I’ve received a number of emails asking about various aspects of the production since the video went live – some from students, some from other filmmakers and some from Spielberg. Some want to know which gear it was filmed on while others want to know how the ‘It-doesn’t-look-like-it-was-filmed-on-DSLR’s’ look was achieved. I’ll try to cover most of the production in this post, but keep in mind that this was a zero-budget project, so everything is done in the most ghetto/cheapest way possible. The following info may not necessarily be the correct/best way to do certain things, but it worked for us…

The aim of the video / why we made it

Apart from showcasing the team’s skills, the video was shot and edited with one main aim in mind. I’ll copy a post Tim ‘Livewire’ Shieff made on a forum as he sums it up nicely;

Our aim really was to try and improve the general publics perception of freerunning. At the moment I feel it’s seen as just kids in cheap sweats that jump around recklessly on walls, and that makes it hard for sponsors to target the sport. We did this to try and show them its more of a lifestyle and that we do care about our image as well. Basically trying to show sponsors that there’s more potential in the sport than some previously thought so we’re hoping this helps benefit the whole community in some way.

Personally I’ve found that a lot of people refer to freerunning and parkour as ‘the art of movement’ yet are happy to slap a fisheye lens onto their cameras for everything they film. In my opinion it’s a totally overused effect, and excluding a few documentaries (Jump London / Britain, for example), adverts and films, freerunning is rarely portrayed in a way which justifies that term. I guess my personal goal was to capture the movement in a way so the viewer didn’t think ‘oh, I’m watching another video of kids fooling around with their camera’. I still <3 my fisheye lens but only like to use it when appropriate!

TL;DR: We wanted to make something cool which could be appreciated by everyone, including those usually critical of freerunning videos.

Pre-Production

One question I’ve been asked by a number of people is whether any of this was planned or not, and how we went about securing permission to film on Trafalgar Square, Westminster Bridge and on the side of a car park. Regarding the latter – we didn’t have any permission, we just went out and did it anyway. We had a few confrontations with security but 9/10 times they were ok with it once we were explaining what we were doing. The other 1/10 was usually around South Bank and the Hayward Gallery, where the rentacops ban you from creating any form of art without a permit – somewhat ironically as it is an Arts centre and they were promoting an exhibition on human movement at the time.

Although people have said a lot of effort must’ve been put into the production, it was in fact quite the opposite – we had a very, very casual approach to the entire video. As there was no budget, we didn’t bother with location recces or do much planning at all. Excluding the Trafalgar Square sequence, shooting the video was a case of arriving at a location and filming from whichever angles came to mind first.  The Traf Sq sequence was planned only because we knew the security would be onto us as soon as we got there (and we were right). I had a list of shots I knew I wanted, but the priority was the actual somersault. To ensure the Square was empty and to minimise our chances of failure, we arrived there at 5am.

Blue was asleep in the car on the way to the location, and within 5 minutes of the demon waking up, we had the shot on 2 cameras. Security approached us just as he landed the flip so we packed our gear and left – only to come back 15mins later, where the security were ok with us as long as no Health & Safety violations were taking place.

For the post-credits sequence on Westminster Bridge we again went out and shot between 5.30-6am. My initial idea was to mask out any vehicles that drove past but we were lucky to find the bridge much quieter than expected. A Police Officer did give us the whole “You’re not allowed to film here” talk but went away after I told him otherwise.

The other talked about shot is Jan’s intro (involving the dropping of the chicken).

Again, there was no planning here – it was a actually last-minute decision right before we were about to head home . I knew in my head that I’d want a shot of him running towards the camera so we shot that first. Once I once happy (and managed to pull focus) 2 camera’s were set up to capture the whole run. The 7D with the 16-35mm was on the stabiliser (with me on skates) while Blue had the 5DII with 70-200mm on a tripod to capture the flip. We actually did 2 takes of this in the end as the chicken didn’t fall out on the first run.

The 5-second rule was followed.

The structure of the video was very loosely planned. The majority of it came through experimenting during the editing process. The only parts that were decided on were starting in the car park or Trafalgar Square and finishing with the team walking on the bridge. The car park sequence was made up on the spot but I did choose to film in an open-top car park as I wanted to finish on an overhead shot (to lead into the opening titles).

