Duck Reaction #3 |Construction et shaders
Welcome to another episode of Tales from the dev from the Duck Reaction studio. In this third instalment, I catch up with Mathieu and Simon after leaving them shortly before the release of Turbo Live (here). This time we’ll be discussing the reception of Turbo Live and their new project, Clay Builder. We’ll talk about the transition between the two games and the new skills they’ve acquired, such as voxelisation. But I’ll leave you to discover this new piece of history from a French independent studio.
Inksushi : Good morning, gentlemen. This is the third time we’ve come together to review the progress of the Duck Reaction studio. First question: how are you?
Simon : Hello, I’m fine.
Mathieu: Hi, very good, thank you.
Inksushi: Last time, we stopped just before the Steam Next Fest in November 2024 and talked about the hassle of setting up a Steam page. Turbo Live was about to be released. What’s happened since then?
Mathieu: It seems like ages ago. Simon has a much better memory than I do.
Simon: So we’ll be back at the end of October, beginning of November. In the meantime, we’ve released Turbo Live for good, and we’ve noticed that streamers are finding it very appealing. On the whole, we’ve made some hotfixes and updates to make the game as simple as possible. For example, Mathieu introduced a whispering system to avoid spamming the chat. All the viewers had to do was whisper to the bot to carry out the desired actions. In the end, the players didn’t like this feature, the bot worried them, it was too laborious. They preferred to spam the chat. We decided to keep the two options to give them a choice. The game, which was released on 11 November 2024, is just a patched version of the demo. We wanted to see how it went. We still had 1,000 wishlists, which isn’t bad for a project with little communication. The game has also been enhanced with free content.
Mathieu: For us, it was good progress, in the sense that Meaningless, our first game, made 1,000 wishlists in a year, with an appearance at the Tokyo Game Show to boot. With Turbo Live, we reached that threshold in three months. We’re progressing at our own pace.
Simon: Mathieu and I agree that streamers are having fun and that’s cool. But we don’t have the visibility we need to reach more people. At the moment, two or three people are sending me messages to say they love the game and asking us to add features or modes. We can see that we need a good update, a big version 2.0, to put in place tools to create as many modes as possible to strengthen the party game aspect. We need to improve the link between the game and Twitch, and even open it up to all platforms. In short, we need a more powerful and modular tool. All this to say that we have a roadmap towards a V2 with no end date, but it will happen. The main variable is the number of players it impacts. If, one day, a big streamer takes an interest in the game and it spreads, we’ll work on this famous update. As the game is free-to-play, we’re just waiting to see. We’re also going to start working on an itch.io page, because free games sometimes gain visibility on that platform, so why not. Here’s the plan for Turbo Live: we’re leaving it aside, unlike Fluffy Doki Sunshine, which we’ve cancelled. With Turbo Live, we still have a mature version. We’ve had some very positive feedback, due to the free-to-play nature of the polish. We’ve reached twelve reviews, Meaningless has four, so we’re still making progress.
Mathieu: It’s not a huge improvement, but we’re getting better. Players don’t spend a lot of time in the single-player mode, but a streamer can easily play for two hours with his community. It’s free and fun, and the game is perfect for a fun evening out with your viewers. If we can see that a lot of people are buying into this first version, with its positive reviews, we’ll be able to say that we’re going back to development.
Simon: Ultimately, it’s a bit of a mobile game strategy. We offer a free game that we let live and if it works, we develop the idea.
Mathieu: For our part, we also have to make an effort. We have to keep talking about it, even if we’re working on something else. A lot of people just haven’t heard about it. We really need to think about communication across several projects.
Inksushi : Turbo Live represents quite a lot of work. The project has enabled you to acquire new skills, perhaps even new production methods. What do you get out of it?
Mathieu: From a development point of view, our main point of evolution has been to get to a stage where Simon can become autonomous when it comes to creating mods. It probably took us too long to get there. Simon needs to have plenty of tools to create quickly, and I’m really in charge of the underlay. In other words, I have to provide the tools. The problem is that we’re starting to have a lot of things to do, but the game has already been released, so it’s too late. We had to ask ourselves whether we should invest more time, or whether we should have been quicker. For the next game, we’re paying more attention to that. For my part in particular, I’ll be making sure that I provide Simon with the tools he needs to prototype more quickly. We’ve looked at lots of things so that he can do little things on his own. He can test things without us even needing to develop them on purpose. It’s not very optimised, but it allows him to work without me.
