As founder and CEO of Rocket Sound, Tom Hays is an expert on game audio production and sound design. Six months after Rocket Sound became a Lionbridge Games studio, Tugdual Delisle, Managing Director of Lionbridge Games, caught up with Tom and asked him to share his insight into the changing world of game audio services and how technology and globalization have impacted the industry. By partnering with Lionbridge Games, Rocket Sound is able to create exceptional original and localized game audio for North American audiences.
Last updated: December 22, 2021 11:00PM
How did you get started in audio production and why video games?
Like so many other people in this field, I started playing guitar as a kid, and got into turning knobs on effects pedals. This eventually led to studying topics like recording engineering, synthesis, and psychoacoustics. When I started dabbling in computers and programming, I realized that there was a field where I could combine my interests. In games, we get to combine sonic creativity and craft with the art of game design, where we use sound to create and enhance mood and gameplay.
What made you decide to create your own studio?
After working as an in-house audio director for 10 years and building Technicolor's game sound services team for eight years, my partners and I felt it was time to make something new that really built upon all we'd learned about the right way to build studios, do projects, and make great games.
Do you remember the first game you worked on?
It was a Dick Tracy game back in the early ‘90s, processing voice files for the Disney Sound Source, a weird device that plugged into the serial port of a PC. The first significant game that I really dove into for months was the PC port of Myst.
What are the key ingredients that created your success at Rocket Sound?
Naturally, a huge part of it is the relationships we’ve built through the years and our track record. There’s a comfort to working with people who have been through the trenches with you on big, crazy game projects, and come out the other end with great results. When there’s as much at stake as there is with a full-scale AAA video game, it’s smart to go with people who not only are experienced, but understand the process and the experience of getting a hit game done on time. Underlying all of this is a boutique mindset, where we never compromise on quality, even if we’re not always the cheapest.
For an entrepreneur, selling is always difficult. Why did you choose Lionbridge Games?
For starters, I've known Matt Whiting for years. Before he became a VP at Lionbridge Games, he was at Microsoft during the time I was at Technicolor. We worked together on the first few Gears of War titles — Lost Odyssey and a few others — and built a strong mutual trust. Beyond that, we've worked with Lionbridge since 2015 on several large projects as a vendor. I've always been impressed with their focus on quality, and on being really smart about process. That's vital if you want creative people to focus on being creative instead of chasing assets around.
How do you plan to keep that boutique touch while working in a global scale organization like Lionbridge?
One main condition of joining was that we would work together to preserve everything that’s awesome about Rocket Sound while giving us the resources to grow and enhance our offerings. Nobody in the Rocket Sound team is interested in creating low-quality work — we all do this because it’s fun and satisfying, and we get to be part of creating amazing things. The people at Lionbridge embrace this quality and enhance it.
You have worked on many AAA games — what are some of the key trends you see there?
The most obvious one is the rise of games-as-service. There are still plenty of large game projects that go through a couple of years of development, with sometimes insane crunch in the final stages before release. But now, soon after that initial release, the team hunkers down for the first live updates, and sometimes get into a rhythm where they push these out every six weeks, or even every week. Then there’s the model of doing a much smaller game as the initial release and then adding on to it if it sells. This can keep development on a game going for years after release sometimes. For us, it means that clients will come to us with urgent sprints when they’re getting ready to release an update, needing a fast turnaround. At Lionbridge, one of the challenges we’re meeting is doing this with multi-language updates, doing hurry-up simultaneous ships within windows of a few weeks.
Some games today also go for a more natural sound than games in the past, including for voice, and some use pretty sophisticated effects to make sounds seem like they are in the environment. This pushes us to deliver a very neutral sound, where you don’t hear anything specific about the room you’re recording in. That can be harder than it sounds, but that’s what we design our studios to do.
Beyond that, the base assumption that players have for quality has risen dramatically in recent years. So many developers have done so much good work that the bar has risen considerably.
Do you see a role for technology to help here?
Absolutely. While you’re always going to need smart, well-organized people to handle these projects, we rely upon software tools to keep us sane. We’ve used Excel as a data-handling swiss army knife for years, along with some things like custom macros and Keyboard Maestro. The last thing you want is to break a game with a typo in a file name or a forgotten asset.
Lionbridge has been taking a huge leap forward in this area, with a cloud-based system that does a whole slew of connected tasks throughout production. Starting by building sessions, then helping us track recording progress, checking duration-matching where needed and other built-in safety measures, they can take what we do in English and drive it to our studios in other countries. This can save our clients and us lots of time and hassle, especially when several languages are involved.
Speaking about technology, what are the key technology-driven trends you see in the audio space for games? What are you getting ready for?
For a while now, the key term in my mind regarding in-game sound is granularity. The more powerful the processors get, and the more evolved the system get that devs put in their games, the more we can attach sound events to individual pieces in a game world, and apply real-time processing to make this sound good. For example, in a rockslide, the old-school approach would have been to build one combined sound for the beginning, middle, and end of the whole thing. A while back, we got to where we could add some sweeteners, so that if a rock rolled past on the left, you’d hear it on the left. Over time, we get to a point where the whole thing can be built from sounds attached to the individual interactions between the stones. Part of this is the evolution of physics engines, which are doing less hand-waving along the lines of “this stone is hitting this stone” and incorporating more stuff like how hard, what angle, and what surface. So we can get to a point where your in-game character gets buried in a rockslide and you get to hear all the details of the stones that are crushing you.
And then of course there’s text-to-speech. Speech synthesis is starting to get really good, and I think that over time there will be more work developing corpuses (vocabularies) for TTS systems. I’m not sure that this will mean less original voiceover right away, but it might stop growing as radically.
And there’s always the overlap between those two examples — using things like neural networks to analyze sounds and scenes and build or enhance sound effects.
You also do a lot of work for Asian customers: What are the key pieces of advice you give them when they consider bringing a game to the U.S.?
If you want your game to feel natural to a U.S. audience, you sometimes need to let go a little. It’s not that the creator’s personality should be erased, but it’s important to understand that certain things come off completely differently to a Western audience, or even between Western countries. For example, humor is crazy difficult to translate. It’s a good idea to allow for cultural review and script doctoring. It’s better for your intent to get through than for your words to be translated literally.
What are the key differences between working on an original English title compared to adapting a version for the North American market?
They can be different on many levels. When localizing, the task is usually to make versions of existing performances that sound right to Western audiences, while preserving the original vision. With an original title, the team is working to create characters and performances that work with a game world that is under construction, and often can have an influence on how that world is formed. Also, when localizing, the entire script from the original already exists, whether or not it’s all been translated and adapted when we start. With an original game, it’s not uncommon for a bunch of material from the first few sessions to get tossed out as the game goes through prototype iterations and vertical slice.
Also, with original games, we can use cool techniques other than a traditional single-actor-single mic setup. A great example would be Gears 5, where we did ensemble recordings, voice recording during motion capture, facial video in the studio, and ADR (replacing lines in existing cinematics).
