Exclusive Interview: ArenaNet’s Bobby Stein

I’ve spoken to ArenaNet’s lead writer, Bobby Stein before, but after playing Guild Wars 2 at the Eurogamer Expo last month he was kind enough to  get in touch to see how I got on.  Despite my initial reservations most were unfounded, but there were still just a couple of things I wasn’t entirely happy with. With interview questions in hand, here is how we got on…

Tap: Great to speak to you again Bobby. I’ve finally gotten my hands on Guild Wars 2 at the Eurogamer Expo and must say it has blown me away. One of the first things I wanted to ask was how you have continued to gauge poor or good acting in the game (we’ve touched on this briefly before).  Some French and American individuals I spoke to at the Expo loved the voice acting in the opening cutscene, but some of the English found it incredibly ‘cheesy’. Are you resigned to not being able to please everyone?

Bobby: Fan response to our VO has mostly been positive, but there are plenty of people who expect our version of the game to match what’s in their heads. We can certainly identify and improve instances where a line of dialogue or an actor’s delivery isn’t working, like we do every time we’re back in the booth, but there’s no way to please everyone since opinions are so diverse.

Voice-over in games is very subjective. You can equate it to a film adaptation of a book. Readers will imagine the characters and setting differently from the author and from each other. If the movie deviates from someone’s perception, they react. Some people like accents, whereas others hate them. Some people feel that our creature voices need more postprocessing, while others are comfortable with what they heard in the demos. It all depends on the person. Perception of VO quality is also affected by context, so if animations, setting, or character models aren’t working to sell the fantasy, it can put more focus on the audio than would normally be there.

To put it into perspective, we average about 6% for retakes with each studio session. Lines are marked for rerecording for varying reasons, such as when a file gets lost or corrupted, a performance doesn’t match context, the delivery is wrong, or because of redesign.

Tap: The opening cutscene sees the player become the actor and narrate their own personal story over some stunning artwork. One of my original concerns was matching voice to face in an effort to improve immersion. Are players going to have the ability to change their voice or can we expect all men and women of each race to sound the same?

Bobby: There are 10 unique voices for player characters, which represent the five playable races and two genders for each. Every PC has more than 300 lines of unique combat chatter in addition to hundreds more lines of dialogue for each story permutation. When you consider that every player line needs to be recorded 10 times, you have to be deliberate in how and when the player speaks. While it would have been great to offer our players more vocal customization, it just wasn’t worth cutting other content types or reducing each player’s lines just to add an extra voice option.

Tap: When speaking to prominent NPC’s during your personal story, the game shifts to a cinematic where your character and the NPC are placed on screen next to one another, surrounded by a cinematic black border as they talk. On most of these occasions the acting has been absolutely fine, but the animations (lip syncing and gestures) have been incredibly distracting. They fail to convey the emotion in the acting and as a result the cutscenes end up feeling awkward. How are you working with animators to ensure that the dialogue you and your team write is conveyed accurately and with emotion, through the animation?

Bobby: Much of the personal story is told through what I call conversational cinematics, with two or more characters speaking against a painted background. The current visual presentation isn’t final, so what you’ve experienced at the Eurogamer Expo will look different by the time we ship the game. Our animators and programmers have been working really hard this past year on a new system that provides lip-syncing in place of the simple jaw flapping animation that’s in right now. You’ll also see a greater variety of gestures that will be timed to the VO, as well as improved lighting. All of this will retain that painterly aesthetic that is the signature style of Guild Wars 2.

We feel that whenever there is a dramatic moment or tonal shift in your personal story, it should be reflected onscreen. We consulted with the animators on what sorts of gestures and emotions we felt were necessary to carry the cinematics, and together we’ve greatly increased our options.

Tap: I ventured over to the Star Wars: The Old Republic booth after spending a great deal of time in Guild Wars 2. The games cutscenes are brilliant and of those I saw, flawlessly acted. Are you confident your new technology will be able to bring the game up to that level?

