28 avril 2020
nDreams shares lessons learned from developing innovative VR shooter Phantom: Covert Ops
To gain insight into how the UK-based studio is designing one of the most anticipated VR titles on the horizon, we reached out to nDreams Game Director Lewis Brundish and Technical Director Grant Bolton. The pair reveal the painstaking steps it took to iterate on their kayaking mechanics so that it would feel realistic yet accessible to those who have never kayaked before. They elaborate on how they designed gameplay that would be highly tactile and physical and leverage VR’s strengths. The duo also talk about how they implemented a slick and minimal UI designed for VR and touch upon how they optimized the game to work on Oculus’ standalone Quest VR headset. There's certainly never been a stealth kayak shooter, let alone one in VR. Can you talk about how the inventive game originated?
Game Director Lewis Brundish: Our main goal at the start of the project was to come up with a movement system for VR that would be totally immersive and comfortable to play. The main strength of VR is the sense that you are really present in the game world, but this is always slightly compromised whenever you have to teleport to move around. We wanted to find a method of navigation that would be smooth and comfortable, allowing exploration without ever breaking that sense of immersion and tactile control. One of the early suggestions was putting the player in a boat, and after a few prototypes, we knew that we were onto something.
At what point during development did nDreams feel like they had a compelling gameplay loop?
Brundish: We were convinced by the idea of our kayak-movement system very early on during prototyping, but what we weren’t sure about was how well this would work with the player in the boat. Our earliest test map just had a single guard on a low bridge crossing a river – the player had no weapons at this point, and their only option was to wait below the bridge for the guard to look away before paddling on. Something about this simple setup worked far better than we had anticipated; the low angle that the player viewed the world from made them feel naturally sneaky, the fact that the guard and player don’t inhabit the same space made the river feel like an inventive hidden route, the wait directly beneath the guard felt tense and suspenseful …even the movements you make with your arms while paddling describe an exaggerated sneaking gesture. Obviously, there was a lot of work ahead of us at this point to flesh out and balance the gameplay loop and combat. Those systems took several months to nail down, but this early test convinced us that the combination of stealth and kayaking was going to work.
Considering militaristic kayaks are real, how did the studio balance making the experience realistic versus fun?
Brundish: Something that really helped us with this early on was listening to a diverse range of voices and opinions across the team. We have several team members who have real-life kayaking experience and were very insistent on the gameplay feeling realistic, while others had no frame of reference at all and wanted the game to control in an intuitive and arcade-style way. We considered both of these groups to be equally important and went through countless iterations of the controls and balance, trying to get it right for both crowds. For a long time, we thought that we were going to have to offer two separate control schemes, but we eventually found a way to combine the requirements of both groups into a system that we found to be accessible and intuitive for newcomers, but with the depth and nuance of a simulation beneath it. Getting this right was one of the biggest challenges we faced and has taken almost the entire length of the project.
The game takes place in the Black Sea region in 1991. Can you establish why that location and time period were chosen?
Brundish: We wanted the player to use weapons and equipment that were analog and tactile, going too modern seemed out of the question; however, we didn’t want to go so far back that some of the stealth gadgets would seem unrealistic. The early 90s felt like the perfect time period. It also provided the military equipment and tone we were looking for. The end of the Cold War was a perfect backdrop for our military espionage narrative.
Regarding the location, it was important that our environments supported the core mechanics of our gameplay. The levels would all be waterlogged, with military equipment readily accessible to someone in a boat – a naval installation made the most sense, and we incorporated the idea of it being previously abandoned to allow for more variations in theming and potential routes for the player. We researched the historical and political significance of naval bases throughout the Cold War, and the Black Sea leaped out as an evocative and appropriate setting.
Considering players can take a stealthy approach or go in with guns blazing, how did you balance gameplay so that both play-styles would be viable?
Brundish: Initially, we intended to balance the game around a more combat-heavy approach, but we found that this made the stealth gameplay too easy to bypass entirely. Making the combat too difficult, on the other hand, meant that it wasn’t a viable approach, and we wanted the player to have as many options as possible.