Simon: We’ve invested in quite a few plugins, including visual scripting. We’re a lot closer to Unreal now. This allows Mathieu to structure things, to make clean code, to make sure things run smoothly. As for me, I’m doing a bit of messy stuff on my own. It allows us to test and choose what to keep. Because there’s only one developer, we have to optimise what Mathieu spends his time on, because he won’t be spending it on anything else. We can’t really afford to throw away the fruit of a week’s work because we realise it’s not what we want. We try to minimise the risk. Typically, on the new project, which we’ll talk about later, I wanted to test two types of movement. I did my fiddling, we tried it out, we confirmed our choice and it’s done. So it’s messy, but it saves a lot of time.
Mathieu: That’s perfect. Simon just does a little thing and we see if we can go further or not. With that alone, we’re improving our development process. That was one of the biggest problems we had with Turbo Live. Another point is the Twitch part, which is a real pain, but we’re going to continue to integrate features for streamers. That’s really the thing we want to keep because we loved implementing it in Turbo Live. Next time, we’re going to try and reach out to more platforms. For example, a game in which viewers take part using their phone via a website, like quiz games. In that case, we’d be independent of Twitch, so players could use it on Youtube, Discord or on the sofa with their mates.
Simon: We’re happy to have made a game that’s really stream-based, but we’ve realised that it’s also very limiting for solo players. For future projects, what we’ve decided is not to go that far. Of course, we want to make games with stream-specific mechanics that have been thought out in advance, but the single-player experience must still be possible. You have to be able to have fun without streaming. At the end of Turbo Live, I started to prototype an idle game in modelling clay. So I tested out a little Tamagotchi concept with some rather dodgy creatures. Mathieu and I decided to go ahead with it without any particular mechanical ideas. We’ll see what happens. The idea is to have a good time with the creatures, take care of them, make them evolve or mutate. Then you start to include exploration to meet other characters and get new mutations. There’s a little Animal Crossing inspiration here.
Mathieu: You can put the game on in the background, watch the world evolve and just adjust the volume of the rain and the music. The basic idea is to have a relaxing game while doing something else.
Simon: Exactly, nothing to do with Turbo Live. Two rooms, two atmospheres once again. But we’re coming back to something closer to Meaningless, so there’s a bit of both games in the end. So we started prototyping, testing animations. Modelling clay works well, we like that. We also got people to try it out. We started thinking that it would be cool to be able to build things, because that’s what modelling clay evokes.
Mathieu: It was a comment that led us to this, it didn’t come from us. Modelling clay is just fun to manipulate and create with. In his prototype, Simon was building a character, a house.
Simon: We posted a video and someone commented that it made you want to build and manipulate. It had to be more than just a theatre to watch. So the first question we asked ourselves was: did it seem feasible? To make up our minds, we made a prototype and a short video that we posted on Reddit and other networks as a proof of concept. We’ve also sent it to Wholesome Games, because that’s really their style of game. At the moment, we don’t have anything playable, but the feedback is already very positive. People want to play it, and they’re really looking forward to the management aspects in particular. All that to say that we were hesitating between this and a horror game, and the public decided in our favour. So the next game will be a little clay builder.
Inksushi: I’m just going to take a step back before continuing with this builder. Simon, do you feel that this independence in prototyping gives you more freedom as a designer?
Simon : Absolutely! There are some bigger productions where two people work together. A programmer makes a few tools and a designer tests them, so there’s an exchange that takes place. As there are only two of us, I can’t monopolise Mathieu and interrupt him while he’s working on something else, like the Twitch interface for example. There was a feeling of guilt. Me, because I was taking up his time, and Mathieu, because he wasn’t allowing me to move forward. Now I can test things quickly and really design. Doing it on paper is nice, but limiting. Mathieu and I test things out and agree on a choice. That takes a load off my mind too. Sometimes I ask myself questions that I can’t answer, so we test and see what happens without really having to implement it. Not long ago, we thought of including a dialogue system without really knowing if it would work, so we tested it. I think it helps creatively. Everyone can concentrate on their tasks and the project evolves in the right direction.
