Sound Staff Interview
Hitoshi Sakimoto – Composer
Tomohiro Yajima – Sound Effects
Hitoshi Sakimoto (L)
and Tomohiro Yajima (R).
—What themes did you want to explore in the making of Vagrant Story’s music?
Sakimoto: With Vagrant Story, I wanted to make music that had a tighter fit with the game (compared with what I’ve written before). That was my first idea. In event scenes, for example, I wanted the music to more closely track and sync up with the climax, and for it to end at the same time the scene ends.
I’m particularly pleased with how the opening came out. In that sequence of scenes, no matter how many messages you skip, it always unfolds in sync with the music. It’s probably a rather plain thing that a lot of people won’t notice, but maybe the fact that people don’t notice it—that it’s seamless—is precisely what makes it a success.
—I definitely felt a unity between the sound and atmosphere while playing.
Sakimoto: For this game, I specifically divided the songs into two types: those where the music was meant to be in the foreground and focused on, and those where the music takes a backseat to other events. In places where you’re just walking around, not doing anything in particular, there should only be “ambient” music and sound effects. In contrast, when you’re battling a boss monster—the kind of creature without any humanity—I wanted the music to be in the forefont, conveying a sense of tension and urgency.
Yajima: After the planners shared the story and concept art with us, the first thing we actually created were the sound effects, before the music. That meant there were some scenes which had these really great, impressive sound effects, so we decided to leave the music out entirely in those.
Sakimoto: We found that the sound effects were perfectly capable of fulfilling the role of “sound” in many of those scenes. That allowed us to choose the best approach in any given scene: some were better with music, some were better with just sound effects.
—How were the sound effects created?
Yajima: I didn’t use any sampling. They were all created by manipulating noise and the basic waveforms (square/triangle/saw). Sampling is a great technique in its own right, of course, but it takes up a lot of memory, which imposes a significant burden on the programmers. Thus, even if it’s more time-consuming, I figured if I could avoid using sampling it would result in a higher quality game overall. Besides, if you’re meticulous, you can get near-identical results to sampling by using basic synthesis. Things like the howl of a wolf—I was able to achieve a very similar sound, almost indistinguishable in quality, from a sampled howl. I was very pleased with the results myself!
—It’s rare to see games these days that don’t rely on sampling for the sound effects.
Yajima: We had another reason for using synthesis: one of our core concepts for the sound design this game was “don’t re-use anything”. For instance, when a character moves, I wanted the sound of that movement to reflect the different clothing/armor they were wearing… or if we did use the same sound effect, I wanted it to sound different depending on whether you’re indoors or outdoors. Basically I wanted different sounds for every scene you were in. When you think about it, there are no two perfectly exact sounds in the real world.
—I think we can safely call you a “craftsman”, with that level of dedication. And for the music, Sakimoto, did you use any live instruments for Vagrant Story?
The opening scene for Vagrant Story,
which Sakimoto was especially proud of.
Sakimoto: For this game, the only time we used live instruments was in the staff roll song. I think this is the case for most developments, but for Vagrant Story we had a tight memory “budget” for the sound and music, so there was no space to use actual instruments.
In terms of my personal tastes, I do prefer using real instruments, but I’ve been doing synthesizer programming for a long time now so I’m probably better at it anyways.
—When it comes to game music composers, it seems like there’s two types: those who play a traditional instrument and can read music, and those who use a computer or sequencer. You sound like the latter type.
Sakimoto: Yeah, I never had any formal music training. I simply loved games, and that’s why I got into this work. When I was in high school I made a doujin game with my friend, and the composition and music programming I did then was a major turning point for me, and gradually I shifted my focus more towards music. So in the beginning, I had to ask other people about how the actual instruments are played, and I tried to copy and learn from them, not really understanding any theory. Even today I’m still studying and trying to learn that stuff, while I work. (laughs)
—What kind of sound/music would you like to do in the future?
Yajima: Sound effects are all about expressing urgency and atmosphere, so for me the ultimate would be creating sounds that are indistinguishable from sounds you’ve heard and experienced in the real world. Of course, I want to continue crafting everything by hand. (laughs)
Sakimoto: Whenever I compose, I always think about how the music will make someone feel while they’re actually playing the game. Is it ultimately a good thing or a bad thing, for the music to be competing for the player’s attention? More fundamentally, in any given scene, what do we want the player to be feeling? And how can the music contribute to that? Those are the questions I want to continue exploring as a composer.
Programming Staff Interview
Taku Murata – Main Programmer
—I understand this was your first “full polygon” game. How did it go?
Murata: It was something I had been wanting to try for awhile, but once we actually got into it, I realized how much more there is to take into account and be aware of, compared with the 2D games I’ve made.
—Programming-wise, what were some of the things you paid the most attention to?
Murata: Well, take the polygon models, for instance. If you know in advance which models can be rendered incompletely (that is, models which are ok to have missing polygons because they’ll only be viewed from a limited angle), then making a 3D game is very easy. The problem is, when you’re watching an event scene and suddenly the camera snaps around and you see an unintentional missing texture, it dominates your attention and takes you completely out of the scene, right? Of course we didn’t want that, so we tried to fix missing textures wherever we saw them, but in doing so it became an obsession… once we started, there was no end to it!