Bobby: Our animation and technology improvements will make a huge difference. We’ve also taken a closer look at the dialogue in our cinematics, and we’ve rewritten and rerecorded where necessary to tighten them up. We aspire to reach the same level of quality as our peers, since that’s what we’ll be measured against.

Tap: As someone who reads quickly and who loves just text, during the personal story cutscenes I tend to skip through the dialogue as I’ve read it before the characters are even half way through speaking. This obviously undermines the efforts you’ve made in crafting the voice over work and I’d question whether using actors is really worth it in the end. Is it more about giving people options to listen or read (or both), or something entirely different?

Bobby: If the characters are interesting and the dialogue is sharp, many people will sit through the show. We never forget that Guild Wars 2 is a game first and foremost, so we try to keep our conversational cinematics brief while also retaining a level of character in them. Cutscenes are successful when they are as entertaining as they are informative. That’s our goal.

For those who like to skip ahead, we’re giving them as many options as we can. You can currently skip a line in a conversational cinematic, or go straight to the end. We hope that people find the cinematics entertaining enough to watch over and over, but completely understand if they’re pressed for time and want to get straight to the action.

Tap: I suggested in my Impressions of Guild Wars 2 that I’d have like to have seen my personal story progressed through animated artwork that was narrated (just like the intro-cinematic) instead of you and NPCs talking to one another in a black bordered in-game cutscene. Was something similar to this ever discussed?

Bobby: It was discussed, but is ultimately an issue of quantity and resources. The stylized, painterly cinematics that you see after character creation are hand crafted by a team of artists and animators. The process requires a lot of man-hours to complete, so we’re limited in how many of these we can create. Because these cutscenes are large scale, they don’t often involve characters conversing, but instead portray lots of action and grand imagery.

In contrast, the conversational cinematics are a lot easier to create, and since there are over 1,000 of them sprinkled throughout the game, we need to make sure they look and sound the best they can. That’s why we’re devoting a lot of resources to giving them diverse animations, lip-syncing, and other improvements.

Tap: Now that Guild Wars 2 has been rolled out to various conventions in Europe and America, how has the response been to the voice acting and scripting?

Bobby: The response has been generally good, but there are people who have different expectations. Many people are first hearing the VO out of context, such as from one of our developer blogs. The audio that we host there isn’t fully processed, and it’s also heard out of context and without visuals. Many people don’t understand that it’s ambient, similar to the street chatter in a game like Grand Theft Auto IV.It’s important for those people to experience the writing and VO in the game setting, which is why I pay much closer attention to feedback from convention attendees and alpha testers over people who have only seen the game through YouTube videos, developer blogs, and scripted trailers.

I’ve asked a number of our alpha testers for their opinions on the writing and VO, and their responses were very positive. There are definitely areas where we can improve, so we take that feedback into consideration as we iterate our scripts and occasionally recast particular roles.

Tap: Overall, how happy are you and the whole ArenaNet team with the voice acting and from the demo shown at the Eurogamer Expo, has there been any major changes? Is there anything you’re still not happy with, that you’re able to share with us?

Bobby: I’m generally happy with our cast and the performances they’ve given us. Sure, there’s the occasional line that I’d like to rewrite or rerecord, but when you think about the impact of one line among tens of thousands, you realize that it’s not a big deal. A cinematic ends and the player moves on for another 200 hours, so unless a line is horrible or poorly delivered, it’ll most likely be forgotten in the grand scheme of things.

People tend to forget that the amount of story and VO content in Guild Wars 2 equates to more than 60 feature-length films. When compared to your average single-player game, it has more than 10 times the average VO and dialogue. Our goal is to maintain a high level of quality across the board, which is very challenging given the scale, complexity, and constraints.
If you look at games like Uncharted or Red Dead Redemption, you’ll realize that they have a small number of cinematics and characters. They can afford to construct their cutscenes with painstaking levels of detail, with all actors present. You can’t really do that in an MMO with literally thousands of characters and hundreds of voice actors. NPCs move around in the world and are constantly interacting with each other and with players, so it’s not feasible to expect each scene to be recorded with all actors in the same room. There are too many variables and permutations to manage the process effectively.

