Pixar's Sanjay Bakshi Talks Bringing 'The Good Dinosaur' To Life

Pixar's Sanjay Bakshi Talks Bringing 'The Good Dinosaur' To Life

Pixar's Sanjay Bakshi visited Emory University in Atlanta, GA to deliver a talk about animation. He is currently the Supervising Technical Director of Disney and Pixar's latest film, The Good Dinosaur. We had the opportunity to sit down with Bakshi after his lecture to talk about the various technology advancements The Good Dinosaur used to bring its characters and environments to life.


Shakefire (SF): You came into the film two years after production began. What was it like coming into the project after it had already been established?

Sanjay Bakshi (SB): A lot of our movies go through changes. Actually, in the last few years I would say most of them have. It’s not uncommon. I think part of it is because to make the movies takes so long so a person who originally started out is the right person to come up with the idea and nurture it but maybe not the right person to execute the film. For this film, a few of us came in to help out. For me I really wanted to be a part of it because I wanted to help out. Pixar, I felt like, needed to have some new energy on that film and some people who didn’t have all the legacy baggage. I was actually really excited to be a part of it. I looked at it as an opportunity to help out. I was able to come in fresh and didn’t have any of those hangups and tried, as a leader, tried to just say, “Pixar has been successful so many times before, it’s going to be successful again. Let’s not worry about any of the past histories.” That was my attitude towards that.


SF: Pixar films always ask “What If?” What’s the “what if” of The Good Dinosaur?

SB: The dinosaur film has been done before, and we wanted to do a different take on it. That idea is really interesting of if that meteor hadn’t hit Earth and destroyed the dinosaurs and humans hadn’t been able to flourish as scientists talk about; the absence of the dinosaurs was an opportunity for mammals to flourish. If it kinda switched. If the dinosaurs developed language and culture and farming and the humans weren’t able to fill that place in the social structure than what would those relationships be like? We’re being quite mysterious about the humans. We’re not delving into it too much. They’re not necessarily cave people. We don’t really know a lot about them and I think that’s also the fun of it. Then the film really becomes about the relationship between this specific and this specific human.



SF: What’s your process for creating the emotion of these characters? How do you capture that perfect balance of human and animal? Do you constantly look at a mirror to see what these emotions look like?

SB: It’s really that. The animators are essentially actors. Their job is to convey their performance. Even more than the voice actors I would say the animators are really the actors. They spend time in front of mirrors. Every animator will have a mirror beside their desk. They’ll have a room they can go to to record themselves to convey their performance. If it’s a physical performance then they’re acting it out and physically remembering the weight and the attitudes of the movements as the scene is being performed. This film is really unique in that there’s not a lot of dialogue so those characters have to be so expressive and you have to be able to convey so much through just their facial expressions and their body movements. It’s so satisfying to see a sequence like with Spot, which has essentially no dialogue, and to see an audience get sucked into that physical performance and feel empathy for Spot and for Arlo. The animators, they’re masters of being able to convey emotion through acting.


SF: You’ve said that a lot of the animation for the setting is procedural in that you can create this one element and manipulate it to fit the need of the scene. Were the characters done in the same manner?

SB: Typically the main characters that are really performing, they’ll all be hand crafted, puppeteering, very careful frame-to-frame animation. But we do use procedural techniques, and we did on this film for the herds of the bisodon we call them, the cattle. When you have hundreds or thousands of characters, we’ll animate one of them doing a run cycle and then procedurally apply it and vary it so an animator doesn’t have to animate each single character.


Another sequence that uses this is the birds that are flying and reacting to Arlo and Spot running through them. We developed a flocking algorithm to procedurally animate them because they would just be too hard to animate by hand. Then we’ll pick out certain ones to put that extra love in, you know. It’s funny how you can kinda, maybe not fool the viewer is the right word, but just give the sense that all of them are carefully animated. That’s the trick of the crowds, too.


SF: Can you talk about using real geographical surveys to make the setting for the film?

SB: The first thing Peter Sohn told me when I got on the movie is, “I don’t want this movie to feel like a walk in the park.” I didn’t really understand what he meant. Over conversations, what I think he really meant was he didn’t want the audience to know that Arlo is going to walk through the forest and that’s the path he’s going to take because everything had been set up to make that easy for the animator to make him walk along that specific path. We really wanted the natural world to feel like nature does; unorganized and chaotic. That’s one thing I think he meant.


The second was that if I’m shooting Arlo from this angle and I move the camera over here it might expose this miles and miles of terrain like in the natural world. That’s what’s so cool about hiking up a cliff and seeing this beautiful vista. He wanted to have those opportunities in the movie. To not necessary plan out the shot so carefully but to be able to move the camera and be like, “wow, that’s amazing! Let’s shoot that.” That’s so hard in computer graphics since we hand build everything that goes into the movies.


Because of those desires we knew we had to do something different. I don’t think it’s the first movie that’s done this. It’s the first Pixar movie that’s done it for sure, and I think it’s the first movie that’s done it as extensively as we have, but it’s a real rich resource. The data itself varies in quality depending on the region. The data that we downloaded I think it’s only one data point every one meter, for instance. It’s actually quite lumpy, but we then put all of this other detail on top of it like the rocks and just a little bit of displacement to make it feel more detailed than the actual data is.


