9 de setembro de 2019
MotoGP 19 - A lovingly crafted motorcycle racing experience
To see how the studio developed such an impressive game, we interviewed Producer Michele Caletti. He talks about how they balanced making a super fast and realistic hardcore racing simulator that’s also accessible to newcomers, elaborates on how the studio used reference images to strive for a photorealistic look, and discusses how the team faithfully recreated all the real-world tracks and motorcycles in the game. Thanks for your time! What do you think makes MotoGP 19 the most advanced iteration of the long-running franchise to date?
Producer Michele Caletti: Thanks for having us! There's a number of technical features that make MotoGP 19 stand out, from the Neural AI, which is a breakthrough that’s incredibly challenging and realistic, to the cool editors, which allow players to create their own helmets, racing numbers, and patches for the first time. There's also the new online system, which uses dedicated servers. 2019 is a great year for the franchise!
How did you balance making a hardcore racing simulator that is also welcoming to newcomers?
Caletti: The trick is to offer a deep simulator at the core, but be able to ease up on it for newcomers. We have a number of options that beginners can adjust, which include tire wear and joint brakes. We also have assist options that include auto braking and simplified physics. There are also numerous race options that welcome all players. In fact, we have two kinds of gamers. One crowd is super hardcore and like to max out the AI, engage in the lengthier races, and disable assist systems. The other is more casual gamers who prefer shorter races and easier AI. Both seem quite pleased with the game.
For MotoGP 19, the studio leveraged machine learning to create what you're referring to as Neural AI. Can you speak to this implementation and what it adds to the experience?
Caletti: It's a completely different approach. Instead of making the AI run on a given trajectory, with explicit rules, we've created neural agents. They are in control of the bike, just like players, and can "see" and perceive the environment: the track, the other riders, feel the tire grip, and everything else. With this, they start massive training sessions, hosted on servers, that make them grow from total beginners to fully fleshed-out MotoGP riders. Everything they do is self-learned, which is obtained when we reward them for accomplishing the best lap times and best behaviors. This creates a fair yet aggressive approach to AI.
The game seems like a love letter to MotoGP history with the inclusion of historical challenges, which include 50 legendary riders and classic race tracks. How important was it for the studio to include this aspect into MotoGP 19?
Caletti: It is important because MotoGP is a story of heritage. There's the new, shiny, contemporary MotoGP dominated by Marquez, but there are also decades of sport, challenge, heroes, and rivalries, and players love to dig deep into this vast aspect. So we not only try to provide players things to play with, but try to provide historical context so they can learn something in the process. People seem to really love it.
With 19 real-world tracks in the game, can you talk about how you faithfully recreated them?
Caletti: We have 19 tracks from 2019, but also some historical ones, namely Donington, Laguna Seca, and an older layout of Catalunya. Our recreation pipeline starts with a drone scan, and a huge set of pictures of all the details, backdrops, and buildings. From the drone scan data, we recreate a "point cloud" and then a game-usable 3D asset that we complete with all the required props. It's a complex path, but the precision achieved is top notch.
With dozens of motorcycles in the game, how did the team set about accurately modeling them?
Caletti: We have four categories for 2019: MotoGP, Moto2, Moto3, and MotoE, which is something new to both the series and the real world, and then we have the historical 500cc and MotoGP ones. For the contemporary bikes, we work with multiple sources, from detailed pictures taken at the first round in Qatar to precise part measurements. There's always some secrecy about these prototypes, so it's not possible to laser scan them, but the accuracy level is very high.
Older bikes are much harder to make: some of them are hardly available for the public now, but museums, older material from Dorna, and lots of cross checking allow for precise reconstructions even 20 years after they've been off the tracks.
For the first time in the series' history, MotoGP 19 features an extensive graphics editor, which lets users customize new helmets and outfit designs. Can you speak to this feature and why it was included?
Caletti: Players entering the career mode or online need to create a personal image, and our experience developing Ride3 tells us that they love to create, exchange, and vote for other player creations. So, for MotoGP19, we decided to enrich this area with a helmet editor. We have some 26,000 helmets created to date, which is growing at a healthy pace of 500 per day! Then there's the racing number editor, so your personal number on the bike won't just be a bland font, but complex artwork, just like what’s on the real riders. Finally, the butt patch: your nickname will stand out and be a clear message to those who follow. We have also integrated a sticker editor to make complex designs easier, and to be able to exchange logos, brands, and patterns.
