After my latest Guild Wars 2 article, “Vocal Coaching”, ArenaNet’s lead writer Bobby Stein got in touch, and put some time aside to discuss Guild Wars 2’s voice acting, its dialogue and the audition process. Here’s how we got on…
Tap: To begin, and for the benefit of those who are new to Guild Wars, can you tell us about your specific role at ArenaNet?
Bobby: I’m the lead writer (or writing team lead, depending on how much room I’ve got). I manage seven writer/editors on the Guild Wars 2 design team. I make sure they have the resources and tools they need in order to work efficiently, effectively, and creatively to draft and/or edit in-game voiced dialogue and UI text. Since I have a background in film and television production, I am also involved in casting voice talent, attending recording sessions, and helping the production team wherever I can.
Tap: How have you approached the script within Guild Wars 2, in comparison to Guild Wars 1?
Bobby: By rough estimates, Guild Wars 2 will have as much text as the entirety of the original Guild Wars series, including all three standalone campaigns and the Eye of the North expansion. It’s absolutely enormous in terms of written content. Because this is such a massive undertaking, we’ve segmented duties for the various content areas. Jeff Grubb heads up the dungeon story. Ree Soesbee is the primary writer for the personal story chapters, along with Scott McGough from my team. Lastly, the writing team is the driving force for event writing and VO, as well as ambient scenes. I give all the major dialogue a “script doctoring” edit pass, and try to make adjustments so that it sounds more natural.
Ultimately, my team looks at all the text in the game, whether it’s written or voiced. Sometimes we’ll author something from scratch with guidelines provided from a content or systems designer; otherwise we’ll take a rough draft and do the necessary rewrites and edits to bring it in line with internal standards and house style.
Dynamic events comprise a large portion of the overall writing and VO allocation, so it’s important that every voiced line is used effectively. Players should feel immersed in the action, and contextual voiced dialogue works much better than text bubbles to get the point across. I’ll give an example.
Say you’re travelling a road between two towns, when a guard approaches you in a frenzy. He emphatically tells you that the estate he’s guarding is under attack, and that the family–including a small child–is in danger. You storm the property and cut through waves of pirates. Meanwhile, you hear the owners down below cowering in fear. But you’ve arrived a few moments too late. You overhear the kidnapping take place, and the pirates whisk the little girl to their hideout for ransom. If you’re quick about it you can even watch them make the journey. An event chain like this is extensively voiced, so everything from battle chatter to context-relevant dialogue is triggered at key moments to make you feel like you’re in the middle of a tense situation. The characters have personality and motivation, and it comes through during each event.
Tap: When was the decision made to use voice acting for quests as opposed to in Guild Wars 1, which made extensive use of text?
Bobby: At the very beginning of Guild Wars 2 development, principals from the design, writing, world building, and audio teams had a lengthy discussion on what we’d like to do differently. First off, we expressed our frustrations. It’s very difficult to tell a compelling story simply using cutscenes as mission bookends. There’s simply not enough time to effectively develop characters in such short bursts, especially when those bits of screen time are often presented hours apart from each other like they are in an MMO. Short of making them longer, which we felt wasn’t best for a cooperative online multiplayer title, we had to look elsewhere for a solution.
We looked at how some of our favorite single-player games handled these problems, and the thing they all had in common was an effective use of voice-over during gameplay. Ambient dialogue can tell you a lot about a game’s setting or the feelings and motivations of its inhabitants, and contextual dialogue can develop characters and move the plot forward in less obtrusive ways, while also providing essential gameplay information on the fly.
Tap: Voice acting is an incredibly expensive process and ArenaNet have stated they have made more efficient use if studio time ‘to get more bang for their buck’. Was there ever a time where you felt returning to text for all but the primary characters would be a better option?
Bobby: It’s certainly easier to leave text unvoiced. For one, it’s cheaper because you don’t have added cost of actors for multiple languages, audio engineers, file cutters, and voice QA. You also have more flexibility in a production schedule to alter things should a redesign impact text content. You can iterate unspoken dialogue more frequently which can be a huge benefit in getting the story, characters, and gameplay right.
That stated, you potentially lose an enormous amount of ambience and character relying on text alone, and since video games are multimedia you could say that you’re not taking advantage of all the tools at your disposal. Ultimately, you need to do what’s right for the game you’re making while also pleasing your fans. When push came to shove, we decided to extensively voice dialogue throughout the game. The exception to this is our conversation system, which is not. Our reason for this decision was twofold.