—A programmers work is as easy or difficult as his obsession lets it be, in other words.
Murata: I think the same is true for designers and planners. The other big headache for us was negotiating the memory card limitations. The Playstation memory card has a 2MB capacity. As the Vagrant Story development progressed there were all these different things we wanted to try, and we quickly ran out of memory. That phase of the development, where we were experimenting to see what we could shave off and still come away with a satisfying finished game—that was the hardest part.
—Were there any mechanics or parts of the game that you had to entirely cut because of the memory limitations?
Murata: There definitely were. For the battles, initially we had planned to use a more complex system of calculations. We had also planned to have slightly more effects for the spells and skills.
—And yet the finished game truly has an incredible amount of volume packed into it, which I can now see was produced through the suffering of the staff. (laughs)
Murata: It is a big game, yeah. But when you’re trying to make a game that allows players to choose from a variety of playstyles, it naturally entails a lot of data. And I think that, as programmers, that’s one of our most important jobs: to find a way to make the game the planners want to make within the constraints of the memory.
—Given the massive amounts of data involved, and the beautifully detailed visuals as well, I was amazed to find there’s absolutely no slowdown in Vagrant Story.
Murata: Strictly speaking, there is slowdown, but we simply tried to hide it so the player wouldn’t notice it. (laughs) We were adamant about not including the kind of annoying slowdown that stresses players out though. There’s so many calculations going on at all times in a game like this, and minimizing the slowdown is therefore extremely difficult, but we absolutely made that a priority—not stressing out the players.
—Speaking of which, I barely noticed the loading times for the CD data as well. Nice work.
Murata: The quickness of the CD load/read times was, actually, one of our secret “selling points” for Vagrant Story. Almost no magazines that covered Vagrant Story picked up on it though. (laughs) We do all the data loading during the event scenes, or while messages and dialogue are displayed. Of course, the planners helped make that a reality.
—The Vagrant Story development seems to have a lot of real perfectionists on the team, people who like to dive into the details: Matsura, Minagawa, Yoshida…
Murata: Everyone here just loves video games. On top of that, we all have our own specific image of the game we personally want to make. When those proclivities are fused and we come together as a single development team, I think it results in a surprisingly solid finished game.
—How would you rate your level of satisfaction with Vagrant Story?
Murata: I think the programmers all handled their individual parts very well, so the overall game turned out good. We succeeded in making something that just feels fun to play… with games, you know, it’s important that it just feels good to control. That simple but pure response—”whoa, this is awesome!”—in the future I want to make games that work even more effectively on that level.
Planning Staff Interview
Jun Akiyama – Planner
Kazuhito Maehiro – Planner
Takayuki Suguro – Planner
—When you first laid out the plans for Vagrant Story, what kind of game did you want to make?
Akiyama: For the event scenes, we knew they were going to be in 3D, so one of our first questions was how to move and direct the camera during those scenes. When it comes to 3D games, lots of them use CG movies like the Final Fantasy series, but we wanted to distinguish ourselves from that crowd, so we came up with the dialogue bubbles which we hoped would evoke an American comic book mood. We didn’t want Vagrant Story to feel like a “Japanese” game—that idea was with us from the start.
L-R: Kazuhito Maehiro, Jun
Akiyama, Takayuki Suguro.
—The camerawork is really well done.
Akiyama: I had never worked on a game with a camera, so I watched a lot of movies and did a bunch of independent research on my own. At first I was moving the camera around way too much, and we had to do a bunch of retakes. (laughs)
—And which scenes, in terms of camerawork, are you particularly fond of?
Akiyama: The opening scene, I’d have to say—the scene where Ashley is introduced in front of the castle gate, in the rain. It was the first scene I created, actually, but when everything was over I felt it still held its own, so it’s the most memorable for me.
—It blows my mind that you produced something this accomplished with your first 3D game.
Akiyama: Well, for me personally though, there’s stil a lot of things I’m unsatisfied with. There were also many elements we had to cut due to a lack of time or memory… such as flashback dialogue between Sydney and the Duke of Bardorba, and scenes explaining Rosencrantz’s motives… it’s really a shame I couldn’t include them.
—I’d have loved to see those. I wanted to ask some questions about the map too. What were your intentions there, at the beginning of the development?
Maehiro: If we were going to do a “dungeon” game, I wanted to evoke an atmosphere similar to the old Wizardry games. I also wanted to get a nice mix of puzzle and action elements, so the puzzle mode was there from the beginning too. People often refer to it as a glorified Sokoban though. (laughs)
—The puzzles are very hard!
Maehiro: They were actually a lot more difficult at first! There’s a guy on our team who is really bad at puzzles, and we tested them on him… if he couldn’t solve it in 5 tries, we would re-do it to make it easier.
Akiyama: But they still came out tough. (laughs)
Suguro: Indeed. (laughs)
—The different names for the individual maps are really cool too.
Maehiro: I came up with those names while I was thinking about the story in my own way. For instance, take the city walls: if you read through the names of the rooms in order, it should give you a feeling of a youth who is maturing as a swordsman. It’s my hope that those names will evoke the rest of the story for players.
