Donkey Kong Bananza is a game that emerged from constant experimentation with tons of different experimental prototypes, such as a Goomba with arms. Its developers didn’t want to give me a ton of specifics about too many early, rejected, or just plain weird concepts when I spoke to them at the Game Developers Conference this week, but I did manage to get them to talk about one: a giant, delicious, bacon burger.
At GDC, Bananza producer Kenta Motokura and programmer Tatsuya Kurihara held a talk entitled: Constructive Destruction: Fusing Voxel Tech and 3D Action Platforming in ‘Donkey Kong Bananza’. We wrote about the talk already, which covered a lot of new details about how the game’s voxel tech worked in tandem with concepts around action platforming and themes of destruction to create a cohesive game.
Following the talk, I was able to speak to Motokura (who I already interviewed once last year, prior to the game’s launch) and Kurihara about its contents, as well as ask them more questions about Bananza now that I’ve smashed my way through the game and seen the voxel destruction for myself. We mostly talked about voxels, but we managed in that to cover a wide range of subjects, including doing research in sandboxes, watching speedrunners break the game, and yes, the aforementioned burger. You can read a full transcript of our interview below, which took place through translators and has been very lightly edited for clarity.
IGN: So to start, in your talk, you spoke a lot about the things that you wanted to let players do with destruction and voxels, but what about the things you don’t want them to do? When you make everything destructible, how do you design to keep players out of things that you don’t want them messing around with?
Kenta Motokura: So basically we just designed the gameplay such that the player couldn’t reach that area or couldn’t do that thing. So unlike a lot of previous game design that we’ve experienced, there’s a lot that we can’t anticipate about the player experience in terms of what they’ll try to do. In that sense, we have to give them the play space they can enjoy and everything else would be essentially unreachable. So for example, in this game, there are some surfaces that Donkey Kong cannot climb and that’s something that we might not have chosen to do in previous games where we wanted Donkey Kong to be able to climb things. So we would leave that up to the engineers to make those kinds of solutions. And again, because we want the player to be able to freely destroy and freely enjoy the game, we’ve created the abilities and the kind of scope of what the player can do to match that.
Do you find it as getting increasingly challenging to keep players from doing what you don’t want them to do? I think about in Odyssey, you put coins way up high because you knew that players were going to find all sorts of ways to get up there. And I think about people in Donkey Kong Bananza have found ways to use the terrain jump to get all the way across the map where they’re not supposed to be.
Motokura: Yes. To answer your question very briefly, it is getting very hard to keep players from going all over the place. But certainly we take those sorts of things into account as well. Sometimes there are sequence breaks in game that you can, once you learn about them, design around so that there is a gameplay experience on the other side of that sequence break. And certainly when we see players who actually get to those areas and experience those parts, we look around at each other and say, “I’m really glad we made that.”
You talked also about having to design not just the external parts of levels, but the internal, because with voxels, players can get inside of things. Are there any principles you have to follow when designing internal parts of levels or rules to make sure that players can actually find what you’re hiding in there?
Motokura: Great, a level design question. So rules, yes, I would say there are certainly ideas that we work with. Certainly all of these structures in the level are designed such that when the chain of destruction is functioning, you end up somewhere interesting. And we also have the ability in this game for players to use the DK hand slap, which has the sonar function, allowing them to see buried objects and objectives within the terrain and adjust their gameplay as they go forward. But it’s hard to create levels in that way because these are all things that you’re normally not expected to be able to see through the terrain. So you have to think in a slightly different way to approach it.
You gave the number of voxels in the canyon layer as over 300 million. Do you know how many are in the whole game?
Tatsuya Kurihara: I actually do not know. But I can say that the canyon layer is a pretty big layer, and so there’s no layer that has more than that.
So Canyon has the most of any?
Kurihara: Maybe not the most, but it’s in the upper limits.
Were there any challenges you encountered related to the way the game remembers every detail about how a player has destroyed a level? So if I break open the ground here and then I run way over there and play for an hour, it remembers and keeps that destruction where I started. Were there any challenges related to making sure the whole level would remember that information?
Kurihara: So when it comes to kind of bring this idea of a level remembering what a player did, it really comes down to memory. And so this was mentioned in the panel, but because we’re able to now develop for the Nintendo Switch 2, it gives us more bandwidth to be able to kind of remember, like you mentioned, something the player broke an hour ago.
Can you talk about making the Switcheroo Goo out of voxels? How did that come about?
Kurihara: So that idea came from some of the prototyping that we did within the team. And so we were trying to brainstorm what other ways, one of the fun ways we can use voxel, and that was an idea that came up in the prototype phase and that made it to the actual product.
Because the Switcheroo Goo is so unique, did using that in levels and puzzles present any unique challenges because you have voxels disappearing out of nowhere?
Kurihara: So first of all, the idea that voxels can kind of appear out of thin air, that’s something that can actually happen in other places in the game as well. So how we deal with that is different and varies depending on the voxel. For example, there’s a voxel that kind of rises up, and how we make that work is that we push up the voxel that’s on top so that this space gets filled up. And so there’s also those times where DK himself might be buried under voxels and we made it so that DK then can destroy himself out of that hole.
