How a World Is Born: Sound Design

Sound is not added to a world to complement the visuals. Sound is co-primary. In an immersive experience, what you hear determines whether what you see is believed. A desert that looks extraordinary but sounds like silence – or worse, like approximate silence – will never feel like a desert. The brain cross-references constantly between sensory channels. When the channels align, presence happens. When they don't, the experience collapses.

This is the guiding principle of Anydream's approach to sound design.

The Sound Brief: Starting from the Script

Sound design at Anydream does not begin when the environment build is finished. It begins – like every other production discipline.

The script includes a sound reference section for every scene. Before a single audio asset is selected or recorded, the sound designer and composer receive a document that specifies:

• The primary sound environment of each zone – what kind of acoustic space it is (open-air, enclosed, semi-enclosed, transitional)
• The emotional sound target – what state the audio should produce in the listener, described in the same language as the visual emotional target
• Time-of-day audio conditions – how the sound environment changes between dawn, midday, and night
• Cultural and geographic specificity – what kinds of sound are authentic to this world's setting and period
• Reference tracks – existing recordings, musical pieces, or sound environments that approximate the target

For Dune Desert, the sound brief described five distinct sonic environments across the world's zones:

1. Open desert – near-silence with wind at variable speeds, distant geological sound, the absence of biological noise as the primary character of the space
2. Canyon – harder acoustic surfaces producing subtle reverberation, the specific echo quality of a confined stone space, occasional distant wildlife
3. City exterior – the ambient wash of a living settlement heard from outside its walls: a low frequency hum of many simultaneous voices, percussion sounds from craft work, animal sounds
4. Bazaar – the most acoustically complex zone: overlapping voices in multiple languages, market-specific sounds (cloth, metal, ceramic, food preparation), wind moving through cloth canopies, the specific acoustic profile of a narrow stone street
5. Oasis – water as the primary sonic element, bird calls specific to the region, palm leaves in low wind, the acoustic softness of a space with significant vegetation

Each of these environments required a different production approach and a different set of source materials.

The Composer's Role: Music as Atmosphere

At Anydream, the distinction between music and sound design is deliberately blurred. The goal is not a soundtrack that plays alongside the environment – it is an audio landscape in which musical elements and environmental sounds are woven together into a single, coherent sonic texture.

The composer's first task is to establish the tonal world of the experience: the harmonic language, the instrumentation, the rhythmic approach (or deliberate absence of rhythm), and the emotional register of the music that will underpin each zone.

For Dune Desert, this required deep research into the musical traditions of the cultures the world draws from – Arabian, Persian, and Mesopotamian. The instrumentation selected for the score included:

• Oud – the Arabic short-neck lute, the foundational melodic instrument of Arabic and Persian classical music, with a particular resonance quality suited to reflective, meditative atmospheres
• Ney – the end-blown flute used across Middle Eastern musical traditions, with an airy, slightly breathy tone associated in Islamic cultural tradition with longing and spiritual openness
• Tabla and frame drums – rhythmic elements used sparingly; in the open desert and oasis zones, percussion is largely absent. In the bazaar, a subtle rhythmic underpinning supports the sense of activity
• Santour – the hammered dulcimer of Persian classical music, capable of both sustained harmonic textures and bright melodic lines
• String sections – recorded with extended techniques to create textures rather than melody: slow bowing, harmonics, sul ponticello passages that read as atmosphere rather than theme
• Processed silence – the composer treated silence itself as a material, using convolution reverb applied to near-silent recordings to create a sense of acoustic space that reads as presence rather than emptiness

The composition process followed the scene breakdown from the World Script rather than a conventional track list. Music was written for specific moments and transitions – the quality of the bazaar at midday is different from the bazaar at dusk, and the score needed to be different too, without the transition feeling like a different piece starting.

Field Recordings: The Foundation of Authenticity

Every Anydream world requires a library of field recordings – real-world audio captured in environments that correspond to the world's geography and cultural setting. Field recordings are what separate an environment that sounds generically "like a desert" from one that sounds like this specific desert, in this specific part of the world, at this specific time of day.

For Dune Desert, field recording work was conducted in and sourced from:

• Desert environments in Jordan, Morocco, and the UAE – capturing wind at various speeds across open sand, the specific acoustic quality of sand dunes that have been heated by full sun, and the remarkable silence of a desert at 3am
• Ancient urban environments with comparable stone construction – the acoustic signature of a narrow stone street is very specific and cannot be approximated from synthetic sources
• Regional markets and souks – recordings of active markets in Marrakech, Cairo, and Istanbul, capturing the specific sonic density of a trading environment in hot weather: the overlap of voices, the acoustic compression of a busy narrow street, the particular sounds of specific goods (ceramic, cloth, spice, metal)

Field recordings form the base layer of each sonic environment. They are the material from which the more constructed elements – music, designed sound, local wildlife – are built upward.

Local Wildlife: Getting the Biology Right

One of the most immediately readable indicators of environmental authenticity is wildlife sound. A desert that features birdsong from the wrong species or, more commonly, the wrong frequency range – is immediately wrong to the listener, even if they cannot say exactly why.