The compromise is to isolate a small number of cinematics with a set of well-developed characters, and get those actors in the same room together. That’s where you’ll get your best performances. Short of that, you have to give actors as much information as they need about a particular scene or their character and let them bring it to life.

Tap: My original reservations about the charr voice acting have actually disappeared since playing the demo. I found them entirely convincing and thought they sounded great, especially in comparison to some of the humans. Is it difficult to find that consistent quality level across four races and from dozens of actors?

Bobby: We’ve recast a number of roles that weren’t working in context. An actor might capture the attitude and essence of the race, but lack the tonal qualities that we’re looking for. In those cases we’ll try to manipulate the audio to bring it within range. If an actor has a great sounding voice but has difficulty grasping the role or creature race, we usually recast.

We’re striving for consistency in our VO, which takes a lot of time and effort. We’ve rerecorded thousands of lines with our newer actors—everything from combat chatter and greets to cinematics and scenes. So yes, it’s difficult to maintain that consistent performance quality across thousands of NPCs and hundreds of actors, but it’s worth the extra effort.

Tap: When you’ve gone back to re-record dialogue and voice overs, is it difficult to break the bad news to actors? Inevitably your telling them they are bad at their job…

Bobby: Our sessions are often months apart, so actors seldom remember if they’ve recorded the same scene in the past. Sometimes they’ll ask why they’re reading the third line in a given scene rather than the entire thing, and we explain to them why we’re redoing it. Sometimes it’s because of a rewrite or redesign, and other time it’s because the original delivery didn’t work in context. Whatever the reason, the actors are usually very professional and accommodating, and they rarely take it personally.

It also helps to remember that these actors often walk into a VO session with little knowledge about their role beforehand, simply because in video games many voices are based on archetype rather than individuality. You voice a generic human soldier differently than someone like Logan Thackeray, who is a more developed character.

Our job is to provide background information so that the audio director and actor can find the proper voice we’re looking for. It’s not usually a case of an actor being bad at his or her job, but rather them needing some guidance and context to make their performances sound believable and entertaining.

Tap: Finally, how are you finding the pressure to deliver on such an important element of the game now we are drawing closer to closed beta, where potentially thousands of players will really get a perspective of the personal story?

Bobby: I feel pretty good about it, honestly. Most of the VO we’ve got works really well in context, and since we’re putting resources into improving our conversational cinematics, I’m a lot happier now than I was a year ago. I want people to look at Guild Wars 2 and say that ArenaNet has evolved as a studio, that we’ve honed our writing and improved our VO over the years. Our fans will let us know if they agree, but most of our alpha testers and convention attendees share that sentiment.

We’ve refined our writing and editing processes throughout development, so any given line of dialogue has usually received multiple passes before we get to the VO booth. We’ve also learned to keep our lines short and to the point, but to also be mindful of technical limitations when designing and writing for the game.

I’m happy we’ve made it this far, and I’m excited to finally get the game into the hands of anxious people around the world. It’s hard to be objective about something you’ve poured your heart and soul into for more than four years, but I’m hopeful people will like what they see, hear, and read.

Also, I’d like to say thanks for getting in touch. I’m glad you enjoyed your time playing our game. It’s one thing to tease fans and press about the project we’ve been working on these past few years, but something else entirely to be giving people the chance to experience it for themselves. The hands-on demos we’ve been showcasing over the past 12 months have produced staggering amounts of feedback from fans and press, and we appreciate the input.

Tap-Repeatedly would once again like to thank Bobby for taking the time out of a very busy schedule to take part in this interview.

Email the author of this post at lewisb@tap-repeatedly.com