Making ‘Monument Valley 3’s Magnificent, Meditative Game Audio – with Todd Baker and Lucie Treacher

Indie puzzle game Monument Valley 3 (developed by Ustwo Games and published by Netflix Games for iOS and Android) transports players into an enchanted world that mixes zen-like beauty with mindbending puzzles. The player takes on the role of a young woman named Noor, who seeks to save the townspeople by keeping the Lighthouse burning bright and making the waters safe to sail.

Returning to the Monument Valley franchise is composer/audio designer Todd Baker as Audio Director. He's joined by composer/audio designer Lucie Treacher.

Here, they talk about creating a lush score with a string orchestra supported by distinctive instruments, designing captivating water sounds, and interesting texture sounds for the puzzle blocks and mechanisms, using Wwise plugins like Soundseed Grain to add dynamic layers to interactions, and much more!

---

Interview by Jennifer Walden, photos courtesy of Netflix; Todd Baker

---

Please share:

• More

Ustwo’s mobile indie puzzle game Monument Valley 3 (published by Netflix Games) is visually beautiful, sonically meditative, and filled with puzzles that require the player to rethink spatial relationships in an M.C. Escher’s “Relativity” kind of way.

Monument Valley 3 earned three nominations at the 28th Annual D.I.C.E. Awards: “Mobile Game of the Year,” “Outstanding Achievement in Original Music Composition,” and “Outstanding Achievement in Audio Design.”

Here, Audio Director/composer/audio designer Todd Baker and music and audio designer Lucie Treacher talk about collaborating as a team to craft the cinematic sound and music of Monument Valley 3, which features a broad instrumental palette that included a 17-piece string ensemble, plus distinctive and evocative instruments like a 10-string fiddle, a West African harp, bamboo flute, and Lao mouth organ. They talk about reinforcing the concept of two different energy sources (‘sacred light’ and ‘nature’s light’) using different sound palettes and treatments, designing interesting transitions for entering different levels using deconstructed sound and music from that level, and combining a physical sound with a tonal layer and adding effects like tremolo to make puzzle block movements feel reactive to the player’s action. They also talk about their technical approach to adding dynamic layers to interactions, altering the water layers’ parameters linked to velocity and rotation speed to create dynamic boat sounds, allowing the player to compose or ‘paint’ with sound when moving a rotator to create scales or melodies, and much more!

Monument Valley 3 | Official Launch Trailer | Netflix

Your beautiful music and sound design for Monument Valley 2 earned several award nominations. Did you feel you wanted to honor what you had achieved in the previous game in Monument Valley 3?

Todd Baker (TB): It’s been wonderful to see the continued love for Monument Valley 2 over the years. Players clearly connected with the way the music and sound design worked in harmony. It gives me confidence that the ‘holistic’ approach I discuss in this article resonated with people, so it was instinctive to carry that approach into MV3.

going into MV3, I was tentative about looking backward or putting pressure on myself to replicate past achievements

I’m proud of the work I did on that game, but going into MV3, I was tentative about looking backward or putting pressure on myself to replicate past achievements. It felt more exciting (and healthier) to focus on how the sound might evolve. Many of the team were new to Monument Valley and I think there was a weight of expectation working on such a successful franchise. Ultimately, I feel you are destined to make a game that reflects the creativity and talents of the people on the team, so you just have to roll with that!

Lucie Treacher (LT): I’ve so enjoyed being invited onto this project to create music and sound design with Todd. My approach has been similar in that I’ve deliberately tried not to think about MV2 in many ways, more for the reason that I felt it was anxiety-inducingly good and the bar was very high!

part of the fun of playing [Monument Valley] is the surprise of a new sound every time you twist a rotator or enter a new space

I knew Todd and the team wanted something fresh, so in a way, that made things quite easy. I just thought, “Well I may as well just bring what I can to the project!” I had the keywords ‘lo-fi’ and ‘ambient’ at the back of my mind, but other than that, I was trying to come at it with an open mind. I think part of the fun of playing MV is the surprise of a new sound every time you twist a rotator or enter a new space and it’s this sense of innovation that needed to be carried through more than anything.
 

