A 2D Game with 3D sound: Ori and the Will of Wisps

by Kristoffer Larson

Formosa Group was asked by Moon Studios to provide end-to-end audio services for Ori and the Will of the Wisps, a much-anticipated sequel to their award-winning Ori and the Blind Forest. We wanted to upgrade the sonic world of Ori to a truly next-generation level by using the latest technologies to provide the most immersive audio experience possible.

One of our core goals for this project was to update the Ori experience to the state-of-the-art capabilities of the Xbox One and PCs while still providing an excellent mix for all audio channel configurations, especially stereo and headphones. By working with the Dolby Atmos and Microsoft Spatial Audio formats we were able to create and mix natively in the immersive audio realm as well as rely on the great sounding downmixing algorithms to deliver equally compelling mixes for 7.1, 5.1, and stereo.

But why opt to spend time and resources on a “3D” audio presentation for what is essentially a “2D” game?

First off, Ori and the Will of the Wisps may have most of the gameplay action in 2D, however it still has a lot of depth within the world. It has so much detail and dimensionality in its environments that alludes to a “there” beyond the playable area. We wanted the audio to add to that sense of space and immersion.

Secondly, there is a large amount of verticality used in the level design and gameplay, so being able to spatialize audio in the z axis really helps to inform the player when there is more to explore “above” them.

Thirdly, the general benefit of having a much wider soundstage makes for a really open and natural sounding mix. There are more panning and positioning options to place elements so they don’t conflict as much, so I don’t need to compress or EQ as drastically just to make things fit.

The next question invariably is — How many people have big Dolby Atmos home theaters anyway?

With spatial audio virtualization supported on most platforms, anyone with a set of headphones has the opportunity to hear a great surround presentation. This was naturally the case with Ori and the Will of the Wisps being developed for Xbox One and PC. Plus, the audio engine was Audiokinetic’s Wwise, so the game was essentially ready to output spatial audio without any additional engineering support.

The combination of the target platforms and the development tools made it essentially frictionless to be able to jump right in to design and mix in Dolby Atmos from the very beginning.

Kristoffer Larson

Matching the visual fidelity

While one of our goals was to create and mix Ori and the Will of the Wisps in a spatial sound presentation, it wasn’t an arbitrary decision to tick off a feature for back-of-the-box marketing. We knew that the extra dimensionality and immersion was exactly the sort of treatment that would match the visual fidelity of this beautiful game. So instead of flying objects around just for the sake of doing so, we worked on finding the appropriate circumstances and level of intensity for spatial audio to enhance and support what was happening in the game visually and thematically.

Ambience with more depth and clarity

Environment sounds are a natural subject for surround or spatial content. Instead of creating wide multichannel beds, we opted for dynamically panned mono and stereo sources, coupled with runtime reverb and delay. This approach not only saved on streaming bandwidth but provided a way to have almost infinite variation. Each element was given its own set of automation paths, dialing in the amount of spread, movement, and randomness that was appropriate for that particular piece of content.

Panning automation in Wwise.

Content was placed logically according to context.

High placement: bird-like vocalizations and wind-in-trees types of sounds
Middle placement: general wind and ambient motion
Low placement: ground-based sounds like grass and foliage rustle

Like in a traditional surround mix, we biased more towards the front of the room in order to maintain the focus on the screen and the game you’re playing, not distracting the player with too much extraneous action to the sides or behind them.

Spot and game object sounds

The next most obvious elements would be sound emitters coming from a specific location in the game world. The props and characters created their own panning based on level design and AI movement.

Vertical gameplay

Level design abounds are where the verticality of the world was enhanced by being able to hear the sounds coming from above you, sometimes as a clue of where to explore next.

Another obvious choice for appropriate usage of spatial audio were the in-game cutscenes and vignettes. We implemented them with a combination of in-engine sources, mostly ambience and characters, and 12 channel stems in a 7.1.4 configuration. Like any game, this combination of in-engine and stems lends itself to a seamless transition in and out of cinematics. The decision of what would be done offline was based on the complexity of the content, if it was going to be a custom one-off, or sometimes just the production schedule. The stems were created in Nuendo and exported as multi-mono files, which were in turn interleaved using the Wwise Multi-Channel Creator.

