How Panic Button became masters of the Switch port

The Austin studio has a knack for optimizing blockbuster games.

The Switch is a remarkable little machine. Part handheld and part home console, the device is home to some wonderful Nintendo-made exclusives such as The Legend of Zelda: Breath of the Wild and Super Mario Odyssey. It's also home to some rock-solid ports, including Rocket League and the Nazi-slaying Wolfenstein II: The New Colossus. These versions are visibly inferior to their PS4, Xbox One and PC counterparts, with murky textures and adaptive resolutions that frequently hover below 720p. Still, they are undeniably impressive, offering the same ferocious gunplay and bombastic humor whether you're playing at home or the back of a bus.

The wizards behind these seemingly-impossible ports? Panic Button, a developer in Austin, Texas with less than 50 employees. It's an unusual specialty -- some companies, such as Bluepoint Games, have earned a reputation on their ability to remake and remaster classic titles such as Uncharted: The Nathan Drake Collection and Shadow of the Colossus. Few, though, are able to match Panic Button's skill at bringing so-called 'triple-A' games to effectively tablet-grade hardware.

The studio wasn't created for this purpose. The four founders -- Craig Galley, D. Michael Traub, Russell Byrd and Aaron Smischney -- had worked together at Acclaim Entertainment, a now-defunct publisher that produced Turok, NBA Jam, and the first official South Park game, and Inevitable Entertainment, a once-independent studio that made a shooter called Tribes: Aerial Assault and a Zelda-like based on The Hobbit. Inevitable was acquired by Midway Games in 2004 and, eventually, swept up in the publisher's push to standardize some of its internal tools and technology.

"And while I'm all in favor of that, there's a right way to do it and a wrong way to do it, and I think they were making some bad decisions," Traub told Engadget. "And Craig [Galley], who was at that time still my boss, also was of a similar mind. So we started seeing some of the writing on the wall, as it were, and we got out while we thought the going was good."

The quartet created Panic Button in 2007, roughly a year before Inevitable Entertainment, now called Midway Studios Austin, closed its doors. The founders didn't have a strong vision for the company. They enjoyed making games, though, and wanted to avoid the "corporate mistakes," according to Traub, that ultimately brought down both Acclaim and Inevitable.

At first, the studio was "very tech heavy" and had little art and game design talent amongst it ranks. Traub, for instance, was a decent programmer and entered the game industry through Acclaim's T&T (tools and technology) division. One of his early projects was an N64 engine that could run games with, at the time, realistic 640x480 "Hi Rez" graphics. Galley had been Traub's boss at Inevitable but was also a software engineer. "That theme, for good or ill, followed us for quite a while, and even influences the company [Panic Button] to this day," Traub explained. "In that we are tech heavy."

When Panic Button formed, the Nintendo Wii was huge. Roughly 12 months into its lifecycle, Nintendo was still struggling to keep up with demand. Families that had bought the console for its approachable pack-in game, Wii Sports, were now hungry for similarly sociable experiences. Panic Button pounced on the console's popularity with two mini-game collections called Go Play Lumberjacks and We Wish You a Merry Christmas. They were hardly original but, crucially, the team was able to nail the all-important Wiimote-waggling. "You could make a relatively small game," Traub explained, "without a gigantic budget, and get it onto the Wii and actually have a shot. So that's what we did."

"You could make a relatively small game without a gigantic budget, and get it onto the Wii and actually have a shot."

The following year, Microsoft released its motion-sensing Kinect peripheral for the Xbox 360. Panic Button's experience with motion controls landed them two projects: Hulk Hogan's Main Event, a critically panned wrestling game, and a decent space combat mode in the ultimately divisive Kinect: Star Wars (yes, the title that also offered a 'Galactic Dance Off' mode.) The company wisely pivoted toward traditional port work, bringing titles such as Injustice: Gods Among Us to the PlayStation Vita and Octodad: Dadliest Catch to Wii U. It also helped Psyonix bring Rocket League to Xbox One and, later, optimize the company's PS4 version.

These projects, and others, exposed the company to different game engines and development tools, as well as target hardware. "It gave us an amazing breadth of experience, and that continues to this day," Traub said.

And then came the Switch. Panic Button got, for a small developer, "pretty early" access to Nintendo's new console, according to Traub, and quickly realized that the hardware had potential. Every company was hesitant, though, after the Wii U's commercial flop. "At the time, when we were looking at the Nintendo Switch it was, 'Well, this could either really go well or it could really go bad,'" Traub said. "It was difficult to tell, but we saw the potential in it. We saw that it had a little bit more muscle concealed behind its small exterior than most people would suspect."