Todd Baker and Lucie Treacher recording strings at Air-Edel for ‘Monument Valley 3’

What was new in your creative approach to the music and sound of Monument Valley 3?

TB: The obvious big addition creatively was bringing Lucie on board as a collaborator. MV2 was essentially just me in my small studio with a laptop and a handful of instruments — guitars, percussion, woodwinds, synths, etc. While I had some experience with orchestras and ensembles, my lack of classical training made me less inclined to pursue that direction.

The composition, sound design, and implementation are all part of the same creative process

Lucie brought to the project her own multi-instrumentalist palette. She also brought a more outward-looking, collaborative mindset as well as experience working with larger ensembles — all of which have hugely contributed to the sound of this game.

More broadly, I needed someone who could take ownership of sections of the game. The composition, sound design, and implementation are all part of the same creative process so if someone has the skills and vision to look at a level from all those angles, that’s when the good stuff happens! I’d known Lucie for a while and we’d collaborated a bit musically. I think there is a lot of crossover in our creative vibes, yet she brings her own energy that compliments the world of MV beautifully.

A Monument Valley 3 game audio deep-dive with Todd Baker

LT: With the game being so visually striking, it’s very easy to make it ‘sound good’ as the visuals also do a lot of work! It’s a bit of a sonic playground in that every chapter immerses you in a different diorama, and typically in MV each chapter has a distinct flavour or sound.

Lucie Treacher during the string recording session at Air-Edel

For me, I enjoyed using colours as a starting point in finding a timbre/instrumental palette or harmonic vibe. We were able to record some distinctive and evocative sounds, including the Kora (West African Harp), Bansuri, and Khaen (Lao Mouth Organ).

A fantastic thing about London is there’s a multitude of players who have access to these instruments. I also got to let loose with a microphone in a room with three different types of gamelan and we recorded a Hardanger D’amore, a 10-string fiddle performed by the fabulous Charlie Grey, who did a remote recording from St Mary’s Space on the West Coast of Scotland. And of course, we were lucky enough to record a 17-piece string ensemble at Air-Edel. The strings add quite a rich element to the score.

TB: Whilst hopefully still feeling like Monument Valley, I think the new game does sound very different. The overall sound is cleaner, more cinematic, and less lo-fi/nostalgic. We have this broader instrumental palette. It also has more edge, drama, and a darker tone in places. We’re in the process of editing the soundtrack so it’s great to hear all the music in context.
 

Behind ‘Monument Valley 3’s Game Audio: Commuters Waltz – Strings x Game

Was there anything new about your approach technically?

TB: We licensed a couple of Wwise plugins — the Convolution Reverb and Soundseed Grain. The latter was especially fun for adding dynamic layers to interactions.

Beyond that, we refined systems from the previous game. Instead of setting up audio on interactions individually, we introduced ‘presets’ for elements like draggers and rotators, which made layering sounds much easier. Many interactions have multiple layers — for instance, a physical scraping sound, a watery swoosh, and a tonal element, all controlled by separate parameters for speed, position, and angle. Having these as presets in Unity streamlined experimentation and reuse.
 

Mover Audio Presets for ‘Monument Valley 3’

Being the creators of both the music and the sounds, you could design a cohesive sound experience for Monument Valley 3. The sonic palette works so well, even for the tension moments like when parts of the buildings crumble into the water. Did you write the music and then create the sounds to fit that tone? Or, what was your creative process like?

TB: A big part of this is, as you say, just being able to work on the music, sound design, and implementation simultaneously. It opens up possibilities to have that harmony and cohesion that would be much more difficult if you split the disciplines. I’m a big fan of working like this in general; it’s trickier to coordinate on bigger projects (and hard to find people who genuinely have the combined skill set) but smaller projects like MV are a perfect opportunity.

