Composing Distributed Soundscapes through Graphical Interaction by Thierry Dilger

Composition FMOD Interaction Osc sound design soundscape Spatialization

This project explores how graphical interaction can support the staging of listening situations within distributed soundscapes. Using GASPR, a custom real-time environment, sound behaviors are organized and distributed through OSC and FMOD Studio across a network of autonomous Raspberry Pi loudspeaker points.

➡️ This talk is part of IRCAM Forum Workshops Hors-les-Murs Tokyo 2026

How can a soundscape be composed without reducing it to a fixed timeline or a centralized representation of space? This project explores graphical interaction as a way to stage listening situations: organizing how sounds emerge, persist, move and become perceptible across a distributed environment.

Staging listening situations

Rather than treating sound only as material arranged along a linear timeline, the project considers the conditions through which listening takes place. A sound may emerge briefly, remain present as an environment, move between listening points or become perceptible through the listener’s traversal.

GASPR is used as a working environment for organizing these relationships. It structures soundscapes through listening fields, sonic entities and perceptual regimes of emergence, persistence and traversal.

The interface is therefore not conceived only as a conventional sequencer. Graphical forms represent relationships between sound behaviors, time, movement and listening points. Graphical gestures can trigger an entity, scan a sound structure, modify its persistence or transform how it unfolds within the listening field.

A graphical environment for sound behaviors

GASPR was initially developed as a graphical audio spatialization program. Its current form extends this approach toward interactive and distributed soundscape composition.

The project uses Rooms as successive listening contexts. Within each Room, graphical structures establish different relationships to duration, movement and interaction. Some sounds form persistent environmental layers, while others appear as events, moving entities or structures connected to the user’s graphical traversal.

The purpose is not simply to position individual sound sources within a virtual representation of a room. Instead, GASPR provides a visual and interactive framework for composing how sonic entities behave and relate to one another within a listening situation.

Distributed spatial sound

In its distributed mode, GASPR sends OSC control messages from a central interface to a network of Raspberry Pi clients through a dedicated local Wi-Fi network. Each client runs FMOD Studio, plays its own sound materials and is connected directly to a small loudspeaker.

Because the sounds are generated locally at each listening point, the system does not depend on a centralized multichannel sound card. The loudspeakers can be placed freely throughout a room, under chairs or within an installation space.

For the live demonstration, the system will use sixteen autonomous loudspeaker points. Each point consists of a Raspberry Pi client, a small loudspeaker and a portable battery. The complete network, including the local router, is self-contained.

This architecture allows a soundscape to unfold across multiple listening points. Spatialization is no longer limited to the movement of a source between conventional loudspeaker channels. It becomes a way of creating situated sensations of landscape, presence, distance and relation.

Research-creation with early childhood

This project is rooted in a long-term research-creation practice involving early childhood, developed through cultural venues, educational contexts and festivals.

In these situations, sound is not necessarily presented as an object with a clearly defined beginning and end. It can become an environment that children and accompanying adults enter, explore, pause within and revisit.

Distributed sound makes it possible to create differences of proximity, orientation and presence without relying only on frontal listening. The spatial arrangement encourages bodily exploration and changes in attention as listeners move through the environment.

These experiences have progressively informed the development of GASPR and the broader reflection presented here: the possibility of composing not only sound materials, but also the conditions through which listening becomes situated and shared.

Live demonstration

The presentation will combine a lecture with a compact sixteen-point distributed demonstration. It will show how a graphical gesture within GASPR is transmitted through OSC and interpreted by FMOD Studio on each Raspberry Pi client, producing coordinated sound behaviors in real time across the room.

The demonstration will focus on three operations:

  • triggering sonic entities;

  • scanning and traversing graphical sound structures;

  • transforming distributed sound behaviors in real time.

Rather than presenting GASPR only as a software tool, the lecture addresses a broader question: how can graphical interaction help stage the act of listening within a distributed soundscape?

Links

Current GASPR 2026 interface demonstration :

https://www.youtube.com/watch?v=aKamnnBJ89g

Foundational publication on the first version of GASPR :

https://hal.science/hal-03112220/document