What about the salt that dissolves the acid? That was another really interesting type of voxel that I was curious where that idea came from and how it was implemented.
Kurihara: And that’s also a product of the many various prototyping we did that resulted in things like the switching and the salt as well. When it comes to the salt, what we were looking for is maybe various ways of destruction. So obviously DK can punch and destroy things, but we were looking for other ways. What other kinds of destruction can we reign on this, the terrain we have? And dissolving something was an idea that came up, which resulted in the salt voxel.
What went into consideration in designing the Feast Layer of Donkey Kong Bananza, after you had done Super Mario Odyssey and the Luncheon Kingdom there, which also involved voxels? Was there a lot of work that was put in to distinguish the two as different from one another?
Motokura: So talking about the Luncheon Kingdom in Super Mario Odyssey, that is an area that has a lot of lava, and our approach to it was actually started with, okay, what sort of kingdom could we build to incorporate a lot of lava? And once we had the voxel gameplay-based ideas present in that kingdom as well, it really started to come together around the theme. And when it comes to Donkey Kong Bananza and the Feast Layer, it actually started with a prototype where an artist had created a large hamburger out of voxels. I saw that and I thought, that looks delicious. We have to put it in the game. I especially thought that the bacon on the hamburger looked really good. I happen to be a really big fan of American-style bacon. In fact, I ate some today. But this is a great opportunity for us to bring together the hamburger prototype and the other prototype of the salt mechanic and put them in one place.
In your talk you spoke about each voxel carrying a lot of different data, like density, wetness, things like that. What work went into ensuring each different type of terrain felt very unique to the player as they destroyed it?
Kurihara: So I’ll start by explaining about one type of information or data that can be contained in a voxel, which is material properties. And that information then contains information about, say, how many punches it takes to break or destroy that material, or if it will dissolve or break away if the character touches it. And so we thought that by giving these voxels different types of material properties, we can create different gameplay experiences. So in the game, there are voxel-clad enemies that walk around. And when you look at something that’s covered in sand, it’s very easy to defeat, but when that voxel then changes to lava, it certainly becomes a much more challenging ordeal.
What makes something satisfying for the player to destroy?
Motokura: Well, that’s a hard question because I think it’s not just one element. There’s a lot that really goes into the satisfaction of an action. There’s the lead up into an action like a punch. There’s the sound it makes, the camera angle, the vibration. A lot really goes into it. And then it comes down to what are you punching? Is it the ground or is it an enemy? What sort of reaction do you create?
How do you test for that?
Motokura: So if it comes to one of these elements like the camera shake, that’s something that I actually check personally. But once you have all of these elements of the structure in place, it still really comes down to, does it feel right? And so we have to play with that in mind. Are we creating a satisfying destruction experience?
Kurihara: And when it comes to implementation in the engineering side, it’s really the experience of playing the game. Game developers play the game just as much as they develop. In fact, so we spend our days developing and breaking and destroying all day long, and that’s part of the experience of how we get to the gameplay experience that we can then provide to the end user.
I did spend quite a bit of time taking a look at sandboxes in parks.
Did you ever have to go outside and dig in the dirt?
Motokura: Actually, it’s funny you say that because I did spend quite a bit of time taking a look at sandboxes in parks and watching people experience that. And so, one thing that I took away from that was how wet the sand is and whether you can dig in it. And so moisture became one of those data points that we wanted to express.
Kurihara: Yeah. So as Mr. Motokura mentioned, in nature, if you keep digging, you’ll find wet or moist ground as you’re digging, or if you take a tree, if you cut into it, there’s water streaming through the tree. So we took hints from nature itself and tried to bring this world to life in the destruction that the player unleashes in the game.
I wanted to ask about one more specific type of voxel. Can you tell me how you came to the idea of the Liftoff Ore?
Kurihara: That’s also another byproduct of the prototyping that we did. And so I think it was mentioned in the panel that we thought about using voxels, for the player to pick up the voxels and then powering up. And it was an extension of that idea. And one way to power up is to grant the player the ability to fly, and that prototype resulted in what you see in the game.
What is the prototyping phase like? Like what determines this is something we want to use versus this is something that we cannot or do not want to use?
Motokura: So when we’re in the prototyping phase, we have to consider each idea in the reference of the game as a whole. What kind of gameplay experience could this add? Where could we use it in the game? So you have what might be an interesting idea. You create a very rough prototype and you place it in the game and see how it is.
Were there any prototypes that did not make it in that you wish did?
Motokura: We got pretty much most of them in, I would like to say.
Is there anything that players have done since Donkey Kong Bananza released that has surprised you or that you didn’t think was possible?
Kurihara: I’ll first start off by saying that it seems like everybody’s enjoying their own brand of destruction. So one example is I see that there’s way more people than I have imagined who would go to a layer and just break everything on the stage. I mean, it takes a lot of work, but there’s people who’ve done it.
