How to... build a LEGO car chase

The newest LEGO spot, Drive What You Love [below], was a lengthy, complex undertaking but which heralded a brilliantly high-performance commercial.

Below, Highly Unlikely producer Alexander Davis, director David Mellor and Framestore VFX Supervisor Charlie Bayliss, give an in-depth account of how they brought all the different aspects of the process together to build a LEGO car chase to be remembered. 

Above: The finished Drive What You Love spot.

This seems unlike most other LEGO spots; what was the brief from the client?

Alex Davis: This was a big deal for LEGO for a couple of reasons; it’s ‘cross-franchise’, so they’re combining vehicles from many different product lines in one ad, which is new territory. But it’s also a brand spot – there is no packshot and no naming of products - they just wanted an awesome, high-octane celebration of vehicles. 

With my experience in VFX I've been on the receiving end of both well-, but more often, badly-planned executions.

Charlie Bayliss: The Fast and The Furious, on a kitchen floor, without the women and guns.

Did you know how you wanted to approach the brief as soon as you saw it/what was the creative process to get to the final idea?

AD: We manage our jobs from creative brief through to final delivery, which means we carry a lot of responsibility. For this reason, we were insistent on taking a very collaborative approach.The whole project was an ongoing conversation with HighlyUnlikely and David [Mellor, director], plus the creative team at The LEGO Agency, Framestore and the DoP, Jordan Buck. We fed into each other throughout and tackled technicalities long before they could become problems.

The creative was fairly tight early on in terms of what it needed to be, including being appealing for both children and adults.

David Mellor: Yes, planning was key. With my experience in VFX I've been on the receiving end of both well-, but more often, badly-planned executions. This film relied so heavily on a seamless workflow from start to finish, we had to plan accordingly to avoid issues post shoot.

The schedule allowed us to scout very early in the process, which meant all creative development was grounded in the reality of a location. This allowed us to tackle problems early, rather than when arriving on set. The geography of the location also fed into and inspired the creative which we were able to experiment with and sell through to LEGO, so they knew what to expect from the finished film.

The schedule allowed us to scout very early in the process, which meant all creative development was grounded in the reality of a location.

CB: The creative was fairly tight early on in terms of what it needed to be, including being appealing for both children and adults. It was more about how we could physically and technically make it work without compromising the creative vision at all.

The team, the clients at LEGO and David were really open to us developing the creative throughout the pre pre-visualisation process. We knew that this stage would be really important for locking down the creative and essentially working as an edit, ahead of the shoot.

Above: WIP images from the spot and one of the final frames.

At what point did Framestore come on board in terms of the VFX/animation and how did that process work?

AD: We involved Framestore from the get go. We knew there would be a lot of points where the VFX and the cinematography needed to work in unison and so we would address each shot collaboratively, making sure everyone had room to flex creatively but never at the cost of shots working for one or the other.

DM: We had very close collaboration. From scout to pre-vis to shoot and post, it was a team effort. We reviewed work multiple times a week providing (meticulous!) feedback on animation, concept, look development, lighting and comp.

Cars are notoriously difficult to animate. The subtleties of their dynamics can be easily overlooked.

I do get obsessive, especially when it comes to the physics of cars. I think the animation was probably the biggest challenge, as it was a key frame exercise. Cars are notoriously difficult to animate. The subtleties of their dynamics can be easily overlooked. The team did a great job in maintaining reality while getting the performance we needed. 

CB: We talked through the idea with David fairly early on in the initial briefing stage. Once we’d made a plan for what we could do within the budget, we did a location recce and mapped it to 3D so we could do accurate scale pre-vis and work out the limits of what we could shoot.

Above: Director David Mellor, producer Alexander Davis, and VFX Supervisor Charlie Bayliss.

Apart from the actors it's impossible to tell what's real and what's not; how much of the spot is CG? 

CB: That would be telling now, wouldn’t it? It’d be hard to say, percentage-wise, but there are both CG and live-action versions of the cars, except for the City GT, which was designed just for this commercial, and the dragster; they’re only CG. 

It’s a real environment, though it switches between live-action and CG. And there are loads of other additions throughout, like the flour and obviously the LEGO versions of each car. 

The aesthetic mission from my perspective was to make it feel as close to a car chase that could have been shot on real streets as possible.

AD: David has an excellent mind for balancing VFX and live-action in a way that is epic but always very slick, and still believable. I think this easily could have become a hyper-real TVC, but actually it stays well-rooted in the world of action movie car chases. 

DM: Thank you Alex! It was certainly a fine balance. The aesthetic mission from my perspective was to make it feel as close to a car chase that could have been shot on real streets as possible. Russian arms and chase cams! This is the outlook Jordan and myself worked to.

I insisted on a speed read out for the cars during animation, just to make sure they weren't trundling along at 60mph.

I love a good car chase, but so often they just don't feel fast enough. Here we had a great balance, to be able to break some of the rules you might run into when shooting in reality, but also consider those shooting techniques while capturing the live-action plates on a miniature level so it feels real.

You want the happy accidents. We got some of those in-camera - the whip pan of the camera, or glint in the lens - and others were added and orchestrated in post. I insisted on a speed read out for the cars during animation, just to make sure they weren't trundling along at 60mph, and more pushing 100+, relative to their scale!

Above: More WIP images from the project.

As a follow up to that, if, in the past, water and hair were the outer edges of where animation and VFX pushed, what are those limits now?

DM: Water and hair were my bread and butter when I started out, before they were really feasible. Nightmare! Now though, I still think digital humans/humanoid creatures and, more often, process and schedule are the limits. We are all experts on human beings, we've all been studying them our whole lives, so we know every detail and micro expression even if we're not aware of it. So, that is a challenge. Capture tech still can't capture it all and dynamic simulation gets you so far, so everything needs the touch of an animator. It's an incredibly hard job.

We're at a point where amazing things are possible, but many projects fall victim to bad planning/timing. 

On the process side, we're at a point where amazing things are possible, but many projects fall victim to bad planning/timing. Far too much being expected in too little time. We couldn't afford to fall into that trap here, which is why we worked so hard to get everyone on board as early as possible. We pushed the available resources as far we could. I think, if this project had been executed in a 'traditional' manner, it could well have fallen very short. 

'Fix it in post' has become a catch-all for bad planning. It should be for those unpredictable and unavoidable situations, or strokes of creativity that happen in the moment. You can never plan for everything, but ideally it should be 'fix it in prep'.

'Fix it in post' has become a catch-all for bad planning.

CB: We’re always working to make things live up to blockbuster expectations with the budgets available. That, and a digital face that looks unquestionably human. Both tough challenges.