Big Memory for 3D Animation and Visual Effects

Download this presentation: Big Memory for 3D Animation and Visual Effects

00:01 Dr. Hank Driskill: Hello, I'm Dr. Hank Driskill. I'm the head of CG for Feature Animation at Cinesite, and I'm here to talk to you about the potential for big memory, and specifically some of the work that MemVerge has been doing for animated and CG films. So, first off, a little bit of an overview. CG movies, there's three major steps. One is the story and design step. That is when the idea of the movie is coming together, and that can take years. They spend a lot of time iterating on the idea of the movie and evolving it, and with this goal of making it the best movie it can be.

00:42 DD: At some point, that gets far enough along that people start to feel good that this is a movie that is worth making and worth spending energy on. At that point, you start to bring on a team to start what's called asset production, and that's where they start to build the components of the movie. Especially on a CG animated film, everything you see was crafted by teams of artists, whether it's the rocks and trees in the scene, the characters themselves, the controls to move those characters around, every aspect of it is crafted by artists. And so, in the asset production phase, you start that process of building the world that the movie is going to take place in and building all the characters that are going to be in the movie. Then eventually, when that's far enough along and the story has evolved far enough that you're feeling good, and around that phase, typically, you've locked in a release date or at least a completion date, then you ramp up what's called shop production. And that's when you're actually iterating on the individual scenes of the movie.

01:51 DD: So, when you're making ambitious CGI movies, which of course most people are trying to do, there's some complications in all of that. When we talk about an ambitious CGI movie, we're talking about two things. We're talking about ambitious in terms of story. You're trying to make the most amazing story possible that really touches the hearts of people all over the world. You're also talking about making something with scope and scale and spectacle, something that looks amazing and often involves imagery that hasn't been seen before or hasn't been seen before at the scale you're trying to achieve. So, what that translates into is, on the technical side, you're building new tools and technologies to enable the completion of the film, and that gets very ambitious. And you're often building things no one's ever done before, either to invent some new look or to enable your artists to iterate on an existing look more than they were able to do before.

02:50 DD: Also from a story perspective, because it's an iterative process, this iterating and iterating on the film, evolving it, trying to make it the best story it can be, what that translates into is eventually, you're going to run out of time. And you will turn the crank on the story, trying to improve that story, probably at least one time more than the people actually executing the shots would have wished you had, because it's going to mean they have less time to execute the actual film, and we call that schedule compression. It's when the amount of calendar weeks that you have to make the shot production phase of the film is less than you had originally planned on. And that's when the movie starts to become very expensive very quickly, because you're starting to do late nights and weekends, and your team is burning very, very hot because they have this hard end date to release the film, and they ended up with fewer weeks than they would have liked.

03:48 DD: So, with that, artist time becomes your biggest cost. By then, you have your entire crew on the film that you can have upwards of 300, 350 people working on one of these films. So, you've got 300 people cranking late nights and weekends and burning lots of overtime and working under a lot of stress to try to achieve this very high bar that you've all set for yourselves, because what you're doing is creating collaborative art. And everyone involved is very passionate about what they're creating and their contribution to it, and everybody's trying to hit individually this very high bar so that the collective piece of art is the most amazing thing that they've created. So, any lost time in that becomes catastrophic. Because you're running short on calendar weeks, any time an artist isn't able to be working on their art becomes more and more painful and more and more costly.

04:44 DD: So that's where the work that MemVerge is doing becomes so exciting. Using Intel and the new Optane memory, there's this concept of immediate recovery. So this notion that when an artist is working, they can have these massive scenes loaded into their session, they can be working on these characters with hundreds of animation controls, intricate subtleties of performance, and then because they're working on bleeding edge tools, there's bugs. They're working on tools that are, at best, beta. And so, things crash. And they can sometimes not only have lost whatever time since the last time they saved when something crashes, but the amount of time to recover from that crash, to load everything back into memory and get ready and working again, can sometimes be 45 minutes or an hour.

05:35 DD: And not only that, but there's this mental set, this mindset that an artist gets into when they're cranking along on achieving a performance or achieving a particular look for an environment or whatever, where they get so hyper-focused on that, and when that crashes out from under them, it isn't just the hour to get the thing loaded back into memory, they've lost their groove. They've lost that momentum that they had on the art they were creating. And it takes them time to regain that. And so, it isn't at all uncommon for them to lose a half a day. And as I said, calendar weeks become very precious, especially as you approach the finish line of the movie. And so, any artist losing half a day is catastrophic.

06:16 DD: So this concept of their session crashes and their scene that they were working on and their character performance that they were working on crashes out from under them, and then immediately, they can recover to a snapshot of that that's seconds old or maybe a minute old, but they can immediately get back to work on that and not lose that mindset of the art they're creating, is just spectacular. I'm so excited about the potential of this because not only in terms of the clock time, in terms of the half day of lost productivity, in terms of the stress and strain on the whole other 350 people that are trying to get the movie made, but also just from a morale standpoint, that the artists can . . . Working on bleeding edge tools and having them crash out from under them is just a minor hiccup instead of this major catastrophe. So, it's very, very exciting what's being built, and I'm super excited about the potential of all of this.

07:17 DD: And then we've had some fun conversations about what comes next. All of these pressures have gotten worse and worse over the years because the bar that the artists set for themselves on the films they're creating keeps getting higher and higher, the level of ambition of what they're putting up on screen keeps getting higher and higher, because the story, the creative aspect is getting more ambitious as well. They're trying to create movies that genuinely mean something to the world, that people will want to watch and show their kids and their grandkids in the years and decades to come. That bar that they set for themselves is so high that that pressure will continue to grow. The amount of clock time, calendar time, that you have to produce the film is going to continue to have pressure on it because of the iterating of the story and the tools and technologies are going to continue to be bleeding edge.

08:09 DD: And so any time you can optimize that and any time you can come up with new workflows, new ways for artists to work together at the same time, any time you can take a serial process and turn it into a parallel process where artists can work alongside each other to craft the art.