I’ve been a bit absent from BMD over the past couple weeks. The reason is largely that my life was taken over by a partially-improvised theatre show called Hamlet: The Video Game: The Stage Show, for which I created over a hundred video cues. It was directed by Simon Peacock, voice and mo-cap director for games such as the Assassin’s Creed and Deus Ex series, Outlast, and Far Cry 3, and a former member of New Zealand improv company The Court Jesters (to which I belong). Not wanting to pass up an interview opportunity, I sat down with Simon and his wife Jennifer Seguin (also a video games director and actor) to chat about directing craft, the role of drama in games, and ruining Shakespeare with game tropes.
How does directing for video games compare to other media? You’re directing for screen, but it also has a lot of other elements.
Simon Peacock: As far as the screen side of it is concerned, like in motion capture, it’s a little bit different. I’m called the “performance director.” I just concentrate on the actors’ performances - also making sure they’re consistent with what’s done in the voice studio, that our dialects remain consistent. On Deus Ex, there’s David Hubert, the cinematics director. He knows how he’s gonna shoot it when he gets it back in the studio, so I don’t have to worry about any of that. For me it’s like, “are the actors hitting their notes?”
But it’s different from film in that it’s just so damn fast. With motion capture, it’s two or three takes and you’re done, because you don’t have to change cameras, you don’t have to touch up makeup or costumes, you don’t have to rearrange lighting. None of that stuff exists, right?
It’s rough on actors as well. If they haven’t done motion capture before, they’ll come in thinking, “I’ll probably do two or three pages of dialogue today,” but it’s like, “no, you’re gonna do 30 or 40 pages of dialogue today.”
How does performance directing relate to game development, as things change and get cut along the way - even late in the process?
It results in scripts that can be really last-minute. There are days when I turn up in the studio to direct something that I haven’t even read yet. It comes down to different companies, too. Some companies will do a lot at the last minute. But one of the nice things about working on Deus Ex is that they’ve been recording for two years now, whilst doing testing and that sort of stuff. It’s a weird industry. There is no standard operating procedure. Even in a company like Ubisoft, there’s no systemic way that we go about it. Each project will develop its own way of doing things. You never know quite what you’re going to be doing.
Jennifer Seguin: I find the reaction of actors is much different to working in any other mileu. They have no choice. They have to do it, no questions asked - be ready to do what the director says, have full trust, and commit to it. When you have something that has more time, actors tend to be more involved in the creation process of the character. But [in games] they have to remember that they’re just one part of a character that’s being built by a lot of other people.
SP: Very very few of the actors will see a full script. Maybe a couple of lead actors, and that’s it. The rest are just getting their scenes - sometimes just their lines within the scene. That becomes a big part of the directing - a huge part of it is context, understanding where it fits in the game, and keeping it consistent with the other performers.
Are actors usually directed separately from each other?
For voice, almost always. For motion capture, obviously you can bring them all in together. That’s half the point of motion capture. A lot of it comes down to the way agreements exist with actors in the unions, as to how long you can book them for. Overlapping becomes really expensive and difficult in a voice studio. Even in motion capture, if you’re hiring someone who’s got maybe a six-line role in the game, you’re probably going to have them do motion capture for six other characters while they’re there, and then you’ll just re-voice them later.
How is performance affected by the technological aspect of mo-cap and games - the fact that you never see the actual actors on screen?
We talk about developing a style for acting called “hyper-real.” For all that technology - facial capture and motion capture is better than it’s ever been - there’s still some stuff that it doesn’t get across. There’s subtleties and nuances to body language and facial expressions that just don’t get captured - especially eyes. You look at video game eyes, there’s never anything in them. They’re always dead and glassy. So that means that information we assume we can take from one another, that you can get from film and theatre, you’re not getting from a video game performance. Everything has to be slightly exaggerated. Especially in the voice studio, everything’s a little heightened and raised, so without the subtleties of the face, you can still get the emotional content across, and you understand what’s happening to the character.
You talked about AI loops. Background NPC barks and other incidental shit has always been fascinating to me - what’s it like churning those out?