It opens up possibilities to have that harmony and cohesion that would be much more difficult if you split the disciplines.

There’s not really a fixed creative process which is part of the fun. Some levels might start with a piece of music that we then deconstruct into something interactive (i.e., stems that are controlled/affected by the player’s interactions). Some might start by adding in some ambience and working on the interactions first to make them tactile and interesting, then we realize little or no music is needed. It keeps it fresh to try to think of different ways to approach a level.

LT: The way in which the focus moves between ambience, sound design, or music I think in turn gives pockets of breathing space in the game for the player. When playing, you can focus on the intricate sound elements of a rotator or simply sit and enjoy the orchestral-like cacophony of birds on a marsh at dusk, as captured in a field recording.
 

Each successive Monument Valley game has leaned more into narrative and storytelling. How did you go about supporting the story of Monument Valley 3?

TB: Whether it’s an abstract tonal journey or something more defined/character-driven, storytelling is the main motivator for me with any project. It’s what makes things meaningful. MV3 introduces a new character journey, but there are also bigger concepts involving changes to the world around you.

The game has the idea of two different energy sources: the ‘sacred light’ that is familiar (but ultimately unsustainable) followed by the discovery of ‘nature’s light’ that is a sustainable alternative. The ‘mandala’ moments at the end of each chapter are effectively the player ‘harvesting’ these different types of energy so we wanted to cement these concepts with different sound treatments.

‘Monument Valley 3’ Game Audio: Sacred Light Mandala

The ‘sacred light’ sounds more synthetic, glassy, and angular, whilst the ‘nature’s light’ uses foliage and other natural sounds as source material. Lead programmer Manesh Mistry and I worked together closely at the end of the project to try to make these as visually and sonically harmonious as possible.

‘Monument Valley 3’ Game Audio: Nature’s Light Mandela

LT: Additionally, core sound elements such as the ‘ping’ you hear when you click to move the character (a consistent element throughout MV games) help to support the narrative and subtly mirror what’s happening at different points. Sometimes this ping sound is subverted, the absence of it reflects the destruction in the world or is used to tell us about a new environment, echoey for instance underwater, or highly textural when navigating through a field of wheat.
 
 

‘Monument Valley 3’ Game Audio: Chapter Select Flow

What went into your sound and music for the interior lighthouse space?

TB: It’s nice that you mention this area. It’s one of my favourite bits of the game and was fun to work on! The main creative cue here was the stained glass aesthetic and the way the lighthouse interior makes the player feel small within an expansive space. I can’t really separate that from the feeling of being in churches and cathedrals here in the UK, which is synonymous with big reverberating spaces. It was fun to make it feel huge. The unlock animations have these big pitched-down mechanical creaks with loads of reverb.

The main creative cue here was the stained glass aesthetic and the way the lighthouse interior makes the player feel small within an expansive space.

The ambiences you hear as you face each ‘window’ are edited from sounds contained within those levels, and likewise, the transitions as you zoom in to the level are deconstructed bits of music and sound from the world you are entering — reversed/manipulated/chopped up, etc. The idea was to echo and subtly foreshadow the character of the world you are about to enter.
 
 

‘Monument Valley 3’ Game Audio: Watery Sailing & Interactions

I love the water sounds in the game! You get to steer your little boat through calm water and rough water. Moving pieces of the puzzles makes the water slosh around. It laps against the buildings. Can you talk about creating those water sounds? Also, since it’s a mobile game, were there challenges in making those water sounds match/react to the player’s actions?