Mathieu: It removes a bottleneck. The code is the most time-consuming part, there are lots of things to do. Very often, I block Simon because he has a need that I can’t meet immediately because I’m working on something else. This freedom allows him to start making prototypes, even if they’re not finished, that we can show. He can do that on his own, testing things, whereas I’m correcting things on Turbo Live. For me, that’s the big positive. We can even get people to test the prototypes, so they can project themselves and give us feedback.
Simon: I can really see the difference in the discussions: you have something concrete to build on, whereas in theory there’s always a bit of a grey area. And you have to solve the basic problems before you go too far. I think we’re making good progress.
Inksushi : And what about you, Mathieu, does that make you a little more sure that you’re spending your time wisely? So that you don’t think you’re working on something that might be thrown away in a fortnight’s time?
Mathieu: Clearly, that’s it. He’s tested it, we’ve decided, and I don’t have to clean it up any more. It allows me to build solid foundations, but sometimes it lacks modularity. Simon can easily plug something into my code with visual scripting or code a few lines. He really does have a solid foundation on which he can build and iterate without me. On the other hand, as a Unity developer, I find it too complicated to make a game completely using visual scripting. In my opinion, you still need a code base, just to automate certain tests. Gamers don’t know it, but the heart of a game is often tested by the computer, because otherwise, as soon as you move a line of code, you risk breaking something. With visual scripting, you’d have to do it by hand.
Simon: And, once again, there are only two of us. When you have a QA (quality assurance) team, you can afford to be more flexible.
Inksushi: You said that Mathieu had developed tools for Turbo Live. Are you reusing some of them on the new project?
Mathieu: Yes, we even have tools from Meaningless. That’s why we’re moving faster and faster every time. When we develop a tool for a game, at the end we ask ourselves whether it was good or not. If it was, we keep it and develop it according to our needs. On each project, we keep small bits of code, which allows us to have a solid base to go faster in our developments. There’s always a part that we throw away, either because it’s not useful, or because we thought it was a good idea and in the end it’s not very good.
Simon: I think that after the next production, we’ll have a good feedback tool, one linked to streaming and also visual scripting systems. We’re starting to look pretty good for two-person projects.
Inksushi: If we go back to the “modelling clay” project, does switching to a construction game bring new challenges, like imagining what players would like to build with the tools you give them?
Simon: From the moment you create a game, the designer’s role is to put yourself in people’s shoes and think things through a bit in advance.
Mathieu: The hardest thing is to scoop, because a builder can quickly become gigantic. There are some amazing ones on Steam where you can do tons of stuff, but that’s not what we want. We don’t have the money. We need to come up with fun ideas. The player shouldn’t feel restricted without the project becoming too ambitious.
Simon: For the moment, we’re going to try and keep it as simple as possible. We’re taking our inspiration from Womp, an online 3D modelling software, which tries to be very accessible. We’re going to build on things that already exist, because our aim is not to reinvent the wheel. As for the scope, we’ve given it a lot of thought. To create a theme, we need an environment with a neutral zone in the centre, which will be the construction area. The elements created can then be placed in a library to create the space you want. It will also be possible to build characters that will be governed by a skeleton system. The game will ask the player what they want to build in terms of a creature, a biped for example, to provide an entire skeleton. It’s a bit like stop-motion. The player glues on the modelling clay, and the rest is done by procedural animation.
Mathieu: That’s the goal we’ve set ourselves and we’re not quite there yet, but we’re making really good progress.