The cool thing that happens when you use real data is you get these side benefits. Like when Arlo and Spot are having a moment together, the terrain is actual terrain so they won’t be on this flat plain that we would naturally build and would put in a few bumps. But it might be a 30 degree angle and animators have to react to that. You might not notice it when you’re watching it, but hopefully the natural world feels even more real because it’s all actual, physical terrain.



SF: How did you find the right balance between these realistic environments and these stylized characters?

SB: It’s a real challenge for us, and we were worried about it because we wanted to make these terrains out of necessity. You have to be there with Arlo. When he’s in the wilderness and he’s alone, you have to feel empathy for him and you have to understand his situation. The natural world has to feel like something we’re familiar with. The scale of all the trees, the scale of the bushes, the leaves, we tried to get that all right so that it feels immersive and real to you.


Then we always caricature our characters because we want them to be able to do non-physical things. We want to be able to use squash and stretch to make the animation feel, at times, comical and funny and use all of the techniques at our disposal. It would look really weird if Arlo was a Jurassic Park level of detail dinosaur doing these outrageous things. Sometimes he’s leaping over rocks and his body’s stretching out in a cartoony way. And he is so expressive in his face. He goes from this tiny little mouth to this huge mouth. I view it as my job to marry those two. That’s why whenever we had to kick up dust from them walking through dirt I was like, “Let’s do it.” because that will really ground them. Whenever they’re moving through vegetation. Let’s do it. I didn’t shy away from that stuff like we do on some movies because I felt that was a tool we had at our disposal to tie them together.


And then the lighting. We really do light them together. They respond to the same lights and hopefully that grounds them as well. I think after you watch the movie you’ll get sucked into it. This is my hope. I don’t think you’re thinking about that dichotomy. I think you’re feeling, hopefully, like it’s a unified visual experience.


SF: When a live-action film is shot, there’s usually changes to the story. You’ve mentioned before it’s the same with Pixar animation in that you’re evolving the story as you’re going along.

SB: I think that’s the biggest challenge for the technical crew, but also if we can do it right we’re providing a real luxury to the story team to have the opportunity to make the film as good as it can be. We’re really careful about when we put something into production we make sure that we feel like it can go into production. And if it can’t, not putting it into production. Invariably, we’ll put something into production and it will still change. The story will still change. Because we have this layered approach where we do layout, basically, and that’s where we just do rough staging of the characters and the camera. We can do that and throw it away and redo that pretty quickly. Boarding is obviously the easiest place to do that layout and as it moves down the pipe it gets more and more expensive. Because it does take a while to get it through all the different departments it actually affords you even more time. You could say, “Okay, this is ready to go into layout” and then you still have like six weeks to make a change. Then it’s ready to go into animation and we do rough blocking. If you change that scene it’s probably okay. We’ve gotten really good about having those conversations and having like a handshake between director and technical crew. Invariably things blow up and we have to change them, but we also are good about honoring people’s work when they’re shot and cut and explaining why.


SF: Were there any scenes in particular that you were really excited about but then eventually got cut?

SB: There’s a scene where they’re herding the bisodon that was out and then in and then out and then eventually got back in and I was thankful because it was this really beautiful scene of thousands of cattle. So that was one that was out for the longest time and at the last minute got back in because there was a good story reason to have it back in. I think we were pretty lean. I was in the story room sometimes accessing technical things, and I pitched a couple gags which I usually don’t and none of them made it into the movie. Pete at that time was like, “Yeah, that’s a good idea” and they boarded it up. I was like, “Wow, I might actually get a tiny little joke in the movie.” Then over the course of the movie it evaporated because of lots of reasons. So yeah, in that way I did experience something getting cut I would have liked to be in the movie. Overall, we’re pretty lean. It was done essentially in two years; production was done. That kinda tells you we were pretty careful about what we put on the screen.



SF: Since you’re designing everything from scratch do you ever get the perfect shot or perfect emotion? How much retooling do you go though?

SB: Yeah, there’s no shot in the movie that’s perfect and that satisfies everybody. We have this process where we fill a room, and we watch it. We encourage people to point out flaws. Left to their own devices, we would show it, point out flaws, bring it back the next day, point out flaws, and we would do that…


SF: And we would never see the movie.

SB: Exactly. So we have to apply this discipline. We use principles that filmmakers have been using for decades of drawing the eye. Where you eye should be over there so that little intersection over there let’s let it go. So we’re always trying to view the movie as an audience member would as well. There’s moments in every movie we’ve made that I’ve been a part of where I still cringe like, “Oh man that could have been better.” I still see that problem, and I expect the same from this movie.


SF: When technology gets faster and better usually costs come down, but in Pixar’s case the costs are actually going up. Why do you think that is?

SB: I think the appetite for just visual spectacle and making the films as visually rich and beautifully as possible is not diminished at all. Not only the story people and the art directors are striving for that. The technical crew wants those challenges, too. There’s always a new challenge we want to take on. We haven’t reached where we’ve done everything we want to do. I think every film is a step forward in some filming of it, but I know the films that are in production now and some of them are going to be super hard and require even more processors to match the desires of the crew to make a more beautiful experience. The next film is using a whole new rendering pipeline that even models the physicality of light bouncing even more accurately. That’s going to require even more processing power.


There’s something to that. When you go to a Pixar film you expect a great story, an emotional story, a story you can relate to, and you also expect to be wowed visually. I think that’s part of the experience.\


The Good Dinosaur opens in theaters on November 25, 2015.

Matt Rodriguez
Interview by Matt Rodriguez
Follow him @ Twitter
Friend him @ Facebook