Motokura: One thing that really surprised me, and this is maybe going back to the discussion of the sequence break that we had a little bit earlier, was the surprising ways that people are using voxels for movement, not just double jump, but other movement techniques entirely that they discovered on their own to get to some very interesting places.
Now that the game has been out for a while, can you explain the timeline? Is this a Mario Odyssey prequel?
Motokura: I was wondering if that question would come up. And I hate to disappoint you, but I’m hoping that we can leave this to your imagination rather than sharing any details at this time.
All right. How was the decision made to bring back K. Rool after all this time?
Motokura: Sorry, this is another one where, because perhaps some players haven’t gotten to that part of the game or played yet, I wouldn’t want to touch on K. Rool just quite yet. But certainly, I would say that thinking of Donkey Kong’s long history as a character, we’re always thinking of ways that we can surprise and delight players old and new who are coming to a Donkey Kong game experience.
I did want to ask a more design-oriented question about the ending, so maybe we can talk around it a bit. The ending sequence, the final fight and everything is designed in a very satisfying way and the pacing is very precise so that you have all these emotional beats as you’re moving through it. I’m wondering what goes into designing something like that where you’re anticipating how the player is going to feel in a given moment and this heightened period of emotions.
Motokura: Maybe we can address that question through just some of the basic ideas of Donkey Kong Bananza. So certainly when we’re approaching game design, we are considering it from the point of view of a lot of different elements. There’s the emotional landscape that you describe. Also, how characters change over the course of the game, how the landscape itself changes through level design, how that affects the player’s pace as they proceed through it, how it affects difficulty of the player experience. These are all different elements that we have to think about in that game design space.
Do you find that players increasingly want their games to be more open-ended and have more freedom just generally, or is that just something that worked very specifically for Donkey Kong Bananza?
Motokura: Certainly, I do think that gameplay experience changes over time. And we saw that in our own experience, even transitioning from Switch 1 to Switch 2, where new kinds of play became possible. So you’re always looking at how you can incorporate those ideas, but make the most out of new possibilities, new technology.
This [DK] is different from the Donkey Kong character that you see in the movie because the one that you see in the game, in a sense, is a double of the player.
So you discovered this voxel technology while working on Mario Odyssey and then turned that into a new game. Did you discover any new technologies while working on Donkey Kong Bananza that you’re experimenting with?
Kurihara: So the honest answer is I don’t know, because even when we were using voxels in Super Mario Odyssey, we didn’t imagine that we would create an entire game based on that technology. And really, it could very well be that we take some of our learnings from Donkey Kong Bananza and leverage it for a future project, but really we won’t know until we actually get down to making it and really ultimately depends on the kind of gameplay experience we want to create and what technology we employ to realize that.
Are you in regular communication with other studios within Nintendo as far as sharing things like the voxel technology and other things that you discover as you work on them?
Motokura: Yeah. So as it becomes necessary in the course of development, we will share technology or skills learnings. There’s certainly times when our team took some of the work that other teams have done and employed it in our project and there’s been other instances where we’ve been able to provide that technology. So that’s something that we can do continuously.
Are there any examples you can share of other things you took from other games for Bananza?
Motokura: Sorry. So it has to do with other titles that we’re not a part of, so it’s hard for me to respond here.
Fair enough. Donkey Kong Bananza came out at a somewhat similar time to the Mario movie and in Donkey Kong Bananza, Donkey Kong doesn’t talk, he just makes monkey noises, but he does talk in the movie. In your head, does he talk?
Motokura: So I would say yes. In my head, Donkey Kong does talk because you come upon these choices in the game and the player has the opportunity to see Donkey Kong’s expression and how he might be feeling about those, and then the player can make their choice. But also, this is different from the Donkey Kong character that you see in the movie because the one that you see in the game, in a sense, is a double of the player, and so they don’t speak, allowing the player to project a little bit more. But I think that is made up for by the fact that Pauline as a character in the game does speak. And so you have the opportunity for this speaking character to deliver hints to the player. Also, as she expresses her feelings, that’s something that the player can empathize with, or they can feel the gratitude that someone is expressing what they’re feeling in the game.
So if this Donkey Kong spoke, he would not sound like Seth Rogan?
Motokura: Yeah, I can’t really imagine so.
You said in your talk that it’s more fun to destroy things that are beautiful. Why do you think that is?
Motokura: That’s a tough question. I think part of it is the surprise element, something that was so elegantly put together, beautiful. When it’s destroyed, it goes to nothing. And that gap between these two states is what creates delight and surprise, I think.
Did you ever during development go to one of those destruction rooms where you get a bat and you just break things?
Motokura: No, I’ve never been, but now that you described it it sounds kind of fun. We would’ve wanted to go. So do they have a lot of beautiful things for you to destroy?
Some, but I also think you can bring your own things. So if you want to destroy your own beautiful things, you can.
Motokura: Interesting.
One last question. Donkey Kong ends his adventure and I won’t spoil it for people, but what do you think Donkey Kong is doing now?
Motokura: Certainly eating bananas.
Rebekah Valentine is a senior reporter for IGN. Got a story tip? Send it to rvalentine@ign.com.