Simon: We did a lot of research into skinning to find the best solution. But before we even thought about that, we realised that building objects means assigning functions to them. A balloon and a house are completely different in size. And if there’s a library, that means drag and drop and orientation. That’s a lot to set up. We’ve gone back to the basis that makes the game original, which is building creatures out of modelling clay. In the end, we’re heading more towards something akin to Spore, which is what really hooks people in. There aren’t too many projects that resemble it, and that’s one of the game’s strong points. I’d sent a message to Mathieu to take up the idea of the Miis, which had a big impact on everyone, with a zone where you put lots of different characters. In fact, it’s going to be a game in which you repopulate a small village that already exists, but which is empty. There will be little missions to create a new character. First a snail, then a biped, a humanoid to which the player gives the appearance he wants. All these creatures will lead their own little lives. We’re coming back to the idle game in a way, it’s an in-between. If you think about it, you’ve got an island and little houses, but if all the quests consist of designing a creature, it’s going to get redundant very quickly. It would be fun to add requests from the locals, like a letterbox in front of a house. In reality, whatever is proposed will work. You have to assume that it’s not limiting. In the same way, you can imagine a table where residents come to eat, but if you make a car, they’ll eat it. That’s cool, because it removes the whole layout aspect, which is pretty heavy in terms of code. The game has a modelling clay editor, so you might as well make the most of it. And it allows you to create little scripted quests that will add a bit of life and interaction with the creatures. It makes the project a little cuter than you’d imagine, and we’re back to the basic idea of developing an island and giving the game a soul.
Inksushi: How do you achieve that plasticine look? There’s the texture, but also the physics. Was it something preconceived or did you have to design and adapt it?
Simon: Before creating the creatures in modelling clay, we did some tests with Mathieu. We have a shader for pure rendering and a plugin that lets you merge shapes, like mixing two balls with more or less blend. It works pretty well. Everything is calculated on the graphics card. To refine things, we add extra layers to create stop-motion effects, cracks and fingerprints. All this is generated procedurally. Once we were happy with the result, we asked ourselves what people would like, and we came to the conclusion that they would want to add detail to their creation. So we need quite a lot of polygons, because if your character has a million polygons, the player can’t have four million. To compensate for this, we use LOD (level of detail), which reduces the number of polygons according to the distance from the camera. The further away a character is, the lower the resolution. We tried a lot of things to automate the rendering without being too greedy or breaking the animation. In fact, we kept the procedural animation technology from Turbo Live for the feet that move by themselves. With all this, we’re starting to get a convincing result with physics and ragdoll. The world comes to life, the creatures move and everything is procedural and runs in runtime.
Mathieu : There’s still one big problem to deal with. When the player has drawn his character in clay, how do you ensure that the clay moves correctly during an arm movement? There aren’t really any tools that allow you to do this at runtime. It’s often done beforehand. Blender, for example, allows you to associate skinning with the skeleton, but not directly in-game (when creating the character). We’ve found a few articles by people who have done algorithms, but still not at runtime. I looked at Blender’s source code, took bits of code from free projects on Github and tested the most naive things possible. Little by little, I’m improving my algorithm by adding things here and there, which isn’t always conclusive. For the moment, the most relevant solution is to use voxelisation, a technique that involves cutting a 3D object into a set of 2D layers made up of cubes called voxels (like a 3D printer). I transform my character, created by the player, into a bunch of big voxels like Minecraft, which makes things a lot simpler. The 3D model becomes a series of cubes and it’s very easy to tell if a cube is next to another cube, if it’s touching a bone, if it’s far away. Once you have your data, you apply a physics algorithm that will propagate changes like heat through the material, and then you apply the result to the original object, i.e. the creature. There are other people who have done similar things, but you can’t find the codes or they’re only prototypes, so there’s a lot to code yourself. At the moment, I’m missing the last stage, so the creatures are a bit deformed, moving the left leg moves the right leg, it’s a bit messy. The big difficulty, when you come up with this kind of algorithm, is knowing why it doesn’t work. The result is a table containing several thousand points, just coordinates, but finding the error inside is like looking for the wrong grain of sand on a beach. I’ve been working on it for quite a while, and that’s the bulk of the work. And that’s why it’s important for Simon to be able to prototype, because it takes up a lot of my time.
Simon: You should know that this voxelisation method is used for auto-skinning in Mixamo. In Blender, some people complain that auto-skinning isn’t very accurate, so they use the voxel method, which is much more efficient, except for fingers for example, but we won’t have any problems with that. As far as the body is concerned, it works extremely well. It’s not perfect, but if the players manage to make a nice humanoid and it animates well, that’s cool.