The inclusion of old-school block puzzles likely reflects the sensibilities of the veteran development team.
—Speaking of the maps, I was also surprised at all the different movements and behaviors of the monsters in each room.
Suguro: In games like this, there’s usually a number of things that feel distinctly unnatural, right? For example, the way enemies will sometimes get stuck on terrain and just sit there not doing anything, or the way you can’t reach certain terrain without a “jump” power upgrade or something like that. We tried to purge Vagrant Story of stuff like that.
—What kind of things did you pay attention to when creating the battle system?
Suguro: I’ve always loved strategy games, so I wanted to include as many strategic elements as possible. Even enemies that appear very strong, for example, if you attack them from behind they’ll have a hard time counterattacking. Lots of stuff like that—I wanted to include a variety of ways for players to approach the battles, so please enjoy discovering them.
—For my final question, what kind of person is Matsuno, from your perspective?
Suguro: His sensibilities, when it comes to making games, are second to none.
Maehiro: He’s a taskmaster, but once you get to know him, he’s really a funny guy. You just have to solve the “Matsuno Puzzle” first, then you can see his true form. (laughs)
Akiyama: When I work with Matsuno, there’s a sense of assuredness that I appreciate. Whenever something seems impossible, either technically or because of scheduling, if you just sit and talk with Matsuno for 30 minutes you’ll see a way to get through it. Then, after you’ve returned to your seat, you’ll realize it wasn’t really all that impossible to begin with. That happened a lot to me. He’s got a certain charisma to him that’s hard to describe… kind of like Sydney. (laughs)
3D Graphic Staff – Interview
Hiroshi Minagawa – Art Director / Character Model supervisor
Eichiro Nakatsu – Character Model creator
Tsunataro Yoshida – Character Motion creator
—What parts of the graphics did you really focus your efforts on?
Minagawa: Making sure everything moved in real-time was very important. In recent Square games, the beautiful CG movies have become a big selling point, with many titles pushing them to the fore in an effort to appeal to players. That approach has its merits, of course, but for us, we wanted the visual appeal of our game to come from the player’s own actions–from the player controlling the visuals in real-time. Maybe it’s because we have so many veteran designers on our team, people who have been making games since the early days.
This development was largely about seeing just what this Playstation hardware could do, visually-speaking. Long-term, our plan was to take the know-how we acquire here and apply it to the Playstation 2 hardware, which is a level-up in terms of power. Those were the thoughts on our mind while we made Vagrant Story.
L-R: Minagawa, Nakatsu, and Yoshida.
—The character illustrations feel like they’ve been very naturally translated into polygon form. I bet that was challenging work, wasn’t it?
Nakatsu: Illustrator Akihiko Yoshida’s characters are very slender, which is balanced by their somewhat large hands—and that gives them a unique aesthetic. I knew if we could translate that in 3D it would look great. However, when we tried animating the models, many inconsistencies showed up, and it caused a great deal of trouble for the character motion creators.
Yoshida: I made a lot of adjustments, like increasing the size of the hands, to make sure that that their movements had a sense of balance to them.
—The character animations (motions) are another eye-catching part of Vagrant Story. Were there any special techniques you used to achieve that?
Yoshida: For the human characters, I would basically try the movements out myself on the spot, and use that as my reference for the motion creation. I think I annoyed a lot of the people next to me, always walking and running up and down the hall. Also, for the event scenes, I would reference other scenes from specific movies and so forth.
Minagawa: Both the character modeling and animation were a big challenge because of the low polygon-count. In FFVII and FFVIII, the texturing was very different, but there we had about 700 polygons to work with for a single character. In Vagrant Story, because the backgrounds also use polygons, we only had about 320-330 for each character—in other words, about half of what we had for the Final Fantasy games. Figuring out how to model the polygons within those limitations was very much like a puzzle.
Nakatsu: The sprite textures we used for the characters’ facial expressions were also extremely finicky to work with. We only had a 32×32 wide pixel space for them.
—That’s almost the same resolution as the Pocket Station. Yet strangely the impression they convey is very realistic… how’d you pull that off?
Minagawa: Well, take a character’s smile. We used a single row of pixels to depict that, but the problem is, say you then want to show the character’s mouth opening: if you suddenly increase that mouth line to two pixels wide, it won’t look smooth. Our solution was to animate it by subtly changing the shading of the individual pixels. Nakatsu here did almost all the facial expressions. We definitely couldn’t have done it without his longtime experience as a pixel artist.
—And having done all this very detailed, craftsman-like work on Vagrant Story to bring the real-time graphics to life, what do you feel you’ve gained from the experience?
Minagawa: This was our first full-polygon game, and what I think we realized is that all of our experience as 2D game designers did not go to waste in the transition. Whether it was figuring out how to make-do with low-polygon count models, or how to raise the processing speed by reducing the available colors—ultimately that kind of work demanded a similar skillset as our older, Super Famicom era pixel work did. Even the Playstation 2 will have memory and processing limits, so ultimately what decides the quality of our games is that old know-how we’ve accumulated over the years.