We settled on a balance that rewards patience and planning – combat is viable as long as it’s planned and executed cleanly. Once you have scouted an area and located all of the enemies and obstacles, you can formulate a plan of attack for the order you will take them out in or think about how you can avoid them entirely. If you go charging into an area without care, you will likely be taken out very quickly.
Can you delve into how you designed the game's various guns and refined Phantom: Covert Ops' shooting and reloading mechanics for VR?
Technical Director Grant Bolton: Lewis described an aspiration for “movie realism” where the player can act naturally but isn’t punished with overly fiddly or cumbersome detail. In Phantom, you don’t reload with a button-press; you have to reach for a new magazine from your pouch and place it in the weapon. We have balanced the detection of these actions to be generous and satisfying. Our goal is to quickly teach each player to learn how to reload so that it feels instinctive to them, as if they were a trained Phantom operative.
We also studied the way real players use each weapon and adapted the software in subtle ways to accommodate – for example, the pistol can be operated one-handed but we found many people use a second hand to steady their grip. Thus, we added a hand animation to the game so that the avatar visibly cups the pistol handle with their off-hand to mirror the player’s real-life action.
Can you explain your approach to designing the game's environments to facilitate varied tactical scenarios?
Brundish: It was very important to us that the player always feel an advantage in the kayak; we didn’t want them to feel like it would ever be a better idea to get out. As such, the environments are designed in a way that present more options than what the enemies walking on land have access to. To ramp up the challenge and variety as the game progresses, we introduce a number of new enemies, interactions, and objectives.
Environmentally, we vary from wide-open spaces to claustrophobic tunnels and everything in between. In a game where the goal is to go unnoticed, minor differences from the number of enemies in an area down to the directions they face can have huge implications for how a scenario will play out. The narrative also introduces unique beats into the gameplay, such as objectives that force the player to raise alarms, or a section where the player can no longer use their weapons.
People have praised Phantom: Covert Ops for how realistic the kayaking aspect of the game feels with players being able to use their paddle to push off walls coupled with the fact that they can lean in a direction to sharpen a turn. How much experimentation and iteration did the studio have to do to ensure this aspect of the game felt good?
Bolton: A lot of experimentation! There were many factors involved – we needed the game to feel familiar to seasoned kayakers but be accessible and comfortable for extended periods of play. During the early stages of development, we tried very “gamey” models with discreet controls – these were very reliable and comfortable but required us to teach players various abstract button presses instead of letting them act naturally. We also tried full simulations, where the player’s every movement and paddle-stroke had a realistic effect on the boat, but this was difficult to teach and could cause comfort issues. It was also much more tiring – making broad sweeping strokes without the resistance of the water takes a toll on shoulders!
The solution we arrived at is a bit like the “fly-by-wire” concept modern jet fighters use; we analyze the player’s inputs, how they’re moving their paddle and body, and translate these into the best physics forces to comfortably move the boat in the way they intended.
Unreal Engine’s visual logging system was incredibly helpful during this process, as we could record the paddle strokes of playtesters and play them back in the editor, zooming in to see exactly how they were moving.
One of VR's greatest strengths is that it can provide gamers unparalleled freedom to reach out and interact with the world. Can you delve into how the studio leveraged VR to make the game more physically immersive?
Bolton: Right from our earliest prototypes, we knew interaction with the paddle would be key. Having it ripple and splash the water was an obvious but important feature, but we ended up needing more subtle systems as well. For example, there's no solid paddle pole between the player's hands in the real world. This could lead to a disconnect with the avatar. We account for this by maintaining grip on the paddle even if it were to be wrenched away by a strong force. We also constrained movements that would intersect the avatar and the paddle.
Weapons and equipment are another important area; we tried to get the larger guns feeling heavy if held with just one hand, filtering movement and rotation slightly per object. A second hand can be used for stabilization, reducing the effect. Magazines must be slotted into the weapon but can also be thrown to create a distraction. We tried to support actions that players would naturally try, to make the game feel more natural and immersive.
