[starttext]
Engel: My favorite example is of the limo driving away from us in a residential neighbor hood it’s also in the trailer; it’s a very quick cut in the trailer though. There is so much stuff going on in just the logistics of that; to look for a mile into the distance and see things shaking and crumble. Just to rig that as a miniature you would have to lay so many parts and composite it together. The triggering of say five hundred events that all have to work at hundred percent would be logistically impossible.
Weigert: The same thing in practical effects not just miniatures. They have the same limitations on set if you shoot something. We had that problem we wanted to shoot some real cars and have a few shots where the freeway starts crumbling and cars are tumbling out. We have the parking garage on the right side that crumbles too and cars are tumbling out. We said hey it would be great to have some real cars falling out too. We worked with Mike Vezina on that the special effects supervisor on that and did some tests. It turned out we could only use very little and do very little because it’s so specific because when that parking garage crumble the cars have to tumble out in a very specific angle. You can drop cars from a crane and have cars swirl through the picture but just to have it at it the exact right angle is a huge set up. If you think about how much it costs to have a second unit team shoot this and get maybe two or three takes a day and you’re destroying two or three cars a day with that, after a while it just isn’t worth while anymore. We know we are going to have to build the cars anyway in CG, we say "let’s just do them all in CG."
Engel: It is a little bit of a mix, in the parking garage sequence there is a real car falling and smashing and one is tumbling over. At the same time the practical effects came in handy for everything where we had to interaction with the actors. That’s when Mike Vezina set up this big rig shaky floor that was six thousand square feet, so that the actors did not have to act like they were on a plate form that was wobbling. Wavy motion like a real earthquake was set up. When they went running from out of the house, the house was crumbling behind them, which was all CG, but they were literally running on this floor and getting into the limousine as the floor was shaking. Roland said as a mandate from the beginning "we can’t do like in the old Star Trek movies where they do the Shatner, where they hold on and act. As if the Enterprise is doing something like this". Roland said "No, the actors have to have the fear on their face and it has to be real." That is why that whole floor was built; which proved to have additional challenges for us later in post, like tracking.
Is it cheaper to make a movie of this scale today with the current developments in the software?
Engel: I recently had a discussion with somebody about that. It’s very fascinating that it is absolutely true that that is happening. The question that came up was can you do things cheaper with the software but at the same time while that is happening where the cost goes down the expectations go up. What you have to deliver with this movie it somewhere evens out. You are always on the same track with what you actually have to show and do.
Weigert: Good point because after every project we think the next one is going to be so much easier because now we know how to do it. Then there is a whole new set of challenges because now everybody knows how to do it. We say we want to do something that nobody has seen before. You now are at the same point you were at two years ago, where you have to reinvent something new. Stretch the computing power and software to its absolute limits. If you look at any sequence in this movie there were horrendous render times. Horrendous simulation time and new software development time.
Engel: There was a nice little statistic we did about computing time. It took days to render the early sequence of the earthquake. If we would render that on a single machine it would be sixteen years to actually do that.
Weigert: If you think back, doubling up every eighteen months, wasn’t that long ago but eighteen years ago according to Moore’s law this would have been double that! It would have been thirty-two years and so on. You can think about how long this would have taken just ten years ago. It wouldn’t have been possible at all. A lot of the scenes we have done here wouldn’t have been possible five years ago.
Engel: There were some new developments that were just happening when we were making the movie like Thinking Particles from Cebas, which we actually implemented. We actually worked with a company that built a plug in for 3D studio max. We said we need it earlier than you guys will have it finished. We actually made a deal with the company to hire additional programmers and whatever it takes to get it done early. Then they can use it for their official version. We have it to use on our movie, which nobody else had at that point.
[endtext]
0 comments:
Post a Comment