LT: Yes, with this strong presence of water in the game it was important to find different ways to present water sonically so it didn’t get tiring on the ear. Sometimes this is as gurgle-y echoes as in Chapter 6, sometimes as gentle ebbing as in the calming marshes of Chapter 3, with complimentary animal calls. In these instances, the environmental sound can create an illusion of a richer environment and a bigger world (beyond the mobile screen’s horizons!). There are also moments when the sea is raging but there’s almost an absence of water (low-passed) and this is also quite a striking effect.

we have different layers of water playing simultaneously as you move the boat, each with different attack and release times

With Boatem, our lovely Totem-transformed boat character, we have different layers of water playing simultaneously as you move the boat, each with different attack and release times, making it feel quite natural. Boatem’s sound is made up of around four different water layers and these are revealed when sailing at different speeds. There are parameters affecting these layers linked to the velocity and rotation speed of the boat. For instance, the pitch might increase with the speed of the boat, and pan as the boat turns. There’s also a bunch of randomised one-shot squeaks and creaks triggered as it’s moving to help make it feel organic.

TB: I’ve built up a decent collection of watery recordings over the years, but gathered some nice new ones for MV3. Most of the lapping and ocean-like waves you hear in my levels are from recordings in Scotland (including Loch Ness, and Cromarty) and some nice beaches on the east coast of Ibiza!
 

How did you create the sounds for the interactive puzzle pieces? Some sound almost wooden, like Lincoln Logs, and others sound like stone, and plants, and some are more abstract and difficult to describe like in the image above…

LT: Nearly all the interactions have a physical layer, often built from some kind of rocky loop for instance. We have a few variations of these loops that alternate so the sound doesn’t get repetitive when moving it a lot. Then we add parameters that affect the sound as you move it, for instance, a high pass filter parameter that activates as you move a ‘dragger’ up. Additionally, we might add an impact at either end of the direction of a dragger, so when you make contact with another block it would trigger a small impact (which can scale with velocity).

Nearly all the interactions have a physical layer, often built from some kind of rocky loop

Then we often pair this physical sound with a tonal layer. A fun effect we used a lot was tremolo on tonal pads, for instance, linking the speed of the interaction to the rate of the tremolo to create a reactive stuttering effect that speeds up or slows down depending on how fast you’re moving a block.

by moving a rotator you can create scales or melodies, and this creates this illusion of almost composing or ‘painting’ sound with the swipe of a finger!

In the case of the example in the picture, I paired a low rumble physical layer with an airy white noise layer which used this tremolo effect, as well as warped vocal layers. Additionally, I used Soundseed Grain in Wwise to ‘scrub’ at random through a tonal sample as you move the dragger. Oh, and a lot of reverb!

In some special cases (e.g., where an interaction might be more of a statement narratively), we might add switches at different steps over a range: meaning that one-shot notes will play out at different angles on a rotator, for instance. This means by moving a rotator you can create scales or melodies, and this creates this illusion of almost composing or ‘painting’ sound with the swipe of a finger! In these cases, the rotators become more like musical instruments.
 

What went into the foley sounds for the player character named Noor? The footsteps are so cute! How about the sounds for the NPCs?

TB: There’s not really anything special going on here! Most of the foley and footsteps are edited from various recordings and libraries I have. We can set surface types, but use them sparingly.

The main thing is making sure you have lots of sample variation in the default surfaces, and just tweaking the mix and reverb sends so they feel right in terms of scale. We use screen space to pan the footsteps/foley rather than actual 3D positioning. Also, the footsteps aren’t actually synced to the animation; they are just randomised one-shots looping with a trigger rate in Wwise.

The main thing is making sure you have lots of sample variation in the default surfaces, and just tweaking the mix and reverb sends so they feel right in terms of scale.

LT: In terms of the NPCs, we did initially do some experiments with creating ‘vocalizations’ for the NPCs, trying sounds like whistles and blips and bloops as if they were speaking to each other. This was an interesting conversation. MV characters have previously always been ‘silent’ but it felt like their silence spoke louder than words in this instance! It gives them this enigmatic, cryptic quality and a consistency with the other games.

Instead, we use things like giving them bespoke speech ‘bubble’ sounds to give them their sonic identity as well as their own musical instruments (playing diegetically in the scene) which is a super nice touch in this particular game.
 

Did you have a favorite cue for the game or a favorite chapter/level for music?

