Bradford Animation Festival (year 2)

Venessa Boyce – 15/8 2:30

CG supervisor Double Negative Visual Effects

Was one of the rare people new from early age wanted to do visual effects

Watched documentary at 15 on Jurassic Park

Graduated from Bournemouth

Got job as matchmover

Worked on Inception, Harry Potter, Total Recall and Skyfall,

Total Recall - director Lenny Wisewood

For Chris Nolan films (for example) there’s beautiful photography to reference with, with total recall there wasn’t much photography.

Very complex shots

Had to create two overcrowded cities, lots of big buildings, power and wealth. Forced to build up into the sky

No real city was filmed apart close up shots.

Had to adapt a lot of the film that was supposed to be based in China, due to the Chinese not liking it.

Apart from 1 lift that was made, everything else was made in full CG.

Had one year to do this

It’s one thing putting something real into cv environment

It’s Another putting cg into real environment  - two different challenges

(UFB) it was meant to be near classical

The director didn’t want glass skyscrapers due to the feeling of them being clustered together on big supporters. It should be magnetic cars hovering on magnetic roads, no trees, no animals, had to be quite serial. Roads and buildings needed making and designing

The car chase; full size cars built, hard to come up for a design for a city when your constantly going through it.

Looked at concept art, was too futuristic, didn’t want tech for the sake of tech. it wasn’t a near classical look he was after.

Pre-vis was useful, 3d storyboarding, directors are using it more today, not just for cg but is an easier way to work for camera moves and for actors to plan everything out.

Everything Double Negative do now works with pre-viz from ‘the third floor'.

Next stage asking director for specific notes from pre-viz. Deciding whats good and whats not. Director looking for negative space, likes unusual edges, nothing too straight.

Went and took references of buildings in London. Always start with photography, more info than concept art.

From these buildings the director like the pics of, they then went to build the environment

Take model of building, chop it up into bits, put into City engine

Constructing city engine

Generate building

Alter the lot shape

Select building and alter perimeters

Building updates dynamically with changes (wasn’t sure what city would look like needed something very flexible)

Designing interesting building shapes

Doing this manually would be terribly time consuming

Extending the process to city blocks

This was a new approach that’s now being used on the new superman

Started mucking around putting buildings in the sky, had to figure out a real shot to work on that could develop ideas in. Had to work blind in sense what final shots would look like from the car scene

Used the third floor pre-viz to help build set in 3D for car chase scene

Put buildings upside down, and underneath buildings

Lighting always backlit

Used very cheap rendering to figure out answers to questions like how quick are the cars going to be etc.

Stones of buildings look very grey scale even though they’ve been textured, so had to start building more interesting brickwork

Start to animate basics, play around with road signs, add more textures and coloures. Director wanted it to be more british, so added in road signs and speed cameras but director didn’t want it. Used holographic road signs instead.

Image from background, photographic background was from New York. Had to make sure their buildings was sat in that world.

Had to build loads of assets for a full City. Team of 4/5. Decided to use trees because City looked too dead.

Next stage, constructing the city, using lego. Putting it together.

Filming was shot in Toronto on an airfield. Director wanted a natural car chase. Made 6 mag cars and stuntmen. Double Negative didn’t have to worry about filming or adding cars it was already done for them. Someone manually had to cut out the cars. Replace reflection in the windows. Had to render CG cars so they could get reflections, and shadows. Lots of CG car involved in every shot

Mapping the roads of UFB

Had to map out the car chase from top of head.

When stuntmen had to make roads fit the movements of the cars.

Designing a shot: director left it up to Double Negative to decide where buildings would fit alongside the road.

Used a shot, and designed on photoshop what the buildings would look like with the cars. The director didn’t like it because it wasn’t backlit.

A month to go and the textures weren’t holding up. In general everything was looking not very nice and just very unfinished.

At this point gets quite difficult to progress, all elements starting to be in place but not looking right

Figuring out composition of image and layer and depth hasing, try get it into shot. Throwing in billboards etc ruined it and felt too cluttered. Took decision to strip back a lot of what they put in. Re-lit it so sunlight was stronger, relighting it helped catch a lot more nice edges, made puddles look nicer. That detail brings the city to life. The DOP added customary lens flare (not their choice) which they didn’t like.

Director was concerned there was a clear connection to the road so the audience would know the cars looked magnetic

Buildings were bigger than traditional buildings (3 times the size)

Took environment to speed past the car because the cars were too slow (30mph after accident on set, health and safety)

Ideas of tunnels, helped with speed element,

After a render they figured out the Tunnels were too dark, and cars were lit in sunlight. Bad compositing, couldn’t look realistic,

So put florescent light in the tunnels, punch more holes in tunnel to expose to sunlight, and changed the position of the sun in the environment

Storm troopers worked for police. Original idea was guys in robot suits and change their waists to CG.

Director was keen to go with the approach or real actors to act with something, and was able to direct them himself and would look more realistic than cg characters. So idea was to replace the stomachs, and would have been easier to make characters with cg.

Problem with cg design was quite easy to mistake guys in the suits for real guys in suits. Wanted to show off that it wasn’t a man in a suit but it looked too much like a man in a suit, very specific. Concept artist had to come up with a few similar ideas. Concept artist, rigger and modeler had to work closely. The man who created iron man helped out. Silhouette couldn’t change too much from original idea of character. Director wanted to be able to see through he spine of the man in the suit so it looked like a robot.

The rig would be driven by snapping the panels on the real and fake character. The idea was to replace the stomach so the cg showed and made it look less human like.  After experimenting with the full CG’s the director said it looked so realistic that they went with the full CG in the end.

Two different worlds and challenges, characters and environments.

Q & A

I don’t keep up with technology, plenty of people who know about software that’s not me, yes there are people that know more than her that she manages and has to supervise but its the nature, theres no pressure on me to know everything about the cg world. The role of cg supervisor isn’t job everyone wants to do, you lose getting stuck into the practical.

Working relationship with cinematographer Paul Cameron, there were conversations with him but wasn’t as collaborative was more like 2 separate jobs, he had his ways on shooting. Sometimes I don’t even know the name of DOP

Best way is always to have photography, if it was up to me I would have taken pics of a city and adapted it, felt more limits making it from scratch. More Cg involved in miniatures, I wouldn’t chose miniatures over visual effects.

Had 4 concept artists for the duration of the year, everything needed concepting. Concept artists were always being asked to come up with new things where as usually there’s one concept artist at the start used for only a couple of weeks

No consideration for motion capture as far as I was aware for the synths, was always going to be real costumes. I was worried CG wouldn’t work for characters and cars. The environemnt is harder to make in Cg

1200 shots used, 1 year to make.

My role is to fit right people with the task and to try make sure the people understand what to do and how their going to do it.  To create and make teams. Downsides you to turn into ‘Mr Manager’, and to have to catch problems before they become disasters. I like the problem solving part and have an over vue of operation.