There was meant to be another section to the video after the Ash handstand / bridge timelapse sequence but there was a little headache on the first day of shooting it. As Ash would be out of action until December we decided to scrap it and just go with what we had. I wont reveal anything about this missing section as I’m sure we’ll get to film it another day.

Production

Although all of the lenses were used at some point during the video, the 16-35 and 70-200 were used the majority of the time – for example in the car park intro, where those were the only 2 lenses used.  Both cameras were using custom SuperFlat picture profiles to make grading a slightly funner task. There are of course discussions on various forums regarding the positives and negatives of using certain picture profiles but these were good for me! I didn’t use any filters to change colour / create light streaks as it was easier/cheaper to do this in post. The only filter I used was an ND8 to keep my shutter speed low on sunny days.

The 7D was stuck on NTSC to enable me to shoot in 60p over the PAL equivalent of 50p – the only reason for this was to get a few extra frames of slow-mo action once interpreted as 25p in post. I always used shutter speeds of 1/100 or 1/125 when the 7D was in 720p60 and 1/50 on the 5DII which was shooting in 1080p25. Both cameras were used on nearly every trick/run filmed so we would cut down on having to do multiple takes. As for usage, the 7D/16-35 combo was always on the stabiliser as I didn’t want to go through the effort of rebalancing for the 5D (and taking off battery grip, etc), while the 5DII was used when I wanted ultra wide, full-frame shots at 16mm. The 10-20mm Sigma was Blue’s and I didn’t always have it with me. Also both cameras were of course on manual focus and shots were focused with the aid of the on-screen magnifier.

I’m not going to review each bit of kit individually but I do feel I need to mention how good the Dakine bag is. It’s quite beefy and is one of the few backpacks I’ve found which can -just- fit my 17.5″ laptop while remaining hand-luggage friendly. I’ve also found it can fit/hold 2 lighting stands, a tripod and a softbox – plus speedlights, triggers, etc… Yassin² infromed me that a few photographers were asking about it at London’s Freeze Festival, so check it out if you’re looking for a big backpack.

All of the tracking shots were done by skating with the stabiliser although sometimes a slightly more advanced technique had to come into play;

Unfortunately this is the amazing tech behind the shot of Blue at 3:04, as much as I would’ve liked to use a heli or cablecam this is all that we could manage. I skated behind Blue with the camera held upside down on the tripod – tilting and lifting the tripod when necessary. The footage did suffer from some intense shakiness and rolling shutter but that was later sorted in post using a combination of VirtualDub and manual keyframing of position, scale and rotation in Premiere. As the footage was 1080p and the final video 720p, I had a lot of room to play about with zooming and stabilising.

We started filming in May and finished in early November. There were a few reasons it took us this long to film but it was mostly down to availability. Livewire was in LA for 3 months shooting for MTV’s Ultimate Parkour Challenge and the others were busy doing jobs here in the UK. I, too, had my own work to get on with so in those 6 months we probably only went out to shoot 15-20 times. I was also very reluctant to shoot on non-sunny days but in the end I had to give in to the Great British Weather™. Aside from that there were also a few minor sprains and stretches, which although not serious at all, still put off filming for 2 weeks or so.

Editing took around 10 days in total, although this was really spread out over the whole 6 months. Excluding the whole intro sequence, credits/outro and grading, it would’ve been a day only! As there was never a solid plan for anything, most of this time was spent experimenting in After Effects and learning 3DSMax.

I’m not going to mention why there are a few hooded Storm guys in random shots in the video – they are there for a reason but I’ll leave it to the others to reveal why

Post-Production

The whole production was edited on an Intel I7 950 (@ 4Ghz), 12GB Ram, GTX480 PC running Windows 7 64-bit.  To those probably wondering why I’m not Mac’d up – I’ve been using DOS/Windows-based PC’s (always home-built to avoid bloatware) since I was young and know my way around them very well.

As for programs, I used Adobe Premiere and After Effects CS5, Boujou and 3DSMax for all editing and effects. The major benefit to using Adobe products was the ability to playback DSLR footage without having to convert or pre-render anything.