Players often read ahead of spoken dialogue and skip to the next conversation state before a character finishes her line. Rather than voice every bit of every interaction, we felt that the ambient voiced dialogue could carry the brunt of the story load, since absorbing plot points and character motivation feels more natural that way. Also, we didn’t want to slow down grouped players by tying the party leader up in a lengthy exchange. That’s not to say that there aren’t other ways to solve the problem, but when we ran the numbers in terms of line count, budget, gameplay payoff, and player investment we felt not voicing conversations worked best for the game we’re making.
Tap: To use voice actors for an entire game must be a mammoth task, but to voice an MMOG is on another level (which Bioware are also currently experiencing). How are you attempting to retain quality but also balance an indepth storyline?
Bobby: It’s enormously difficult to generate content for a game this size, but it’s made even harder when you add voice to the equation. I’ve had many conversations with audio folks, writers, and designers in and around the industry over the past few years. Not long ago, a single-player game with voice acting might have several thousand lines of recorded dialogue to carry you through 10-hours of focused play time. So something like Half-Life 2 or Portal can effectively tell a compelling story with rich characters, all in a relatively controlled environment with a conservative amount of voicing.
When you look at open world games like Grand Theft Auto IV and Red Dead Redemption, or extensively voiced RPGs like the Elder Scrolls or Fable series, it gets a lot more complex. You not only have principal characters to worry about, but thousands of ancillary ones roaming the world. Unless you’ve got an unlimited budget and years of time to devote solely to writing and voice acting, you have to determine just how many voices you’ll need to flesh out the world, and how many lines these NPCs will need to say to avoid the inevitable repetition and ear fatigue that sets in after you’ve spent 25 or 50 hours wandering the world. To put it into perspective, there are Guild Wars players who have spent several thousand hours (and counting) in Tyria, so you can imagine how much additional voice you need to keep the world interesting.
Tap: Do you consider using voice actors more restrictive creatively in comparison to written text? Especially considering the writing talent within ArenaNet…
Bobby: It’s restrictive in a positive sense. If you give writers a blank page, they’re inclined to fill it with a lot of potentially unnecessary information. By putting some guidelines into place about where and when voice can and should be used, and the average length of a line, it forces us to be more deliberate in context. If you watch a film or television show with closed captioning, you see a distinct difference in how dialogue is delivered compared to a game that relies heavily on text. Lines are shorter and more impactful, and characters often build off a previous statement rather than reiterate it.
Development of Guild Wars 2 has forced us to rethink our approach to dialogue. Since writers make up a relatively small portion of the larger design staff you can imagine that it’s sometimes difficult to break old habits or enforce newer styles when everyone previously wrote for the page or the monitor rather than for the ear. This change is most noticeable in the human starter quests that we showcased at Gamescom back in 2010, and in the idle scenes that trigger around Divinity’s Reach.
Tap: Have you personally had an active role in auditoning and hiring actors and if there was an audition process, who determined their quality and how was this measured? The quality of a persons acting is obviously very subjective…
Bobby: Yes, absolutely. I, along with Jeff, Ree, Eric, Colin, and Ben, reviewed casting submissions for every voice role in the game. As you can imagine, we’ve had a lot of actor interest so I wouldn’t be surprised if we listened to nearly a thousand different actors over the past few years. We don’t always agree on a particular actor, and we’ve had to recast certain roles when the voice talent didn’t quite work for a particular character. Thankfully, we have the time and resources to fix those mistakes throughout production. Just because you’ve heard a particular actor in a convention demo does not mean that person will be delivering those lines in the final shipped game.
Tap: How did you prepare your voice actors for delivering their lines? Did you provide them with background material to the race they would be acting or images of the character(s) they would be playing?
Bobby: We provide a bio that contains some essential character information, a picture, and some sample dialogue to give the actors a reference point. They use this for their audition, and refer to it during the recording process. We also send the scripts to our studio partners in advance of the session so the actors can better prepare for the scenes. I know Troy Baker, who plays Logan Thackerary, studied his character in advance and it shows in his performance. His reads for the human quest chain cinematics are very natural because of his personal investment in the role.
We also give our actors a bit of leeway, so if they want to try something different with a line, be it speaking in a different tone or ad-libbing the dialogue, we let them experiment. Sometimes it doesn’t work and we go with the original take, but many times they bring something to the performance that really brings the character to life.