For Dune Desert, wildlife audio research covered the specific fauna of the Arabian Peninsula and North African desert regions. Animals whose sounds were incorporated include:

• Hoopoe (Upupa epops) – a bird found across the Middle East and North Africa with a distinctive three-note call, one of the most recognisable sounds of the region. Present in the oasis and city exterior zones.
• Rock pigeon (Columba livia) – the specific cooing of pigeons in stone urban environments; one of the most reliably authentic sounds of a working Middle Eastern city.
• Desert lark and wheatear species – sparse, intermittent calls used in the open desert zone to suggest biological presence without undermining the dominant character of the space, which is emptiness.
• Arabian sand gecko (Stenodactylus arabicus) – a small, high-frequency chirp used in the night scenes at the oasis, barely audible and placed spatially at ground level.
• Dromedary camel – vocalisation, breathing, and footfall sounds for the city and caravan scenes.
• Wind through specific vegetation – palm leaves produce a very different sound from broadleaf trees; this distinction matters and was recorded specifically.

All wildlife sounds were sourced from field recordings of the actual species in their actual geographic range. No substitutions or approximations were used.

Spatial Audio: Making Sound Behave Like a Real Space

Having the right sounds is only the beginning. In an immersive environment, sound must behave the way it would in the physical space – which requires spatial audio processing that goes far beyond stereo or even conventional surround sound.

Anydream uses binaural spatial audio processing: a method of encoding audio so that it carries information about the direction, distance, and acoustic environment from which each sound originates. When experienced through headphones or a compatible playback system, binaural audio produces a full three-dimensional sound field – sounds that are clearly in front of you, behind you, above you, or at floor level, and that move realistically as you move through the space.

This is implemented in Unreal Engine through the MetaSounds system – Epic Games' procedural audio engine that replaced the earlier audio system in Unreal Engine 5.

MetaSounds: Procedural Audio in Unreal Engine 5

MetaSounds is not a playback system. It is a node-based audio programming environment that allows sound to be defined procedurally – meaning that sounds are generated and composed in real time based on conditions in the environment, rather than played back from fixed recordings.

For Anydream's purposes, MetaSounds enables several specific capabilities:

• Dynamic wind – wind in the open desert is not a looped recording. It is procedurally generated based on parameters including wind speed, terrain exposure, time of day, and the presence of structures or vegetation. The result is wind that never repeats in exactly the same way, that changes character as you move from an exposed dune to the partial shelter of a rock formation, and that decreases in intensity and frequency content as you enter the walled city.
• Acoustic environment modelling – every space in Dune Desert has its own acoustic profile. The open desert has almost no reverb – sound dissipates rapidly in open air over sand. The canyon has a specific reverb character based on its geometry: stone walls at a specific angle, at a specific distance. The interior of the bazaar has the particular acoustic compression of a narrow space with heavy canopy coverage. These profiles are computed in real time using Unreal Engine's audio propagation system and adjust dynamically as the listener moves between and within spaces.
• Positional wildlife – wildlife sounds are not ambient loops. Each animal is placed as a positional audio source in the environment, at a specific location, with realistic spatial behaviour. A hoopoe in a palm tree at the oasis is in that palm tree – its sound comes from that direction, that height, at that distance. If you walk toward it, it gets louder. If you turn away, it moves behind you.
• Layered ambient construction – the ambient sound of each zone is built from multiple independent layers that can fade in and out based on time of day, location within the zone, and proximity to specific elements. The bazaar at dawn sounds different from the bazaar at midday not because a different recording is playing but because the layers have different weights – more bird, less crowd, less vendor activity, different wind intensity.

The Sound Mix: Final Integration

The final stage of sound design is the mix – the process of balancing all of the audio layers so that the sonic experience is coherent, emotionally appropriate, and technically correct across different playback systems.

Key considerations in the Dune Desert mix included:

• Dynamic range – the difference in volume between the quietest and loudest moments. The open desert at 3am is genuinely quiet. The bazaar at peak activity is genuinely loud. Maintaining this contrast is essential to the emotional impact of each zone.
• Frequency balance – ensuring that the low, mid, and high frequency content of each environment is appropriately distributed and doesn't compete destructively with the music layers
• Transition design – how the audio transitions as you move between zones. The shift from the bazaar to the open desert happens gradually, with the city noise fading and the wind building, rather than cutting abruptly
• Headphone and speaker compatibility – Anydream experiences are designed to be heard through headphones, where binaural spatial audio is most effective, but must also work on standard speaker setups. The mix is validated on both

The complete sound design process for a single Anydream world – from brief to final mixed audio – typically runs parallel to the latter stages of environment construction, taking between six and ten weeks. Sound design and visual construction are iterated together in the final weeks of production, with both disciplines informing adjustments to the other.

A visual decision can change a sound decision. If the lighting shifts the bazaar from a narrow shadowed space to something slightly more open, the acoustic needs to change too. The integration of these disciplines is one of the most complex and most important aspects of Anydream's production process.