TB: I’m a fan of the papery trip that is Chapter 3: “Origata Atelier” and the lovely playful interactivity that Lucie built into the music for that level. It’s a great example of visual art, music, and sound working together to feel cohesive. Those lovely warm, glitchy piano chords were from one of the first sketches Lucie did for the game and I think it captures that fresh, initial inspiration.

Those lovely warm, glitchy piano chords were from one of the first sketches Lucie did for the game

LT: This changes all the time, depending on how immersed I’ve been in a particular chapter, but at the moment I’m digging some of the post-launch content we’ve been working on (coming soon!), which evokes some of the feeling of MV2. Todd’s got a beautiful string cue in there and has conjured some both dreamy and slightly dark soundscapes over which I’ve been creating the interactive sound design. The palette is essentially ‘hollow wood’ so think guitar harmonics, cracking twigs, pizzicato strings, etc. This game is honestly all about ASMR!
 

What are some challenges, limitations, or considerations that factor into creating and implementing sound for a mobile game?

TB: The great thing is that the limitations are very few these days because phones and tablets have so much power and memory. In MV2, we split the soundbanks and loaded them per level, but for MV3, we just load everything when the game boots. We’re running all the sound channels, DSP, etc. that we need and don’t feel restricted.

Some lower-spec Android phones might be difficult to account for but the team has a target low-spec device for both Android and iOS, and we just work to that.

the 30-40Hz sub-frequency range is much more ‘available’ creatively now so there’s quite a lot of 808 weight to many sounds in this game

One thing that has changed since MV2 is the proliferation of really good quality headphones and earbuds. We hope for the best case that people listen on headphones, and the boom in this market means that many people will be getting a pretty hi-fi experience as a result. I feel like the 30-40Hz sub-frequency range is much more ‘available’ creatively now so there’s quite a lot of 808 weight to many sounds in this game that wouldn’t have been ‘worth it’ back on MV2!
 

What have you learned while creating the music and sound for Monument Valley 3? What lesson(s) will you carry forward in your work on future games?

TB: Lucie and I have said countless times how much we’ve learned from each other on this project. It has been such a nice collaboration and I think we’ve inspired and pushed each other in different ways. For me personally, this game solidifies how much I like to feel part of a wider team. Even with all the ups and downs of development, it’s always just very meaningful to feel part of a collective effort, and I’m grateful for that. Seven years ago MV2 changed the course of my life and career in many ways, so to revisit that world and pull something together that we are all proud of means a lot to me.

sound design is most certainly a kind of music-making that must be approached with the same rigour, sensitivity, and musicality as creating a score.

LT: MV3 is the first proper game I’ve worked on, so really this has been a baptism of fire in so many ways, both technically with the implementation and with the music production with which Todd has been so generous in sharing his skills, and for which I’ll always be grateful! I think the game has reinforced my feelings that sound must sing, and that enabling players to interact with sound in such a way creates another layer of puzzling for the players. I’ve also felt in this process that sound design is most certainly a kind of music-making that must be approached with the same rigour, sensitivity, and musicality as creating a score.
 

Footnote from Todd: To finish – a quick confession from me. Shamefully, I ended my last article for ASoundEffect by asking people to guess in the comments what percentage of players were listening to Monument Valley 2 with headphones, promising to reveal the analytics at a later date, but then never following up!

The data (specifically for MV2) showed around 10% of players listening with headphones. Apparently, this is more like 2% for the wider mobile gaming market, so hopefully the ‘best with headphones’ messaging that Ustwo adds to their games is helping somewhat. The figures might sound a bit depressing for audio folk, but when a game is played by tens of millions that’s still a lot of people getting the full experience. In my view, the only way we’ll get those numbers up is by continuing to actually remind people to grab their headphones when they play, as well as just striving to make great-sounding games that people want to hear.
 

A big thanks to Todd Baker and Lucie Treacher for giving us a behind-the-scenes look at the sound of Monument Valley 3 and to Jennifer Walden for the interview!

Please share this: