800 Carnarvon Street, #323 New Westminster V3M 0G3

Cost Control in the CG Department

Hey there Producer. Ever wonder where your profit margin went on that CG-heavy visual effects / animation project? There’s a good chance much of it went to over-engineering; wild, crazy, out of control over-engineering, overbuilding and overthinking. How did this happen?

Mostly it probably happened due to poor oversight of the crew. Likely because you have managers who are not experience CG artists, so when artist say “the flibflabble is conjammed at the fronculous”, they respond anxiously “well….how long will that take?” Or you have managers who may be great artists but have no management training. They are busy trying to be the crew’s buddy or they’re on a power trip or whatever. Either way, the crew pays them little attention.


Over-engineering

 

Here’s a real world example of unchecked over-engineering/overthinking gone wild:

Once upon a time, a CG FX team was assembled with the task of creating mortar explosions for a war film. This was a team of very talented, smart individuals. The show needed several dozen mortar explosions and, of course, each needed to be a little unique. So the crew set to work and six months later presented an incredible explosion tool that could not only make mortar explosions, it could make any kind of explosion. Just by tweaking a few parameters it could be anything from a fire cracker to a nuclear blast. This was a tool that could be used by the company for explosions for years to come. Awesome, no?

Awesome? No. Not really. You see, nobody asked them to build a company-wide tool that could be used on many productions. If they had, it would have involved meetings that calculated how to amortize the cost of development over several productions. It would have involved looking into what projects were coming down the pipe and were likely to need explosions. It would have included a discussion about how often a tool like this would be needed and weighed against how costly it would be to build another, unique tool for another project.
But the crew didn’t ask any of these questions, didn’t tell the management what their development plan was, and in fact didn’t even show anything until the tool was near completion. So instead of a calculated development plan where several projects share cost, one project unexpectedly bore the entire cost of development, and that money had to come from somewhere.

The mortar explosion that was required by the project would have taken one FX TD maybe a month or two total. Instead, five FX TDs were tied up in development for six months. In other words, the tool cost fifteen times what it needed to for the project. Did they use it for anything other than mortars? Nope.

Was it the FX crew’s fault? Certainly not. FX T.D.s work hard in my experience. They mostly love what they do. They are enthusiastic, excited, passionate, smart people. In other words, they need someone overseeing them who will tone down the enthusiasm and just get the job done to specs. In this case, the management team failed by not providing that oversight. Remember, dear production manager, if the team fails for any reason, it’s your fault. It’s not just that you have to do the honourable thing and fall on your sword for your men; it is actually your fault for not setting the team up to succeed.


Building for the project

 

The first and most obvious cost control measure that comes out of this example is to always build specifically for a project. Not just in the FX department but everywhere. If you need a 1957 Chevy model, don’t build an “every-car” that can transform into every model of car in existence. Just build a damn 1957 Chevy. Or better yet, find one for a hundred bucks and upgrade it to the level needed. Sometimes you’ll find one well enough done for your needs and save several thousand dollars with a quick internet search. But I digress. There are many things we do need to build from scratch or alter/rebuild significantly.

Unless you have strong reason to believe you will need more than the basics, just build the basics. If your bidding producer has done a good job, your contract came back with specifics about what you are supposed to be doing for each shot. Knowing the scope of work puts you in a great place. It means you know the extents of the work you need to do. Don’t do more than that. If you aren’t getting specifics, or if nobody is defining the scope of work for you, then you’re in trouble. Your bidding process needs to be looked at.

Any experienced production sup/lead is aware that artists love to go further, sometimes much further than is actually required. It’s really fun to build some of these things. But their job is not merely to build stuff for the show. Commercial artists have a responsibility to be cost-effective and to help protect your budget. They should be aware of the project needs and build to that level. The Supervisor and/or client will tell them if they need to go further. Also, if the client wants you to build much more than your contract specifies, there should be a change order. I know these days many clients expect you to throw in a live unicorn with a bid to make CG dice, but don’t do it. This is the path directly to failure and bankruptcy.

Always build to the needs of the project.

But how do we do it? CG artists are always trying to baffle you with bullshit, sometimes succeeding. How do you combat this? Obviously you need someone to manage the teams who cannot be baffled by bullshit. An experienced artist can do this. One of your senior artists would make a perfect team lead or department supervisor…but only if they have real management training. You should always be looking at your team to see who you can throw into team lead positions so they can get their feet wet with low-level management skills. Then you should get them real management training from professional management trainers. Anecdotal training, mentorship from other senior artists without management training will not do the trick. Grow your own awesome mid-level managers. They’ll keep the bullshit to a minimum.

Refining cost control: Build to the shot

There’s a second level of cost control and this one requires the expertise of someone who knows how to build shots and understands the CG process in detail. This level involves not only building to the needs of the project, but also building to the needs of individual shots.

For example, if you need a bunch of 1920’s cars built and you look through all the boards, finding that two of them get pretty close to camera, but the rest of them are parked in the background, then fairly obviously, only two of them should be built to a close-up level of detail. The rest do not require such high-level detail. Many details can be painted or faked in many ways.

Furthermore, if the hero cars are seen driving past camera, you almost certainly do not need the engine, transmission, axles, brakes, and most of the interior built. Simply do not build things that will not be seen in the shot.
Granted, we don’t always know precisely what will be in every shot when we begin building our assets, however we can make a reasonable guess. And you know what, if it turns out we need an extra few details, we can always add them as needed later! Magical!

There is one magical ingredient needed to succeed at this level of cost control. Good planning. Yes, you’ll need to know every shot in detail. It will take a good deal of homework, but in the end, you’ll be the master of your budget.


Arguments against building for the project / shot

 

I’ve had this debate more times than I can remember. Arguments against usually revolve around “but what if the director later decides to do a shot where he pushes the camera up the tailpipe, through the engine cylinders, out the carburetor, through the firewall and into the driver’s seat? If we don’t build the whole car to detail, we won’t be able to do that shot. Also, if we do build the whole car to detail, then we can respond to client requests like this very quickly and we’ll be superstars.

Well sure. Bankrupt superstars.


Counter-argument

 

Once again, if there’s a change that big, it should be a change order with a new budget and schedule. We don’t magically pull an entire engine interior out of our asses.

Speaking to the philosophy about building detailed assets for everything “just in case”, let’s do a little math. If I have a hundred assets in a project and I build all of them to detail when I really only need two high-resolution assets, it will probably cost me twenty or thirty times as much to build them. Now, let’s say that the director does actually ask for the engine shot. Yay us, we turned that around in a couple of days instead of a couple of weeks. But the cost of that feather in our cap (which the client almost certainly doesn’t care about) is that we had to build every single asset to a high level of detail. So, in fact, very little gain for very huge expense. Furthermore the client now expects you to respond to ridiculous requests in unreasonably short time frames. I’m not sure that’s a good business model. The reality is if the client adds something big to your scope of work, they don’t expect you to turn it around instantly. They may ask for it instantly. They may beg for it instantly. They may insist that you turn it around instantly. But they don’t actually expect it instantly.

It is too easy for CG asset builds (and every other aspect of production) to spiral out of control unless it has oversight from competent people. A business degree alone does not fulfill that need. It’s very useful in other aspects of production, to be sure. But on the production floor, team management needs to be able to sort the fronculous from the flibflabble.

Nicholas Boughen is a veteran VFX digital artist and supervisor and founder of CG Masters School of 3D Animation and VFX. He runs the only training program that provides advanced, comprehensive vfx & animation production training in a studio environment.