Tap: Of the voice pages released so far during racial weeks, many of recordings demonstrated are to be used by NPC’s as conversational dialogue. Is there a large quantity of these recordings to ensure players won’t hear the same lines repeatedly?
Bobby: The conversational dialogues we’ve been releasing are mostly idle scenes. These are background exchanges between two or more NPCs around cities, towns, and out in the world. What’s potentially misleading about these audio samples as they’re portrayed on the developer blog is that they’re not of the front-and-center variety, but ambient and generally heard in pieces as you’re exploring an area. When you hear things out of context such as on our developer blog, it’s sometimes jarring. People jump to conclusions because they don’t have a sense of where it fits into the experience. Most of those fears are unfounded, though, because when those scenes are viewed and heard inside the game they sound a lot more natural because of where they happen and at what volume they’re played.
Most of the idle scenes were written to be race-specific and are unique to an area, though there are some common ones that can be triggered anywhere. There are well over 20,000 lines of voiced dialogue just for idles and events, with more being added as development progresses. This represents just a fraction of our overall voice budget. Any given map should have plenty of dialogue variety.
Tap: How did you go about constructing the dialogue for these types of conversations?
Bobby: Idle scenes were developed early on since they’re considered relatively stable from a design standpoint. Most of what the characters idly talk about is independent from the player story or dungeon areas, so unless there are sweeping lore changes we can write them early and iterate them over time. The first ones were recorded back in 2009. If, after reviewing them we found major problems with the dialogue, acting, or context, we had plenty of time to go back into the studio. Ultimately, if something wasn’t working for us we cut it, whether or not the public had seen or heard it already.
We’ve dropped some scenes and added new ones as necessary. Home cities are especially dense with idles, so just by roaming around Divinity’s Reach, Rata Sum, the Grove, Hoelbrak, or the Black Citadel, you should get a pretty good sense of a given race.
Tap: As a writer and the team lead, what is the most difficult challenge you’ve personally faced whilst working on Guild Wars 2? Is there a constant battle between time/budget?
Bobby: The writer in me would love nothing more than to focus solely on the dialogue, but my job carries managerial and developmental responsibilities which fight for my time. I can always use more resources (more writers, editors, and actors), but game development is frequently about the struggle between time, money, and quality, and as any producer will say, “Pick two!”
We’re fortunate that Ncsoft has given us the time to finish the game, so quality is of utmost importance. That’s why we haven’t announced a ship date yet. We don’t want to release a game that’s half-baked, and that holds true for the writing and voice acting as well.
Tap: Guild Wars 1’s acting left much to be desired and I must admit the quest givers in the recently revealed norn starting area I found to be lacking quality in places. How happy are you with the current voice acting in game? Can we expect extensive post-processing and further editing?
Bobby: What, you weren’t a fan of Shiro Tagachi’s award-winning performance? (I kid!)
Overall, I’m a lot happier with the Guild Wars 2 recordings since they represent a shift in how we write for voice. That stated, we knew going into the last studio session that the writing and design for the norn personal story (what we showed at PAX East 2011) weren’t ready for voicing since the design wasn’t finalized. We decided to record it anyway so that folks on the show flow heard something during their session, even if it was placeholder. We figured it was better to have that than “robo-voice” temp files, even if the text was of first draft quality.
We’ve been taking inventory of the various cinematics and scenes in each of the personal story chains and rewriting and rerecording where appropriate. We’re iterative in our design, and that’s reflected in our writing and voice-over processes. The final, shipped version will look and sound different from what players saw earlier this year.
As for post-processing, all voice will have it to some degree. Human ones will go through the usual cleanup like file trimming and normalizing. The other races will have more extensive work to really make them sound unique. For example, the charr voices in Scott’s recent blog post had multiple effects on them. Everything from pitch adjustment to slight distortion was added to make them sound less human. We have to be careful that we don’t go too far with voice processing, since the human ear has its own concept of the Uncanny Valley. If you pitch drop a female voice too far it sounds unnerving. The key to getting a race to sound good is to select versatile actors who are comfortable speaking in that range. When we force them out of their comfort zone it doesn’t work as well. In those cases we either take a different approach to the writing, the vocal qualities of that creature race, or the casting selection.
Tap: How has the iterative process ArenaNet adopt effected your ability to write and record dialogue?
