two objects are related, or it will warp it so badly in between that it seems like we’re watching a shape-shifting amoeba. I decided to add a Chinese flair to the video by adding Chinese for ‘wushu’ in the title sequence and splicing Chinese ‘proverbs’ in the intro segment. These ‘proverbs’ are generally not genuine, as I merely searched on google for idioms that described martial arts or wushu (which isn’t easy to come across). If I hear/am offered any better alternatives to some of the quotes, then I’ll be glad to change them in the future. With the timing of the music, however, the text is pretty hard to read in the beginning especially because viewers are not necessarily ready to read at such quick intervals; though the text itself isn’t really as important as ‘feeling’ the intention. Post-Production probably took around 14 hours to complete (though 35-45% of it was waiting for rendering). For those interested in Behind the Scene Photos, view below the twirl menu:

Behind the Scene Photos

1 3 4 5 6 7 8 9 10 11 12 1314 15 16 17.1 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 bts4small bts3small bts2small

V. Reflection

The final cut of the video ended up being pretty high octane and satisfying to watch. Despite some obvious warping in the Twixtor sections that I couldn’t avoid (perhaps I can edit it later in After Effects in the future), I am quite happy with the choices of clips and ultimate pacing. I probably could have equalized the RGB curves a bit better and more consistently as to not over or underexpose certain sections, but the general reaction when screened in class was relatively positive. If I could choose one thing to improve on the next time (besides my communication abilities) if I were to work on a project like this, I would most likely want to improve my understanding of camera angles and framing everything correctly. I ended up having a lot of flexibility with scaling and translating the image within Premiere Pro in post, so I could adjust some framing issues within reason, but it’s always good to shoot good raw footage first as well. Perhaps, additionally, a better understanding of lighting would’ve helped improve the quality and graininess of the final product (although it’s not TOO bad as it is).

In the end, though, the project was immensely fun to complete and I’m pretty satisfied with the final product. One thing that I realized during the entire process was that I failed to really plan ahead of time thoroughly. To maximize on both EFuzzy’s and my energy, it would have been optimal to have a much more firm idea of exactly what kinds of shots I would require and shoot that way there is less confusion. I ended up being lucky this time around and having nearly all the clips I needed to edit together to make it look great; but that was a product of spending multiple hours just shooting things over and over again.

Thanks to everyone who helped out! The final video can be found here.

Shot with the Canon EOS 7D.

Edited in Premiere, After Effects (Twixtor)

Directing and Editing: Christopher Hsing
Wushu Artist: Edward Gan
Music: Foreign Identity by Two Steps from Hell

Inspired by Stunt Poetry by Rishi Kaneria (vimeo.com/27612305)

EXERCISE III: A Study of Time for ARCH5424: Direct Cinema Media Fabrics (arch.virginia.edu/arch5424/) course at University of Virginia’s School of Architecture

Leave a Reply