There are so many. Sometimes there’ll be 20,000 of those in the game, and we’ve got two months to get them all recorded. So you can’t spend five minutes on any line polishing it. Most actors get two takes at each line. If after two takes you haven’t nailed it, that’s when you get direction. If after the third take, you still haven’t nailed it, you get a line read from the director. “Here’s how I want it; replicate it.” Done. Moving on. Next one. The guys who are really good at it? You’ll get a couple hundred lines an hour out of them on those barks. Especially in a game like Rainbow Six, where it’s really short and always at the same intensity, like “Incoming! Stacking! Frag and clear!”
JS: Often the writer will be present in case someone can’t deliver a line. Often the same script goes to every character. Every background soldier, say, will get the same script, but if it’s becoming too repetitive, sometimes Simon will say “let’s change this line.” As long as it’s okay with the writer and they don’t have to animate the mouth, they’ll change it just to give it more variety.
SP: It’s an argument I’ve had at a lot of production companies. When you’re doing voiceovers for cutscenes and walk-and-talks, they’ll take a lot more time with it - twenty or thirty lines in an hour instead of two hundred. But if you actually look at the practical effects of it, you only hear those lines once. Those background barks, though, you’re gonna hear those hundreds of times each. So if you don’t spend the time making sure they’re of a high quality, it’s like “you can smell the river from here” from Thief. Every single character in the background records that line. It’s one of those things that bugs me. You can’t do jokes in barks, unless you have some system of implementation whereby you can guarantee that line only plays once. With the project I’m working on with Eidos at the moment, they’ve been really protective about that sort of thing. They’ll identify, “that’s a funny joke. Once one character says it, it never gets played again.”
Do you frequently get lines that show up in the final game and just clunk?
Oh, yeah. It’s hard. If the actor comes in and does 200 an hour, and as a director you’re in the studio for eight hours, you’re doing 1600 lines a day. It’s really hard for one of those not to get away. I worked on Rainbow Six: Vegas, and it was exactly the same barks for every single character. 800 of them. But then you had to record them in stealth mode, combat mode, relaxed stealth mode, combat stealth mode, combat relaxed...so many permutations. You’re recording the same line eight times for each actor, over sixteen actors, 800 lines. By the time you’ve done that, you’re just starting to go nuts. You can’t tell up from down sometimes. Tango? What’s a tango? That’s a funny sounding word.
And pain grunts and things?
We call them “oh-noes” - onomatopoeia, and also the reaction of the actors when they see them. Those ones, we record like 300 an hour. We always schedule them at the end of a session, before we trash their voice.
JS: Simon was actually instrumental in making sure that the union had a limit on how long an actor can do oh-noes, because it really is hard on the voice. If the director isn’t like Simon, who’s sympathetic towards that, they’ll book the actor for a whole day. I mean, that’s impossible. You can’t use your voice like that, screaming nonstop for hours.
SP: They’re good to record. That’s where you find out if someone’s a gamer or not. If they’re a gamer, it’s really easy and fast. It’s like, “give me ten impacts, arrow to the torso.”
They know the function of it.
They know the function of it, and they know the way it sounds in games generally. I mean, oh-noes are horribly unrealistic. I remember a WWII veteran talking about Saving Private Ryan, saying it wasn’t like that at all, all that screaming and stuff. If someone got hit on the beach, they were silent. That was the creepiest part of it. They’d just collapse, crumple, and go into deep shock really quickly. It’s like if you spend time in an emergency room. Very few people come in making a lot of noise. If they did, they’d be the ones with minor injuries. Major traumatic injuries are quite quiet. But oh-noes can’t be quiet, because it’s about feedback to the player. The player needs to know if they’ve shot somebody, they need to know if they’re going to die. So it’s heightened and sort of ridiculous, which gamers will get straight away.
It’s also the language Hollywood’s established in movies. Everytime someone gets punched in the face in a film, it’s a big sound effect. They’ve created this weird vocabulary of sound that games have latched onto. They’ve already established to people what it “sounds like,” so we’re not going to mess with that. And then we have to build variety into it as well, so you’re not just hearing the same impact or the Wilhelm scream every time you get shot.
If you get an actor who doesn’t understand how that stuff works, it’s harder to explain. “So...you’re partially on fire, and you’ve just made that realisation. It’s not a bad heat yet, it’s gonna ramp up, but there’s also pain on the inhales, your chest’s on fire, so when you inhale, you’re burning your lungs” - it’s just easier to give them a demo. “It’s just going to be this!” [makes noise just described] Some actors hate that stuff. Others love it. I like doing it - it’s really fun, dying for an hour.