Bobby: It’s mostly a blessing, but sometimes a curse. The upside is that we can refine dialogue until it works. The downside is when we “finish” writing and recording something, only to turn around and throw it away. You can’t be attached to your writing in game development because it will be edited, rearranged, added to, cut, or otherwise changed as the game evolves, but since it’s in service of crafting a great experience for our players it’s worth the effort and mild emotional distress.
Tap: Have you concluded all your voice acting sessions within the studio, or are you still finding your having to re-record where (and for example) you may be unhappy with certain lines?
Bobby: No. We are continuously recording new dialogue, while also revising existing stuff if it’s not doing the job. If a line is written well but poorly performed (or vice versa) we identify it and try to remedy it the next time the actor is in the recording booth. When we see or hear something that doesn’t work, we either fix it or cut it if we can.
Tap: Writers and directors often clash in creative media such as movies and television; in these instances the writer is usually on the losing side. What level of input do the writers at ArenaNet, and in your experience throughout the industry, have in casting and directing voice talent? What is your opinion on voice directors in general in the games industry?
Bobby: I’d be willing to wager that 95% of all game companies don’t employ full-time writers on staff. ArenaNet is somewhat of an anomaly in that regard. Maybe it’s because RPGs require more written and voiced content. Whatever the reason, I’m glad to work for a place that appreciates our craft. We don’t win every battle, and we sometimes have to make concessions on the type and tone of the content we’re working on, but the fact that we’re at the table in design meetings is hugely empowering.
I report to Eric Flannum, our lead designer. My people are often embedded on production and strike teams where the writing is a significant component. That level of involvement is rare in a development environment, but is absolutely necessary if we wish to evolve and improve our writing and voice acting with each game.
As for voice directors, I feel they are absolutely necessary for getting quality VO. They know how to talk to actors. They understand the process. Even when we might disagree on a particular take, I value their opinion.
Tap: A well-established franchise like World of Warcraft has a huge backlog of lore and history to draw from, something that new MMOs often have to create from scratch. Do you (alongside Jeff Grub and Ree Soesbee), find it problematic to develop lore in advance of a big release like GW2, rather than having it grow organically through years of development in the same world? What can be done to make that history seem natural?
Bobby: Ree Soesbee and Jeff Grubb handle Guild Wars 2 lore. They’re world building experts and seasoned novelists, so I trust that they are “playing nice” for continuity’s sake. It is hard, though, building on a monster of an IP. You need to be faithful to the source material, but also take some chances to evolve it in new and interesting ways. That’s one of the many reasons why Guild Wars 2 looks and sounds a bit different from the original series. It also helps when you have a 250-year gap in storyline to play with.
Tap: Alex Garland (Hollywood screenwriter who did 28 Days Later and later worked with Ninja Theory on Enslaved) has said that before he found Ninja Theory, he approached many developers to offer his writing services and was told they had no interest. What degree of crossover in writing is possible from linear narratives like film to branching ones in games? Are they completely different skills?
Bobby: I wouldn’t say they are completely different, but they definitely require a different mindset. A screenwriter is generally creating 90-minutes of visual action and dialogue in a script compared to 100+ hours of content for an MMO. It’s especially hard for a non-gamer to grasp the challenges and limitations inherent in a persistent online game. For example, he might want to kill off a main character from a story branch, but you can’t do that in a persistent zone if that NPC is critical to someone else’s experience or story chapter.
It helps if the writer has a cursory understanding of branching narratives, or at least a willingness to learn. It’s difficult, though, for a contract writer to fit into the game development process near the end of a project. The earlier that person can collaborate with designers about characters, story options, and technical features, the better.
The way I look at it, if game companies want to tell compelling stories they need to have at least one professional writer/editor on staff. Writers look at character arcs and narrative structure through a different lens than content or systems designers, so it’s very important to have a writer around who can tell you if your awesome gameplay mechanic or scenario makes sense with the current story structure. If it doesn’t, you need to find ways to make the story and game work together.
Tap: Writers of fantasy often try to avoid falling into the Tolkein/Dungeons & Dragons stereotypes. Do you think it’s possible to go too far in this? Do audiences demand a certain comfort level with the world created by the writer? Was there any intentional effort in the GW2 creative process to avoid the “Dungeons & Rings” stereotypes?