Although the majority of the edit was done using the included filters/effects in AE/Prem, some of the effects and animations did require additional plugins. Trapcode’s 3D stroke and infamous Shine as well as VideoCopilot’s Twitch were used in the intro and credits sequences, while Twixtor was used to create the ultra-slowmo flips during the intro. I also plead guilty to using a fair bit of Optical Flares – again from VideoCopilot – my neighbour gave me a hard time over my decision to use them so I tried to use it as sensibly as I could.

VirtualDub (with the Deshaker filter) was used to ‘remove’ the rolling shutter and general shakiness of some of the footage. In my opinion it does an amazing job, and best of all, it’s free.

All grading was done in Premiere using its built-in colour correction tools – with the majority of shots corrected using only curves and colour balance.  Both of these filters were also CUDA-ready so I could playback graded footage in realtime. Have to give some recognition to Premiere’s Mercury Playback engine as, for fun, I tried applying 20 layers of colour correction to a clip and it still played back flawlessly. Below are a few before and afters showing the raw footage and final output.

Tim’s handstand shot was originally very noisy as it had to be shot at ISO 12800 on the 7D. Ash’s climb was filmed over 2 days as the car park was closed when it was sunny and I couldn’t get the shots I wanted from inside.  Unfortunately the 2nd day was cloudy so I had to fake a tiny bit of sunshine using Trapcode Shine in AE.

Although I used Color Finesse in AE for a few test shots here http://www.youtube.com/watch?v=XE1PvzJfYrU I found that I could recreate the same look I was after (and save a lot of time) by using only the colour correction options available in Premiere.

The overhead car park shot which zooms out into a city view is accomplished by overlapping the video footage onto a recording from Google Earth. I also made Tim’s car into a seperate image so I could animate it over the Google Earth footage. Once the animation was all done I gave the car some headlights for added effect. The screenshot shows the comp minus the Earth footage.

The SupremeBeing shot was firstly tracked in Boujou before the data was imported into 3DSMax to create the text. I won’t go too much into 3DSMax as I had to learn it as I went along but the scene consisted of the camera (+ movement data from Boujou), a few lights for the shadows and some bevelled text. Once I was happy with the look of the text I imported the rendered files into After Effects and masked out Blue where necessary. I was doing a very rushed job of the masking so ignore how bad it is!

For the MaxiMuscle shot I first created the 3D-looking logo in Photoshop before importing that into AE for another messy masking session. I animated the logo moving slightly forward while masking out the necessary areas to make it appear as if it were coming out of the wall. The exploding bricks were done using the Shatter effect on a duplicate layer of the logo.

The clip of the hooded Storm guy walking down the bridge at the end of the video was created by combining a photo-timelapse with video footage of the actual walk. I firstly filmed the Storm guy walking down the road, and then using the same composition started shooting the timelapse. Due to shutter-speed constraints in video mode I shot the timelapse as still images to get the light streaks from the cars. My memory’s a bit sketchy but I’m fairly sure I shot around 200 photos at 1.6s exposures. If I were to do the same shot again I’d probably experiment with shooting at 2 or 2.5s to get longer streaks. Although I was shooting on a tripod I didn’t have an intervalometer so the timelapse was a bit shaky – AE’s tracking/stabilise tool fixed it up after a few passes and luckily the rotobrush tool managed to create a near perfect mask of the guy walking so I didn’t have to spend much time combining the 2 shots.

What’s next?

Although not for a while now, Volume 2 and onwards will happen. As there were a number of shots we couldn’t achieve due to the lack of a budget, we’re all hoping to secure a financial sponsor for the next one. We all know we could (very) easily outdo ourselves were we to have some £ to film a video. We’d also like to avoid filming in London or the UK as it’s been shot over and over again by so many people now. For the time being however, we’ll continue making and releasing mini videos when we get time. Unfortunately we didn’t have anyone doing behind-the-scenes coverage but I will be releasing a quick edit of outtakes and whatever other footage was left behind.

If there’s anything I’ve left out, or any other questions you want answered, comment below and I’ll get back to you there.