Throughout Phantom's missions, there are areas where players can rip off panels, pull levers and shutters, pull out plugs, and more. To make these feel substantial and physical, we combine animation, audio, and haptic queues with kinematic tricks. We've learned that every real-world movement a player performs should have an effect in the software; this is the key to immersion. However, this doesn't necessarily have to be true 1:1 movement. We can trick the brain (to an extent) by filtering real-world movements before they appear in the game. For example, to represent pulling a rusty lever, we might translate the first few inches of pulling a lever to a much smaller movement, with an accompanying creaking sound. After breaching some threshold, we play a “crack” sound and trigger dust particles, then allow the avatar's hand to catch up with the player's real-world position. This provides the feeling of wrenching a stiff object until it gives.
This system also allows the player to pull themselves towards or away from things they can grab in the environment, which is useful for getting up close to ammo pickups, control panels, and more. When the player grips a piece of the environment, their virtual hand locks in place, and we begin to map the movement of their controller to simulated pressure in-game. This pressure is applied as a physical force to the avatar and boat, such that moving the controller away from the body pushes the kayak away from the environment, and pulling objects pulls you towards it.
Why was Unreal Engine a good fit for the game?
Bolton: Unreal is fantastic for multi-platform VR development as we were able to quickly iterate using the editor preview mode and a PC VR headset. Within seconds of making a data change, we can be playing the game in VR, and have confidence that the gameplay systems will translate well to other platforms.
The sophisticated asset pipeline and lighting tools are favored by our artists, and Unreal’s profiling and scalable render features have been essential for delivering on standalone VR headsets like the Oculus Quest.
With no traditional in-game HUD coupled with slick motion graphics that relay mission objectives to players as they float downstream, can you share how the team designed the VR title's minimal UI?
Brundish: As Grant mentioned, we felt that rather than treating the experience as a simulation, we wanted players to feel like they were inside a movie. As such, the general rule was that giving the player on-screen information was fine as long as it felt cinematic and immersive. We looked at opening film titles to inspire the style of our objective text, which appears in-world and doesn’t interrupt the flow of gameplay. When we were trying to draw attention to ammo clips in the equipment bags, we didn’t want to draw highlights around them as this didn’t feel cinematic – eventually, we settled on dramatically spotlighting the relevant ammo whenever you are holding a gun, as dramatic lighting choices are consistent with a cinematic visual language.
Another discovery we made with UI elements in VR is that they are less obtrusive if the player opts into them intentionally. Phantom allows you to tag enemies using a viewfinder (as is the case in many stealth games). Originally, we had these effects visible on the enemies whenever the player looked through the viewfinder, but we felt that the additional UI elements looked too distracting. As soon as we added the requirement to take a photo of the enemies first, this feeling disappeared – we found that more information on screen feels appropriate if the player has specifically requested it first.
What was the most challenging aspect of designing the game, and how did you overcome it?
Bolton: The large explorable environments of Phantom were the most challenging aspect to achieve on the Oculus Quest. The Quest is a very capable piece of hardware, but as a standalone device, it requires extra care and attention around performance. Our world-builders were meticulous with scene composition and encounter design to make the best use of the device, ensuring that sightlines, dynamic objects, destructible lights, and enemy patrol paths wouldn’t result in any view or scenario being too resource-intensive to render. We also developed several bespoke tools to help with optimization and profiling, as well as relying heavily on existing Unreal Engine features.
nDreams has developed many VR games. What has the studio learned about the burgeoning medium thus far that you're building upon for Phantom: Covert Ops' development?
Bolton: At nDreams, we talk a lot about the power of VR for immersion and try to place the player in a believable world that behaves as they’d expect. We try to focus on the space immediately around the player first, as this is where detail and depth perception have the greatest impact. With Phantom, we’ve placed the player in a kayak and loaded it with powerful military hardware, so they begin to explore their equipment and locomotion before they’ve even paddled out into the wider game world.
Does the studio have any game-design advice for aspiring VR developers?
Brundish: Don’t be afraid to try something new. VR is still in its infancy, and many of the tropes and genres that we are familiar with from traditional gaming don’t translate directly. I think we need to explore new solutions and ideas that wouldn’t necessarily make sense in other mediums but might make perfect sense in VR. It’s a scary proposition to develop a game that doesn’t sound familiar to people or match expectations carried over from other platforms, but I hope we’ve shown with Phantom that unique ideas can be worth following through on!