JS: It is fun. Gets your frustrations out.
Game protagonists get a lot of criticism for being samey. If you’re doing a series of gritty protagonists on a path of revenge, how do you differentiate them?
SP: It’s really hard. Because the protagonist in most cases is also representing the player, it’s hard to make bold choices with that. It’s one of the things we talked about with [Deus Ex: Human Revolution protagonist Adam] Jensen. On the first Deus Ex, we spent so much time recording him and re-recording him, putting him in, testing it. We ended up using a pretty narrow emotional range. Jensen never gets really upset or really freaked out by anything. It’s really subtle, because you’ve gotta remember the player has to be feeling the same thing to get to an extreme. So if you have your protagonist absolutely freaking out, and the player hasn’t had that same experience, then it breaks them out of that immersion.
That’s the other extreme - when people make fun of it for seeming over the top.
“It’s stupid! It’s dumb!” Yeah. But you can take those risks with the other characters. Like with Far Cry, with Vaas, we went really crazy with him. He can be really ridiculous, because he’s not representing the player. And if it’s done right, it adds more colour to the protagonist as well - the way you interact with them and reflect back on their actions.
You get to have more fun with NPCs?
Definitely more fun. It’s where you can make much more interesting choices. I work with Nolan North quite a bit, and he’s a different case as the protagonist, because he’s just an incredibly funny guy. Really quick, really witty. I’ve directed him as Desmond in Assassin’s, I worked with him on Teenage Mutant Ninja Turtles. [laughs] He’s a freakin’ genius. When you get him in the studio, you can try a lot of things. On a lot of Uncharted, they’ll get him to deliver the script, then say, “and now give us one however you want.” That gives you a much more natural-feeling result out of it as well, which means as a protagonist you actually can go a little bit further, because it’s coming straight from him. And they get themselves organised enough that they can show him what’s happening on screen as well, so he can just improv off that. It’s a great gift. Even different barks as well - it’s not all coming from the imagination of just the writer. You get another voice from someone who also understands the character. Elias [Toufexis] is great for that too, with Jensen in Deus Ex. He knows that character so well now, because he took a couple years developing the character before voicing him. The writer says that if Elias struggles to deliver a line, it means the line’s wrong.
How much improv do you get in the studio, or is the script hard and fixed?
It depends on the project, and the writer for the project. But a lot of them, especially for barks, actually like it when you can get improvisers in. Because again, with just one writer writing all those barks, who’s going to come up with 150 different ways of saying “what was that?” There’s only so many ways, from your own frame of reference, you can do that. But if you bring in other actors and say, “say whatever you would say,” suddenly you’ve got some varieties you hadn’t thought of. It’s a valuable asset.
Some people skip every cutscene, and others dial down difficulty so they can just experience the story. Where do you stand on the balance? Your work can be completely ignored by some people.
The balance at the moment is still that gameplay rules over storyline every time. You can have a game with great gameplay where the storyline is crap, and it could still be a massive hit. But the opposite doesn’t work. If you have a great storyline but the gameplay’s crap, no one’s gonna play it. So you see a lot of times in the studio when you feel those decisions being made. The level designer’s decided to scrap a whole chunk of dialogue because they’ve put you in a mine cart for that stage instead. But it is changing. I think The Last Of Us is a good example of why it’s changing. When games like that come out, where the sales start snowballing because people are going “oh my god, the story and characters are amazing,” that gets held up in the industry a lot at the AAA level. It’s like, “look! You can tell good stories. You can have interesting characters, and still have good gameplay.” And that should be the holy grail, right? Get both of them up to that standard.
And ideally, they should support each other.
We’ve got a lot of writers who have come over from film and television as well, so it’s taking them time to understand the medium themselves. How do we make that mesh? How do we bring the two of them together? As companies, when do we have level designers and writers work together? What’s the process? With a lot of games in the past, the level designer did everything first, then said, “here are our twelve levels; make a story.” Now there’s a lot more sitting down at the beginning of a game, where level designers say what their ideas are, narrative can say what they’re thinking of for an overarching story, they discuss what works and what doesn’t work, and slowly but surely they start slotting level design and story together into some sort of coherent whole. Hopefully.
What would you hold up as great examples of melding drama with gameplay?
Again, The Last Of Us does it really well. Coincidentally, also zombies, The Walking Dead episodic stuff was really good, because it gives them a bit more freedom with the storyline. They can take a pause, see what’s landed well, figure out where the plot can go, and be open to possibilities. They can also see what was popular. A lot of game developers spend way more time on forums than people realise, looking at feedback and what players are talking about and liking. Sometimes it’s to the detriment of some things - a writer freaking out over what 0.001% of players think of something.
What kinds of things get changed based on feedback?
The most common ones are if players are having a hard time understanding objectives or gameplay mechanics. Often you’ll have to go back and get lines from characters and make them more clear.
With the stuff I’m working on at the moment, they won’t find out half the things that can be done in the game until the testers go in. If it’s a stealth game, say, you can drag a body around and hide it. That also means you can drag a body anywhere, right? So you can drag a body up to another character and drop the body. So now every AI character has to have ten lines covering a body being dumped in front of them. And then - like, we have secure zones where you can’t do that. But a tester will find a way. They’ll pick up a body, take it into the ventilation shaft, go around security, find three ways of doing this, this, and this. So our protagonist is now presenting a dead body to their boss, which we didn’t think was possible. And now the boss has to have ten lines to cover how weird that is, even though in actuality, when the game is played, there will probably only be seven or eight people in the entire world who will ever actually do that. The amount of time that goes into those little details is ridiculous.
That’s what you get when the player gets to choose what happens.
Especially with open-world games. The level designers cannot think of all the possibilities until you get people in to start playing it, and you realise, “oh crap, you can do that! Oh no! Now we have to have dead animal lines! Now we have to have foley of animals being killed!”
At least in Assassin’s you can pat animals as well. I really appreciated that! But you can only pat the sheep, the pigs, the dogs…
Not the bear.
I tried it.
You should never pat a bear. We’d be sued for that if we had. We’d have kids going up trying to pat bears in Yellowstone, getting mauled to death. “Another Assassin’s Creed getting someone killed! Video games!”
JS: Didn’t you kill the bear in the woods there, and you’d have a little prayer for the bear? Or was it a stag?
SP: I think both. Connor used to give little prayers for pretty much everything he killed.
I think you killed the stag first.
JS: You killed a rabbit...
Just working your way up the animal kingdom.
SP: Jenn did a lot of directing on [Assassin's Creed III] as well, directing all the Mohawk voices.
JS: That was really interesting! The Mohawks have only learned [their language] by speaking it with their elders. They don’t read it. So there had to be a Mohawk teacher there to tell them what it was they were saying. And even when they spoke it, it was hard for them to grasp what they were seeing. It was a really long process. And they all had to be authentic Mohawk actors to actually pronounce those words.
SP: That’s one of the things Ubisoft, with the Assassin’s franchise, has done a really good job of. They do so much research. There’s always one or two history professors they bring in to consult with. They’ll try and find real people who can actually speak these languages, as opposed to just “here’s what it sounds like, now say it back,” which is always gonna sound horrible. And from the cultural side of things as well, it’s just not right to bring in a bunch of white guys to read Mohawk and pretend to be Mohawks. Before you know it, you’d be in a 1960s Western again. Or an Adam Sandler film.
How do you deal with cliches when you’re directing?
We teach video game acting, and one of the things we talk about is that stereotype is actually a shortcut to character. If you’re an AI character in the background, you may have two lines. In those two lines, the player has to be able to identify who you are and what you are. And what’s the easiest way to do that? It’s a stereotype. Here’s the blacksmith. We know, because he’s got a big burly old English accent. “Ooooh, you need yer shoe fixed therrre?” Same with a pirate. “Oh, oi’m a poirate. Ooh, arr, arr.” You don’t want your pirate coming in with a soft French accent and this beautifully-worked characterisation - oh, and by the way, we need 200 lines in the hour for that beautifully-rendered character. There’s only so much subtlety and nuance you can get there. So for AI, cliches happen a lot, but you embrace it. For your protagonists and your bigger roles, you’ve got the room within those characters and their scenes to try and do something more interesting and show some more colour.
Check back tomorrow for part 2 of this interview, in which we discuss Shakespeare, improv, gender, and Hamlet: The Video Game: The Stage Show.