The Switch holds a surprising amount of power, but it's still drastically different from the PlayStation 4 and Xbox One. Unsurprisingly, some studios -- especially those pushing for bleeding-edge graphics and silky-smooth performance -- didn't want to dedicate resources to an in-house port. "That created a business opportunity for us," Traub said. Once again, the company started working with Psyonix on a Rocket League port. With a little bit of "finessing and romancing," according to Traub, it also persuaded Bethesda to hand over Doom, an iconic shooter that was rebooted to universal acclaim in 2016.

"To find ourselves working in that space, on that title, on that IP, was pretty impressive, and pretty intense," he said. "But we worked very hard not to let ourselves get psyched out by that. It was, 'Keep your eye on the ball, work on the tech.' Every time we could save a millisecond, every bit we could push, pushed us a little closer to, I don't want to say glory, but pushed us closer to our objective of making our own contribution to the franchise."

How to make a port

Panic Button uses an eight-stage process to complete most of its port work. The first step, called the "evaluation phase," starts with a quick check to see if anyone inside the studio is familiar with the game. "If I don't play that particular game, and I'm just picking it up and noodling, I'll get some insights but I may not pick up on some of the subtleties," Traub said. "Checking around the studio and seeing who loves that game, and chatting with them and getting some insights is a very good step."

Then, the company will typically get an "evaluation drop," including the source code of the game, from the original developer, publisher or rights holder. Some senior Panic Button engineers will then go "spelunking through that code," according to Traub, "to see what makes it tick." At this stage, the company is trying to get a general sense of the title and whether a Switch port is feasible. The team wants to know the resource-intensive hot spots and what technology, including middleware, was used to build the original game. "It's the equivalent of doing a few x-rays to see what's in there," Traub said.

At the end of the evaluation phase, the team will present its findings and suggest some performance targets. Everyone wants their favorite game to run at 1080p resolution and 60 frames per second, but that's often unrealistic for the Switch. Panic Button will then say whether they think the project is worth pursuing at all. "There are times when we've looked at titles and went, 'We just don't think that's viable without a lot of compromises,'" Traub said. "And neither us nor our partner is looking to make a long laundry list of compromises. So if we determine that it's not going to be a win-win, then we can knock off [the port] right there."

Bethesda, obviously, said yes.

The company then enters a deeper "technical due diligence" phase. Every game is different, but this stage typically lasts three months. "So this is x-rays, CAT scans, all of those metaphors," Traub said. The team will try to compile the game's code on the new platform -- in this case, the Switch -- by stripping away almost every conceivable element, including the graphics, audio and networking tools. "It's an exercise in thousands of compile errors, followed by thousands of linker errors, followed by 'it's running but the screen is black, why is the screen black?' Just tons of investigating," Traub explained.

Panic Button will then provide another report. Sometimes, the due diligence will expose something that didn't come up in the shorter and shallower evaluation phase. Panic Button will need to explain what it's found and how that could affect both its projected timeline and the compromises required to port the game. The two parties will then reflect and discuss whether they're still happy to pursue the project. "If it's not to everyone's liking, then no harm no foul," Traub said. "[We] shake hands and [say] see you next time." So far, no one has ever used this escape hatch. Still, it's a vital step that gives both companies confidence.

"It's very comforting to both sides, because we're coming in with a lot of unknowns," Traub explained. "so giving us these three months to really get in there and mosh around a bit allows both sides to have a much higher degree of comfort and assurance that the project will actually be viable."

Panic Button then enters a phase called "first light." As the name suggests, the aim is to light up at least one pixel. It doesn't matter what the color is, or if you can tell what it's representing in game. The team just wants a sign of life. "When you run it the first time it's a black screen," Traub said. "It's always a black screen, and it's a black screen for days, possibly weeks."

"When you run it the first time it's a black screen ... It's always a black screen, and it's a black screen for days, possibly weeks."

But the team is persistent and, somehow, always finds a breakthrough. "We're trying to get it so even garbled, horrifying, 'I have no idea what the hell is on the screen' pixels start lighting up, and then suss through that," Traub added. "Well, we went from a black screen to purple screen, why is it purple? I don't know, and what is that oblong thing in the left? Just trying to get it to the point where it may be ugly, but at least we recognize that's the game running."

Once the port is visually recognizable, the company barrels into another self-explanatory phase called "first playable." The team will start re-connecting the controls and enough subsystems so the player can walk around the world. You might not be able to fire any weapons, though, or kill anything dangerous. "But at least you're moving around in the world, and it's starting to be more of an interactive experience," Traub said.

By this point, the team is intimately familiar with the game and how it was made. They get to see every trick and workaround that is, ideally, imperceptible to the player. "It's not a spaceship, it's a Hollywood set of a spaceship," Traub explained. "And if you turn the camera 15 degrees to the right, you see a grip standing there with a microphone." For Traub, it's like watching a 'making of' documentary for your favorite movie. You can appreciate the time and craft that went into its development, but you'll never truly believe in its world again.

