The Adventures Of Tintin: visual effects supervisor Joe Letteri interview

When signing up to adapt any much-loved literary character, there’s a certain amount of risk involved. But when Steven Spielberg and producer Peter Jackson announced their intention to make a computer-generated version of The Adventures Of Tintin, the challenge they set themselves seemed almost insurmountable.

Given the tepid reception that greeted previous motion-captured outings such as Robert Zemeckis’ A Christmas Carol, The Polar Express and Beowulf, and the outright hostility that greeted the ill-fated Mars Needs Moms earlier this year, it was widely questioned whether a computer-animated Tintin could encapsulate the warm, joyous spirit of Hergé’s original stories and art.

As it turns out, we needn’t have worried. Not only does The Adventures Of Tintin possess much of the sense of adventure found in the books, but it also stands as Spielberg’s most satisfying mainstream blockbuster since Jurassic Park. Tintin’s a timely reminder, too, of just how good a motion-captured film can look.

As visual effects supervisor, Oscar-winning industry pro Joe Letteri brought all his skills to bear on bringing Tintin into the third dimension. And as The Adventures Of Tintin arrives in UK cinemas, we sat down to chat with Letteri about his work on the film, and the particular challenges it presented.

First of all, congratulations on Tintin. You must be very proud of it.

Yeah, yeah. It came out okay.

People often mention the dreaded words Uncanny Valley in relation to motion capture. How did you manage to avoid that in this instance?

By never thinking about it. [Chuckles] It’s something people focus on in retrospect, but while you’re working, your attention’s on the details. Does it look right? Have we captured all the moments of action on the faces? I think all those things  add up to giving you a proper performance.

When you’re so involved at that minute level, it’s hard to step back and get an accurate first impression any more. So, really the idea is to focus on all the actions over and over again, at all the aspects of performance. You know, does Tintin actually make me feel the same way I feel when I watch Jamie [Bell’s] performance.

But did you feel a weight of expectation, collectively, when adapting books as loved as Hergé’s?

Well, yeah. There was always the sense that we wanted to do this right, but we have that on every film. What’s nice about working with Hergé is that his style is very similar to the way we work as effects artists. We try to make everything as realistic as possible. He’d take something that looks really interesting, like a palace or a motorcycle or a plane, he’d modify it a little to fit his own style, and that’s very similar to the way we work when we make characters or creatures for a fantastic film.

So we just had to apply that one step further in trying to figure what that third dimension was. In a lot of cases, you could go back to his original sources and figure out what that was. For the characters, that was a little bit harder, because he was less realistic. He was a little more elastic when it came to drawing his characters, and they had real style to them. So we had to evolve that a little bit more to get the performances a little bit more with the actors.

This was especially true with Tintin, because from a drawing perspective, he is the simplest character. There’s very little detail in the way Hergé draws him. So we had to turn that into a character that actually looks realistic, that can actually support a performance. When you look at Tintin when he’s drawn, you can’t tell if he’s got cheekbones or not. But if you’re making a realistic character, you can’t make him smile if he doesn’t have cheekbones.

So it’s realistic details like that that we have to start putting in. But once you start putting cheekbones in, does he still look like Tintin? Not quite, so why not? So you have to go through all these changes to make sure that, when you looked at the character, you know that’s Tintin, but he could still do all the expressions that Jamie did.

I thought it was quite interesting, quite playful, that at the start of the film, you don’t see Tintin’s face right away. Instead, you see an artist paint an impression, and then Tintin holds up the picture and says, “There, what do you think?” Was that you saying to the audience, “What do you think? Does he look like Tintin?”

Yeah. Exactly. We really wanted to have a nod to the world of Hergé, but we didn’t just want it to be a film for fans. We wanted it to be, even if you don’t know who Tintin is and never read the books, an enjoyable film. It should also work at that level.

I remember reading a few years ago, I think it was an interview you did for King Kong, you said that when you do facial animation, you still have to use artists, because of the complexity of muscles in the face. How has that come along? Has it evolved a lot since then?

It has evolved. We’ve got a lot better at understanding how the face moves, and the technology for reproducing those movements has gotten a lot better. But we still rely heavily on artists to get it right. It’s still a mix of science and art for everything that we do.

What mixture, would you say, is there between art and science?

It’s all captured, but what you have to understand is, when you ask a computer to translate the motion of an expression, we have to do that by analysing what you see on the outside. We’re trying to see what the muscles do, and then getting the muscles in Tintin’s face do the same thing. That’s a very elaborate process of analysis and refiguring information. So what we’ve done is developed a learning system.

Early on, we’ll program in some ideas of what we think it wants to be, based on what artists will actually sculpt as key frame poses – Tintin smiling, or Tintin happy. Then we compare those to Jamie smiling and Jamie happy, and we put all those things together, and as we start seeing, frame by frame, the different expressions, especially in scenes of dialogue, we ask the computer to interpret what we think we know. And we look at it, and if it’s not right, we add more.

So we build up this whole system of learning, so it’s all completely artist driven. The net result would be that, if you worked on it for years and years and years, you wouldn’t need the artists’ input anymore. But during the course of a movie, you’re only capturing a small part of what a human face can actually do, even though there’s hours’ worth of film. What the human face can do is infinite.

So you always need artists involved in that process to focus those results – even if it’s just standing back and looking at it, and saying, “Technically, this is supposed to be correct, but does it feel the same to me?” And if it doesn’t, we go back and start over again. Sometimes it’s a small adjustment, and sometimes it’s a house of cards that we have to rewind back to the beginning and look at the performance one more time.

So from a technical standpoint, the challenges were reversed on Tintin, when compared to a live-action blockbuster. The intimate moments of dialogue were a greater challenge than the action sequences.

It’s the intimate back-and-forth between the characters is the most difficult, because the difference between an expression gets down to millimetres. If you look at an eyelid, the difference of a few millimetres can make a character look surprised or tired. Whereas with a big set-piece, if you have the choice between three explosions, they all look pretty good – you can just pick the one that’s your favourite.

Was this a bit of a baptism of fire for Steven Spielberg, since this was his first fully CG movie? What was his learning process like, do you think?

His process was going through that translation. Understanding what the actors did, working back from that, and seeing how that relates to your characters.  Because you’re by definition constraining your characters to not be human. Otherwise, if you wanted Jamie Bell exactly, you would have made a live action film and just shot Jamie. So there’s an on-going process of understanding your characters as you go, besides the usual process that actors and directors go through, which is just learning the character as you built the movie.

What struck me, in a film as action-packed as this, is that the explosive sequences didn’t take me out of the film as much as they would if I was watching a live-action movie with lots of CG special effects, because it’s all integrated. Do you think mo-cap is actually the way forward for big action movies?

For doing any kind of character that’s not a human, it’s a great way to mix, because you keep actors in the mix, and it’s easy for directors to work with them in a way they’re used to in a live action film. So it’s a great tool. Obviously, there are moments where you can’t use actors, just like in any live action film – extremely dangerous stunts and things like that, you just have to go all digital with it. But those are more the big action sequences, where you’d probably use CG anyway. But for the subtle drama moments, the interaction between actors, those are the moments a director and actor can really give you on the screen.

And with that, we were sadly out of time. Joe Letteri, thank you very much.The Adventures Of Tintin: The Secret Of The Unicorn is out now in the UK.

Follow Den Of Geek on Twitter right here. And be our Facebook chum here.