Monitoring

Naturally throughout the entire project we needed to be able to monitor the audio to ensure that our work was translating well in all of the common formats and devices. Between myself and my two audio leads, we were constantly going back and forth between 7.1.4 Dolby Atmos speakers, stereo speakers, and headphones (with and without spatial virtualization). Additionally, during the mixing stage of development, I ran compatibility testing in 7.1, 5.1, and stereo in various professional and consumer environments.

Formosa Group immersive editorial suites

Overall, the translation of the mix into different configurations was excellent and we did not need to make many changes or compromises. The only thing that did come up a few times was related to panning spread when comparing speakers or virtualized spatial renders to regular non-spatialized stereo headphones. Because of natural crosstalk in acoustic spaces, whether virtual or physical, our left/right panning sounded a little too extreme when monitored in headphones without virtualization. We ended up finding a nice middle-ground for our panning that still provided good directionality on speakers or virtualized, but without sounding too hard-panned over stereo headphones.

Mix goals and tools

As with most game projects, while there was a mix phase near the end of the development cycle, the bulk of the mix work started early in the development process and continued every step of the way as the game came together. As systems came online or content was implemented, we were always mindful of how it impacted the soundscape and the relative mix. Early in the process we established some loudness guidelines for classes of content such as music, character sounds, environmental audio, cinematics, etc. Using these as a rough skeleton of the mix, we created busses and structures within Wwise to dynamically adjust the mix according to what was happening in the game, both sonically as well as design or gamestate driven.

Busses and structures within Wwise.

I wanted the game to sound very dynamic, with the big stuff hitting hard and the environment to be enveloping but not intrusive. It all needed to work with the amazing music being a central component of the mix. Because we were using every tool and trick for populating content into the spatial landscape, the mix phase was spent almost exclusively tweaking and manipulating 3D attenuations, emitter locations on game objects, and bus dynamics. The goal was to find a comfortable and natural space for each element with as little need for brute-force compression and EQ just to make things fit.

Mixing for spatial platforms allowed me to use so much more of the room to work around the music without it sounding oddly placed or unnatural.

Kristoffer Larson

Because the music was composed and mixed in stereo, it needed to be up front in the left/right mains. Mixing for spatial platforms allowed me to use so much more of the room to work around the music without it sounding oddly placed or unnatural. Just a little bit of positional spread helps items be heard clearly. Remarkably, the downmix whether via Dolby Atmos renderer or Wwise’s internal routing still sounded great and felt like they made the right “decisions” to preserve the aesthetic intent of what I was trying to do.

Final thoughts

I’m really happy that we were able to create such a rich and immersive audio presentation for Ori and the Will of the Wisps, we’re really proud of how everything came together in the end. The combination of the target platforms and the development tools made it essentially frictionless to be able to jump right in to design and mix in Dolby Atmos from the very beginning.

Award-winning sound veteran Kristoffer Larson joined Formosa Group in the summer of 2016. With over two decades of sound experience, including post, games, and VR/AR, Larson truly understands the needs of developers. He’s happiest when a project is collaborative across all levels, from concept to integration. Recent projects with Formosa include Ori and the Will of the Wisps, Marvel Strike Force, State of Decay 2, and Mission: ISS.

Prior to joining Formosa, Larson founded his own company, Tension Studios, based in Seattle, Washington. After having spent years in various internal audio roles with companies like Cranky Pants Games, Konami, WB Games, and Dolby Labs, his goal was to create a game-centric audio studio that specifically catered to interactive development. His unique approach and experience can be heard on Tiny Bubbles, Fable: The Journey, State of Decay, and Halo 4.

Get in touch with our Developer Relations team.

Have questions about how to make your game in Dolby?

Our Developer Relations team is here to help!

Text

Selection