With impressive visuals that feature distant mirage effects on hot days juxtaposed against realistic-looking rainy nights, how did you deliver the game's great graphics?
Caletti: We tried matching real pictures against the game, which is something that's not easy to do. Sometimes in games, reality is edulcorated, contrasts are tamed, colors are saturated, but we wanted to match what you see on TV during a race weekend. We found that a faithful approach can achieve visuals that can be stunning and realistic at the same time. The goal was to convey feelings, from the extremes of a hot summer day in Mugello to the hard weather of a stormy Sepang.
We chose picture sets, analyzed colors and temperatures, spotted those details that make every track unique, and iterated until we were satisfied with the result.
MotoGP 19 features an incredible sense of speed and makes you feel like you're actually out there on the race tracks with a palpable sense of danger. How did you manage to make the game feel so real?
Caletti: That's right, the sense of danger. Some games miss it, and try to cope with the sense of speed with tricks like excessive blur, but we went in the opposite direction. When you're on the track, time flows differently: on a straight, you're fast, but the track is wide and you feel still. Then the corner approaches, and time moves faster, and you need to catch the breaking point to [slow] the bike. Reference points flow fast at your side, and you feel the risk. Into the corner, speeds are not [overwhelming], but you’re at the edge of control. The sense of speed is not only visual, it's multi-sensorial and comes from the sense of risk, of challenge, from the sounds and from all the little cues that build up to generate what it feels like to be on a motorcycle.
Lots of people are praising MotoGP 19 for its realistic physics engine. How did you create it?
Caletti: It's the result of years of experience. We start from some real measurements; again, MotoGP bikes are super-secret, but some data is available. Then we create a generic model that takes into account all the forces and all the reactions. Gyroscopic forces play a key role and are very hard to model correctly, for example. Then we fine-tune a single class or bike, and we cross-check speeds along the track, braking points, and all that's available using onboard laps and timings while keeping in mind that we want to offer a great sensation for players.
Sometimes simulators go by the rule, "the harder, the more realistic," but that's not necessarily true. Racing bikes are conceived to be very usable and consistent up to 95 percent of their performance - then comes the challenge. That's what we want to convey: bikes that feel real, look real, and are a blast to play with.
How large was the game's development team?
Caletti: We have a core team of programmers, artists, and designers reporting to leads. These individuals are relatively few, around 20 people overall. Then, depending on the production phase, the team expands considerably, adding both developers, but also outsourcers that are used for assets like bikes. It’s worth mentioning that Milestone is also a publisher, so we have in-house professionals that work on marketing, PR, promotional content, and social media. We can top 100 people at peak capacity.
What made UE4 a good fit for the MotoGP franchise?
Caletti: Unreal Engine has clear strengths, from the straightforward multiplatform support, to the powerful tool sets available to artists and designers. It was a winning choice for MotoGP as it offers great visuals, ease of integration for third-party plugins, and offers a well-known pipeline for outsourcers among other advantages. Overall, the package is powerful, stable, and flexible, and allows us to reach great quality quickly and on budget.
Does the team have any favorite UE4 tools or features?
Caletti: Generally speaking, it would be the speed in which you see things in the editor and have them realized in the game. Being able to have a clear idea of complex assets, lighting, animations, and being one click away from seeing all of those aspects in motion together is a powerful concept that’s brilliantly executed.
For the first time in the franchise's history, MotoGP features dedicated servers. Can you speak to your efforts in bolstering the game's online functionality?
Caletti: It's been a mandatory step for us. Players wanted to have a stable, high-quality experience and only the use of servers, provided by Amazon, allowed that to become a reality. Our framework is very complex: different platforms, complex races with flexible sessions and regulations, the need for precise physics and collisions, but also the use of player-generated content, and then we have the eSport season. There's a lot of ground to cover. We had an R&D team allocated for the whole development cycle, and some MotoGP team developers working together to build the infrastructures and the game essentially at the same time. It's been quite a challenge, but it paid off in the end.
Thanks again for your time. Where can people learn more about the game?
Caletti: You can find out more at https://motogpvideogame.com/.