Bobby: Everyone has a different idea of what fantasy dialogue should sound like. Some are at home in the Renn Faire zone, whereas others would rather stick flaming Q-Tips in their ears than hear it. My take is this: nowhere is it stated that a fantasy game has to be in a Medieval setting, or that its characters must all speak with British accents. I’m not saying that either of those are bad, just that it’s odd that we find ourselves confined to such a tight box.
Guild Wars is our baby. We can–and should–try new things. If we’re fearless in our design philosophies, why shouldn’t we experiment with the dialogue? The important thing to remember is that we treat each race differently in regards to how the characters speak. Humans in our game sound the most natural. Charr speak efficiently. Norn talk in grandiose language. Asura dabble in techno/magi-babble. Sylvari speak in a more traditional fantasy tone, though their sentence construction is mostly active. In essence, people shouldn’t worry that our characters speak in contemporary street slang, because they don’t. There’s something for everyone in a game this huge; every race sounds a bit different. Ultimately, it all has to work together to simulate a living world with believable characters.
Tap: ArenaNet have regularly said that they aren’t just trying to create the best MMOG ever made, but the best game. In my article ‘Vocal Coaching’, I’ve made various comparisons between the use of voice acting within the single player game and in an MMOG. Do you think the differences in quality can ever be equalled within an online videogame? The budgets for MMOGs are often surpassing that of a single player…
Bobby: MMOs should have writing and VO on par with a triple-A single-player game, top-notch television show, or big budget film, but as I mentioned it’s a mammoth task that requires a lot of time and resources. You have to balance technical issues and limitations, and work with available tools–or develop new ones.
MMO budgets are significantly higher than the average single-player game. Not only do you have to generate more content and assets, but you have to put more resources into network infrastructure and customer service. You have to be smart about how and where you use writing and voice-acting to make the best possible experience. Obviously, you need to fund them both adequately, but you also need to plan ahead and adjust your production schedule in ways that are conducive to iteration.
Tap: In film making it seems that writing is always given a secondary position. What’s the famous quote? “Screenwriters are just secretaries with a flair for dialogue.” Do you see a parallel attitude in games?
Bobby: Auteur theory holds that a film’s director is responsible for the creative vision, which overrides everything else. If he decides to alter a script so that it conforms to his ideas, then the screenwriter will make adjustments. What about in cases where the film’s director is also its writer? Are there internal conflicts?
Some films, especially independent productions, rely more on dialogue than your typical summer blockbuster. In those cases you can argue that the script is often more important than the visual effects budget. It really depends what type of film you’re making.
Games can follow a similar pattern because sometimes a writer is controlling the vision, but more often than not design is the driving force. In those cases, the writer needs to work with designers and not against them to create compelling stories and interesting dialogue.
Tap: Visual mediums like film always seem to subjugate text and audio. They sweat blood over art direction and photography and often let sound and dialogue slide until the very end. Like film, games are very much a visual medium (though gameplay trumps even the visual) and it seems that writing and dialogue are given short shrift. Do developers feel more pressure to up their game, so to speak, as audiences become more and more demanding of quality voice acting?
Bobby: Developers do feel more pressure to improve storytelling, writing, and voice-over in games. This is partly due to the evolution of the medium. We can now render exquisitely detailed worlds in three dimensions, emulate real-world physics, and simulate day/night cycles and weather patterns. We can surround you with sound and provide you with hundreds of hours of voice acting. With these new tools, we are required to entertain our audiences with higher quality content. Gamers are savvy, and they appreciate a good story that’s told with style, flair, and substance.
Tap: What role, if any, do game developers see for using actors early on in the process? So many gamers seem to resonate with well crafted characters in a game. Do actors have a place earlier on in the future of game development?
Bobby: It depends on the game you’re making, but if you can afford to involve actors earlier in the process and get them invested in the characters they are playing then you will likely get better performances. As more and more games utilize professional actors for not only voice-over, but also motion capture, there will be an increasing demand for skilled talent.
Tap: Finally, besides meetings (which I know your a big fan of!) what have you planned for the rest of your day and can you give us a sneak peak of when we can expect the next racial week?
Bobby: I plan to drive home, kiss my wife and kids, and decompress. It’s been a crazy month. As for the next player-centric racial week, it’ll either be asura or sylvari. I won’t spoil the date.
Tap-Repeatedly would once again like to thank Bobby for his time.
Email the author of this post at firstname.lastname@example.org