Sometimes, Panic Button will bring in quality assurance (QA) testers at the first playable stage. The game will be in rough shape, but the team wants to be sure that what has been implemented is working correctly.

Next, the company aims for feature parity. That means implementing and polishing every system that is necessary for the final release. The difficulty of each addition, and the order they're tackled in, differs depending on the project. Rocket League is a predominantly online game, so that functionality might have been required for "first playable." For a title like Doom, however, which has a single-player campaign and a fun but secondary multiplayer mode, it could be tackled later. Panic Button will often use this time to improve the visuals, adding shaders, lighting effects and complex particle systems. The team will also tackle the audio and figure out when, and how to load different parts of the game.

"Some games just try and do a static load upfront," Traub explained, "and other games are doing streaming as you progress. The different bandwidth challenges may complicate that, so we might have to do levels of compression to make it load faster, or we might have to be more aggressive about background processing." Video games use lots of tricks, including elevators and slow, narrow passages that force a cinematic shimmy, to hide sizeable load times. Panic Button has to match the timing of those sequences or add minimal obstructions, such as a jammed elevator door, to keep its load times from standing out. "We definitely have to watch for things like that," Traub said. "It's always entertaining, all the tricks and airlocks that games put in."

Once the port is effectively complete, Panic Button will go through pre-submission and submission with Nintendo. Finally, once the game has been released, the team shifts into a live support mode, reacting to player feedback and releasing any optimizations that weren't quite ready for release day. "As much as everybody tries to QA things internally ahead of time, the reality is modern games are obscenely complex animals, and once you throw a million players at something, they're going to find even the tiniest crack," Traub said. "So paying attention to what's being reported out there, if anybody's having problems, or if there's chatter about problems, then that's something that we can investigate."

Doom was released for the Nintendo Switch on November 10th, 2018; Rocket League came out four days later. Most critics praised Panic Button's technological achievements, while noting that the definitive version, with fewer technological compromises, would always lie on other systems.

The team then moved on to Wolfenstein II: The New Colossus, another first-person shooter published by Bethesda. The experience with Doom "gave us a little bit of extra boost out of the gate," Traub said, but ultimately the game had its own challenges and systems to contend with. There was also one instance where the developer just had to insert a new building into the level to hide objects that would have required Herculean rendering. "There was one shot, one camera angle, that was just killing us," Traub admitted. "So we just simply put a building in there that was a big occluder, and that saved the day."

"There was one shot, one camera angle, that was just killing us. So we just simply put a building in there that was a big occluder, and that kind of saved the day."

Again, the port was considered "a technical marvel" and "something to behold." Panic Button's reputation then skyrocketed. It started receiving more inquiries and RFP (request for proposals) from other studios and rights holders. That meant the company could be more selective and pick up projects that were both interesting and challenging for its employees. "What we're not doing is picking the ones that make us go, 'That's easy, we know we can knock that out of the park.' We're actually picking the ones that make us go, "All right, that's hard, but we're pretty sure we can do it,'" Traub explained.

Panic Button has since ported Warframe, a free-to-play online shooter, and Hob, a colorful adventure game by Runic Games, to the Switch. It's also ported Subnautica to consoles and helped Electric Hat Games release To The Top, a first-person parkour platformer, on PlayStation VR. The company has kept busy, though its fans are understandably excited for two upcoming projects in particular: Wolfenstein: Youngblood and Doom Eternal. These two games are still in development and considered by Panic Button as co-development ("co-dev") projects, rather than a typical port.

The company's eight-stage process remains the same. Working in a co-dev model does provide some unique advantages and disadvantages, though. If Panic Button has a question, for instance, it's easier to reach out and talk to the person who worked on that particular level or system. "I say potentially, because some projects cross the finish line gracefully, and other projects cross the finish line on fire," Traub says with a chuckle.

With a typical port, that person might have moved onto another project or left the company entirely. "It depends on the amount of time that's passed," Traub said. "We've had some ports where we're working on something that shipped years ago, and there's a lot that's been lost to time and legend." A co-development model also means that the game is still being worked on. It's a moving target, and if the other developer makes a sweeping change it can impact Panic Button's work too. "Every project is different, and every group is different, so every one is an adventure," Traub added.

Wolfenstein: Youngblood, developed by MachineGames and Arkane Studios, has two protagonists and will offer a co-operative multiplayer mode, unlike its predecessor. Doom Eternal, meanwhile, will have new locations, weapons and traversal mechanics -- including a quick-dash and deadly grappling hook -- that increase the tempo and blood-pumping action. "They are easily among the most challenging projects that we've worked on," Traub said.

Games of this size and visual fidelity are normally considered beyond the Switch's modest hardware. Panic Button has demonstrated, though, that it's possible to bring 'triple-A' games to the portable system. The company has already set a high bar -- the question now, it seems, is how much farther it can be raised before Nintendo gives them a hand and releases a Switch successor.