Mathieu: Blender’s method is good, they use the same principle of heat propagation that I was inspired by. But there’s also a lot of calculations involved in shooting rays from points and that requires a lot of resources. Still with the idea that our players don’t all have big machines, we can’t afford ten-minute loading times. Especially for character creation, which requires a lot of back and forth, people try things out and then change them because they don’t like them. We keep in mind that the experience has to be fluid for everyone. We work on big machines, so we have to be vigilant. From a technical point of view, the latest big advance is the management of all the voxelisation by the graphics card using a Github project that wasn’t yet operational, but which provides a good basis. I’ve used it to switch everything over to the graphics card and it’s almost real time. The time saved in executing the algorithm is enormous.
Simon: The next step for us is to offer a small demo containing a biped or a humanoid on which we can add modelling clay and see it articulate. That’s what we’re aiming for in the first test version, to see if players like the manipulation.
Inksushi: And do you have a name? How far do you want to go in terms of scope?
Mathieu: The code name is Clay Builder, but there’s no official name for it at the moment.
Simon: I’m going to read you what we’ve written down. We’d like to propose a game in which we’d have a scripted tutorial, a presentation of the features via dialogue. Given that it’s on an island, we want to do a little story with a tale or an exchange between characters to guide the player. A bit like Minami Lane, we’ve tested lots of cozy little builders to get an idea of the atmosphere and a character editor. So, we’d like to be able to integrate the characters, obviously, have a predefined neighbourhood that would be a small island, and be able to create an object that is automatically placed and replaced. The character will be able to interact in a limited way with his environment, wander around, enter houses, take a nap, and so on. Basically, the character will live his life. There will also be interaction between characters. New areas and skeletons to unlock through objectives. You start with a slime, then molluscs, and so on, opening up new possibilities as you go along. We’re also thinking of a photo mode, so you can share your creations in the hub or in a photo studio. For streaming, we’re thinking of a viewer who could take control of a character. And as a bonus, we’ll be able to edit an object rather than replace it completely. We’ve also thought about a parenting system, a home system to create families and the possibility of infusing a personality, like in The Sims. After that, we’ve also got player interaction with the environment, shaking a tree to make apples fall, a day/night cycle, and weather, a bit like Animal Crossing. And if the game works, we’re thinking of events linked to the calendar, and customising skeletons, like adding arms.
Mathieu: We’d like to go back to the Spore principle by assembling arms and legs to give us greater freedom. If you want to add wings, you model them and add them. It’s also pretty much inspired by A Short Hike, with an island that’s not very big but well designed, and we want it to be fun to explore in a cool atmosphere. It’s a game that really made an impression on me, one that I played with a friend. She played and I watched. That alone was really enjoyable.
Simon: For the loop, the idea is to start with a little creature that’s already there, who you’ll have to make a friend with, a slime, which gives access to the next zone. It’ll be a bit scripted. You’ll be able to wander around in little biomes where you’ll come across other creatures to build, and you’ll be able to switch characters. For exploration, there will be collectables and secret passages. There will be doors that require a certain number of creatures to open or a certain level of happiness. The quests will require you to build things, which will make the inhabitants happy. In short, you do the quests to increase the number of characters and their happiness so that you can move on to the next zone. There will be a very slight metroidvania aspect to unlock secret passages with this or that creature, which will be optional and not explained to encourage curiosity. All this with lo-fi in the background to create a cool atmosphere.
Mathieu: As a gamer, I love the concept of construction games, but I was a bit worried about the redundancy of not having any objectives.
Simon: With modelling clay, we want to keep the connection with childhood. A child who plays at sculpting doesn’t stop to contemplate his work once he’s finished, he wants to put it into action. In our discussions, we felt that it made more sense to make an effort on small quests, which can be done in visual scripting, which suits us, than on the layout which, once finished, doesn’t allow the player to enjoy it. We prefer the Animal Crossing aspect, where you invent stories that you can put into situations in the game. And we save ourselves a lot of trouble.
