The tools and techniques that gave the sequel its hand-crafted look. With EXCLUSIVE breakdown videos.
At VIEW Conference 2025 in Turin, befores & afters got the chance to sit down with The Bad Guys 2 director Pierre Perifel and DreamWorks Animation visual effects supervisor Matt Baer to discuss the level of stylization in the film.
DreamWorks had, of course, already established a look in The Bad Guys, but were able to take it further for the sequel. Here, Perifel and Baer break down what lessons they took into the new film, and what new approaches they could adopt in terms of cinematography, procedural tools and enhancing the mix of 2D and 3D techniques.
b&a: I’d love to go back to the beginning and talk about the earliest conversations that were had about giving the first film a stylized look. As an industry, we have moved to some stylized looks for a bunch of films now and DreamWorks has been doing amazing stuff in this area. But I wondered whether there was any hesitation for this kind of approach on the first Bad Guys?
Pierre Perifel: Not from me [laughs]. When I saw the cover of Aaron Blabey’s book, it immediately elicited the idea of these guys driving cars and robbing banks with suits, like Reservoir Dogs mixed with The Blues Brothers. I was like, it has to be stylized. I’ve always wanted to see a stylized CG film. Spider-Verse came a couple of years before and opened the doors for us. We then said, alright, it can be done. I come from 2D animation. I really wanted to see CG done differently, but not so pushed and so crazy that the audience would be lost. So, it was always there brewing inside of my mind.
b&a: By the same token, from a more technical level, was there any trepidation at DreamWorks about whether you could actually achieve this level of stylization?
Matt Baer: Well, each film or each franchise at DreamWorks has its own look. Our artists are definitely used to jumping onto a show and shedding the techniques and aesthetic from the previous film and ramping up. I think that is an exciting part for all the artists. Everybody’s up for the challenge. It was really our first film at the studio where we were going to push a ‘simplicity’ idea or a graphic idea even further.
The biggest challenge was time. We didn’t have a lot of time to develop the look when the artists got on and we knew that we weren’t quite sure what the eventual look would be. So the strategy for implementing that look had to have the ability to adapt six months later after we started it. We wanted to keep this cohesive look, but we didn’t know what it was going to be. We knew that we would probably settle on the look after we finished about 30% of the film, but we would be given no means to go back and change anything.
So we went in with a strategy that is held true today, which is, we want to get as much out of the renderer as possible. We were worried that just a 2D approach would mean the gag would start to show later on. So we went in with that strategy and that served us well because it allowed us to quickly iterate in lighting. We were able to pull a lot out of the render, but we had enough tools in the comp to really refine things to get the look that Pierre really wanted. A lot of it was a scheduling challenge, working out which parts of the look needed to lock in first. Others could come in a little bit later.
We created some new roles on the film, basically making space for some look developers who are not reporting to any department—they don’t have any artists working for them and they’re basically innovating all day and bridging and building brand new tools.
b&a: Pierre, did you feel on both films that you needed to communicate differently with artists that were more used to just 3D animation?
Pierre Perifel: On the first move definitely there was a shift. It was a little bit of a learning curve for the teams. I felt it mostly in the animation department. That style of animation was really new for them. It was anime-inspired. It was using less live-action reference. Animators got really used in the past few years before that to really use real life reference, or vidref. I was asking them just to lose those poses, and craft new poses and push them a bit further. One thing was the eyes, even, which are trickier to pose than the typical eyes we usually have. We also had to think of things a little differently in lighting and effects.
Matt Baer: Yeah, there’s certain tricks that you can’t use anymore. If you think of the effects department, well, they’ve become better and better at 2D effects these days. Instead of simulating things, they realize that they can just paint the frames and be done. I mean, previously, say you’re doing some steam and it needs to fade out, or you’re simulating some dust behind the car, usually we use opacity to let it slowly fade out. That’s a trick we didn’t want to use anymore. We want this to feel 2D, and that means we had to change the rules. Now they would have to figure out ways that a 2D artist would use to carve out dissipation shapes, for example.
