John Lau, lead producer at ustwo games, discusses how the studio fostered collaboration between artists, designers, and programmers for optimal productivity and visual fidelity.
John Lau, Lead Producer, ustwo games
My name is John Lau. I am a lead producer at ustwo games. Monument Valley 3 is the latest in the Monument Valley series, and it is the story of Noor, who is an apprentice to a lighthouse keeper, and players will need to guide Noor through a world of impossible architecture and optical illusions in order to rebuild a community after a flood.
With Monument Valley 3, what we really wanted to do was try and broaden out the gameplay space a little bit more, so you'll see, in addition to new mechanics in our chapters, we've included new ways of getting around. So, we've kind of broadened things out with a boat that you can sail, as well as a new story with new characters.
The thing about Monument Valley and Monument Valley 2 was that gameplay was centered around identifying paths and walkways and, you know, joining up walkways in impossible ways.
That encourages the player to really focus on identifying lines and paths on the architecture itself. But now that we have a boat, there's an entirely new way of thinking about the space around these architectural structures, because that's now fair game. So I think that really helps the player to think about the spaces in a different way.
Integrating the look and interactive design
So, one of the team's biggest challenges is that when we make a Monument Valley game, one of the biggest guiding principles is that we want every level to feel as though it is a piece of art that you could print out and put on your wall. Now, that is also true, as well as us wanting to make them compelling puzzles and interesting worlds to kind of move through. This means that both the interactivity and game design of the levels themselves and the look of them has to be very, very closely integrated.
Often, you'll find that you're not really sure whether a level looks the way it does because it has to play in a certain way, or whether it plays the way it does because it has to look a certain way. That involves incredibly close collaboration with the art and design departments, and the biggest challenge is that, often, game development processes don't facilitate that kind of back and forth particularly well.
What you want to do, or the most efficient way of doing things, is to have game design, white box, block out a level, and then pass it on to art and never think about it again. But that's not really possible when you have art and design so closely intermingled.
The way that affected the team and development is that it became very difficult to predict how quickly you're going and how well you're doing in terms of making the amount of game that you want to make.
When you are constantly iterating in this loop between passing a level between design and art, you don't know how many rounds you're going to go. You want to close that as much as possible in order to make sure that you can actually make enough game and make enough good game, you know, to live up to the Monument Valley promise.
Maximizing tooling
Overcoming those challenges was a multi-pronged approach. I mean, we fostered collaboration in lots of very varied ways. I think a really important part of it was the way in which we used our tooling in Unity to facilitate that kind of collaboration, and we've had a long time to perfect that process, because Monument Valley, the first Monument Valley was released in 2014.
It's been maybe 10, 11, or 12 years since development started on that game, and we've had time to iterate on the codebase and develop the kinds of processes and tools that might make that collaboration easier.
So, in the case of level design and art, how that's evolved over the three different games has been very interesting, but it's always been with the view to closing that loop and making the barrier indistinguishable between design and art.
In the first Monument Valley, all of the paths, all of the kind of blockouts were made with Unity cubes, and that's, you know, cubes stuck together, kind of arranged in the level in order to create the blockout.
Now, each of those cubes is an individual GameObject with nav on it as a separate GameObject. So, what you'd have is a scene full of hundreds of GameObjects, which all need to be kind of managed and kept track of.
Then, when it came to applying the art, what we had was an automated process which would generate quads on top of the visible faces of those GameObjects or of those cubes. Then, those meshes were combined, sorry, those quads were combined into meshes that we could then begin the art-ing process with.
Now, that is not a particularly reversible process, so what we'd have to do with the first Monument Valley was we'd have to say, right, this is a design complete level, move it onto art, which may well be a familiar workflow for, you know, another different kind of game.
With Monument Valley, where there has to be a lot of interplay between design and art, that gets very tricky because if an artist, say, wants to make a walkway a couple of units longer in order to make the composition look a little better, that's a difficult situation to kind of reverse.
So, with Monument Valley 2, what the team did was, inspired by ProBuilder, create a kind of cubic modeling, a cube modeler, essentially, in Engine, so that you could have a game designer extrude and essentially shape a cube-based form in Engine.
That was all one GameObject, and that was the same GameObject that an artist was then able to art and apply textures to, and embellish with all the things that make a Monument Valley level look beautiful. So, if an artist, at that point, wanted to say, you know what, can we add a little bit here, or can we adjust it slightly whilst keeping the puzzle intact, that was an entirely possible thing to do very, very quickly. So, when you have an artist updating the same GameObject as the designer, it's very difficult to tell where design ends and art begins, and then you're automatically kind of closing that loop, which is brilliant.
With Monument Valley 3, we've built on that, and when we’ve got that initial kind of decorated base structure, we updated our decorating tools. This meant that we could have an artist spend, you know, a couple of days making lots of different kinds of roofs, and then import those roof meshes into the game. And then, just kind of use them like a brush, as a template, to just dot them around, bop, bop, bop around the level in order to create this wonderful, architectural kind of effect.
Over the course of the three games, you've got this gradual tendency to collapse that barrier between design and art so that you can iterate many, many times over. Once you go around a couple of times, it starts to become very unclear who's designing and who's art-ing.
Play Monument Valley 3. Find more Made with Unity titles on our official Steam Curator Page, and visit unity.com/resources for more content featuring Unity developers.