Mathieu: It adds life to have creatures that come and chat with you, and we avoid being in direct competition with games like Tiny Glade, which use procedural creation and are very accessible, and quickly very beautiful, because everything is calculated. You can make beautiful or ugly things, it doesn’t matter, because you’re wandering around in your universe, there’s life, the objective is different.
Simon: And how would you like to be as a player?
Inksushi : I’m not going to hide the fact that this isn’t my preferred style of game, I’m more of a rogue-lite player. But I do know that I have a few games in this genre and that, when I’m working on an article, for example, and I need a break, I’ll build something for half an hour and then relax a bit and get back to work.
Simon: That’s what we had in mind, already with the Animal Crossing inspiration and the minimalist side of Minami Lane, not much content, but a bit of scenario and cute interactions. That’s enough for the amount of time you spend playing, and it’s really satisfying to reduce the game to a small neighbourhood. A bit of storytelling and you spend thirty minutes relaxing. But if you want to do things right, create a beautiful environment, that takes time. What reassures us about our game is that we estimate that you can have up to twenty characters, but if you want to make them with details, that can easily take time.
Mathieu: And that’s not counting the objects that you might have to redo to keep things neat and tidy. The cool thing is that, even though we’re moving away from the original idea, you can still develop the game by adding an empty neighbourhood once the island is finished, where you can build houses, new objects and cars. You can really expand the game without any problem if the players want to.
Inksushi : This is the third project we’ve discussed, and each one is very different. Is it interesting to change the type of game? Is it interesting to change the type of game? To move away from the streamer/viewer relationship and not fall back into an ultra-dynamic concept?
Mathieu: Personally, I think this freedom is really cool. It wouldn’t be the case if I worked in a studio that was a bit specialised, for example in mobile casual, you just do that over and over again, it’s redundant. It’s true that we did Turbo Live, which was very edgy, and now we’re moving on to something cosy. I’m happy to change the atmosphere and style of the game. As long as I’m excited about the game, it’s very cool to work on it. I want to create it so that people can play it.
Simon: I know that some people like to have a clear editorial line with their studio. We’re already going to make a game that works, and then we’ll talk about it. On a more serious note, I get the impression that indie gamers don’t care. They just want to play cool games and, from time to time, they think it’s the same studio as someone else. But if you look at a studio like Supergiant, who made Hades, they really have their own style, a recipe that they improve as they go along. We’re more into experimentation, we ask ourselves what kind of game we like and we launch ourselves by taking into account previous failures. We ask ourselves if people will like it and if we can find mechanics that will hook them to make the project original, it’s not the same approach. Like Mathieu, I’m happy to have this freedom. It’s a complete change and, at the same time, it lets us try out new things.
Mathieu: The only thing we have in common is our desire to create mechanisms for streamers. That’s the thing that sticks in our productions. Here, for the moment, a viewer can control a creature. We’d like to add a ‘draw to win’ mode where you have to try and shape a word to make people guess it. It would be great fun to see a streamer trying to do something – it’s ugly, but fun, just like drawing.
Simon: Ideally, before the project comes out, we’ll do a publicity stunt with a plasticine version of Skribbl. What’s more, if people get hooked, we’ll have a big test session. All the players who love this game, when I told them about the concept, they were crazy about it.
Inksushi: This brings us to the last question in episode three. As we’ve mentioned several times, the atmosphere is very important for this style of game. How are you going to manage the music?
Simon: On this game, I’m in charge, as I was on Meaningless, because I like making music. The intention is lo-fi, which lends itself very well to this kind of game with a little orchestral edge, a bit like the lo-fi playlists in Nintendo games. I’d like a mix of fairytale vibes and chill at the same time.
Mathieu: As you can see, at the moment, while I’m struggling to develop a tool, Simon can be prototyping, managing the graphics and the music.
And it’s with these words that this episode ends. The brotherly duo have no time to slow down and are constantly looking to enhance their skills with ever more surprising concepts. We’ll be back with them in a few months’ time to see where they stand in their never-ending race to find a project that will find its audience. For those of you who can’t wait, I suggest you join their Discord (and that of Point’n Think at the same time). I hope you’ve enjoyed your reading, and I’ll see you soon in print or on podcast.