Pierre Perifel: These were all issues we faced on the first movie. For the second one, a lot of the team was the same, so the ramp up was a lot less difficult. That allowed us to build up on what they already knew. The one challenge, though, was that we had a Sony Imageworks team helping us out, and so they had to catch up. Of course, we already had a full movie to show them as reference that they could match to.
b&a: Armed with the experience you had on the first film, did that now let you change the workflow? Could you proceduralize things more? Maybe on the first film a lot of something was done in comp, but now it could be done instead in the render or in lighting? Is that what happened?
Matt Baer: Yes, we did that. At DreamWorks, because we’re moving from show to show and these all need different looks, a lot of our tool strategy involves building modules that can be reconfigured. So when you go onto a new show, you can actually come up with something unique, but without reinventing everything. For The Bad Guys 2, because it was a way more ambitious film than the first, we wanted to make sure our artists had time to continually bring new ideas. On the first movie, our artists did keep adding a lot of new ideas but they turned out to be more manual than we had anticipated. Little things like just the rim light on Wolf’s shoulder. We always wanted a little bit of hatch lines going through that. Well, that became a very per-shot manual task. Or, for character effects, the folds in the clothing have this kind of chunky rounded shape—almost all those are hand-sculpted on the first show.
On this new movie, each department went through what they had previously done and said, okay, this is going to be a bigger show, how are we going to do more of each thing? We took a lot of those things that became more manual, we analyzed those, and we said, can we fold those back into the library setup? For example, we rebuilt the simulation setups for all the clothing and the folds. Now, automatically,
80% of them come out with folds that Pierre, JP and the head of character animation really, really liked. That allowed the character effects artists to then add additional ideas on top. That went for every single department.”
b&a: There’s also something heightened in this film in terms of camera work, speed ramps, the use of anamorphic lenses, etc. What was that process like for you, Pierre, in referencing examples of live-action to include in the film?
Pierre Perifel: Because we already had a stylization of the rendering, the effects, the animation and the character design, the idea was that on top we could build in a sophistication to the camera work and the lensing. That was done mostly by referencing all the live-action films that we love that were either speaking to the genre we’re playing in like Mission: Impossible or James Bond. Or individual sequences could have their own little winks to something else.
One of the examples that I often give is that scene where they arrive in their crappy car and the whole car collapses. I literally gave a reference from a movie that I loved as a kid. It’s an old French film, Le Corniaud, where the guy is driving a Citroën 2CV and is bumped into another car and the whole thing crumbles.
b&a: Matt, in your talk here at VIEW, you mentioned the implementation of some of those live-action lensing techniques is about taking the ‘CG curse’ off a 3D animated film.
Matt Baer: Yes. That makes me think of an anecdote from the first film. We were in some technical reviews and we were looking at the footage and thinking, oh, it looks too clean, there’s something different about it from what Pierre had earlier seen and really liked. And we realized that what he had seen earlier were lighting renders that had not sampled the image completely, meaning there was a little bit of render noise in it, and that’s what had been approved. The technical render later had made everything too clean.
What Pierre had been perceiving in the lighting renders was film grain, basically. It really led to a whole bunch of conversations about what you can do to break the CG curse. We do it in all the departments in different ways. So, in models, I mean, there’s not really a straight line in any of the models. On top of that, we used anamorphic lenses. The Batman came out and they showed how grimy some of those lenses were, I just thought that was wonderful, and how can we bring more of that look into our film?
Getting back to the film grain, we actually did figure out a way to add that in. What’s really nice is when you end up with something that’s really clean, say a really clean grad that instantly feels too clean, just by putting that film grain on it, it breaks it up just enough.
We really relied on the artists from each department to take their part in all this. One of the notes we would say in lighting dailies was that a shot might be missing just a couple more aspects of that illustrative touch. What we did was, we went and observed how our production designer and our art director actually worked. Our goal was for our artists to get into that same mindset. We’d say, if you just went back into Photoshop, what would you do? Are the reflections too real? Are the glints too real? And they would go back in and do what you might do in one of those tools to make it more illustrative.
b&a: You’re reminding me of a period in CG feature animation filmmaking where the ‘art of’ book for an animated film would come out and I would buy it and it was all beautiful, stylized or painterly. And then the film—which certainly still looked amazing—didn’t actually look like the artwork. And I’d like, oh, I wish it looked like the art.
Pierre Perifel: Dude, you’re a thousand percent right! I think that’s where my desire of stylizing it more came from, too. I was like you, frustrated. And honestly, with our studio, too. The artists were so talented and in the end, the result felt synthetic, it didn’t feel organic or as pretty as we had got to in the artwork.
b&a: I would even say sometimes the results didn’t feel like a person had touched it, like a computer had done the work, even though I know how much artistic work had been done.
Pierre Perifel: Yes, there’d be so much work behind it, and yet you watch it at the end and you’re like, wow, I would’ve loved the style from the art of book instead. So, getting the final result in these films closer to the art of book was definitely something we set out to do.
b&a: I think an example of the great combination of styles is the rocket launch. I was smiling when I saw the 2D effects work for the smoke and fire, which feels hand-crafted, but it’s also nicely combined with effects simulations and 3D work. This doesn’t seem easy—how did you accomplish that here?
Matt Baer: On the first film, we didn’t do a lot of bespoke 2D effects animation and shots. On this second film we said, let’s do way more if we can. Now, we don’t have a lot of pure 2D effects artists available, so it really became a matter of, how do we allow them to do this? What we would do is partner them with some of our really great CG artists and they’d be collaborating together.
b&a: Because there was so much of that 2D work, you mentioned in the presentation that you were planning ahead with that rocket launch sequence a lot more than usual in a sort of a pre-previs, and sticking to the set camera angles way earlier than normal so that you could lock in the 2D look. What was that like for you, Pierre, to approach it that way?
Pierre Perifel: Well, what was interesting is that I wanted to shoot the rocket taking off in a very traditional way. We were really inspired by NASA footage of rockets taking off and all these big films that we love, like Apollo 13, where they really do have a lot of fixed cameras anyway.
b&a: The other thing I loved in the rocket launch sequence was where you could really go hard on smears, like when they’re leaping between areas outside of the rocket, or Wolf is spinning Snake, to give that sense of dynamism and motion. Tell me more about that side of the work—I feel like it’s something you wouldn’t necessarily notice, but if it wasn’t there, you would notice it was missing.
Pierre Perifel: Oh totally. It’s exactly there. What it does, again, is it brings very much a sense of traditional craftmanship. You’re feeling it. You’re not supposed to notice it, but you feel it. JP Sans, who co-directed this film, said very early on, let’s really reduce the amount of motion blur so that it doesn’t feel as CG. It lets you read every single pose really well, but then in order to not strobe, we would use traditional smears. We had to train our animators to do this because it’s not in their classic skillset anymore. When you animate in CG, the motion blur is there automatically for you. The smears are more of a traditional animation trick.
Matt Baer: This approach really helps with the graphic quality of it. It keeps the silhouettes really clean. It gives another sense of the illustrator’s hand. Now, because it is intended to look hand-drawn, the trick is to make sure every one of those lines feels like they’re all done by the same artist. They all have to be on-model. On the first film, JP did all of them! On the second one, we built a post-lighting 2D effects workflow. We used it for anim lines, a lot of the tears, the drips, sweat that you see in the rocket launch—all that was hand-drawn as well.
Pierre Perifel: It really just had to feel hand-crafted. Just seeing the hand of the artist directly on screen, because that’s literally what it is, that is, somebody just spending the time drawing over the frame.
