Generative sound creates music and sound that can change over time.
Instead of playing back one fixed composition, a generative system defines relationships: rules, probabilities, behaviours, inputs and constraints.
The result is sound that can evolve, renew itself and respond to context without repeating in exactly the same way.
In installations, exhibitions and adaptive environments, this makes sound feel alive: present over time, responsive to change and connected to the behaviour of the space.
Generative sound describes sonic systems that produce variation through rules or processes.
A composition may define the material, the structure and the limits of the system, while the exact result continues to unfold over time.
This can include algorithmic composition, probability, data, real-time input, machine learning or simple rule-based behaviours.
The important point is not the technology itself. The important point is that sound becomes a system rather than a fixed timeline.
Generative sound is useful when an experience needs continuity without exact repetition.
A museum room, installation or public space may run for hours, days or months. A fixed loop can quickly become predictable or tiring. A generative system can keep the sound world coherent while allowing it to change.
This does not mean endless randomness.
The system needs a clear musical and dramaturgical frame: what can change, what should remain stable, how dense the sound may become, when it should withdraw and how it belongs to the atmosphere of the space.
Generative sound can be designed through different kinds of systems.
Rule-based composition
Musical or sonic events are created through defined rules. The system decides when something happens, how it develops and how elements relate to each other.
Probabilistic variation
Sound changes through controlled chance. This allows variation without losing the identity of the composition.
Data-driven behaviour
Data influences how the sound system develops over time. The data may shape rhythm, density, movement, intensity or transformation.
Machine learning and AI
Trained systems can generate or transform musical material. In spatial contexts, this becomes interesting when the output is shaped by a clear artistic and scenographic frame.
Real-time adaptation
Sound changes in relation to time, movement, sensors, environmental conditions or other live inputs.
Generative sound is closely connected to interactive sound and data sonification, but each term describes a different emphasis.
Generative sound describes a system that produces variation over time.
That system can become interactive when visitor action, movement or presence influences its behaviour.
It can become data-driven when information, datasets or live conditions shape how the sound unfolds.
In many installations, these approaches overlap. A sound system may be generative, shaped by data and responsive to interaction at the same time.
Generative sound becomes most meaningful when it belongs to the spatial and dramaturgical structure of an experience.
The question is not only how the system generates sound, but what role that behaviour plays in the space.
It may create long-duration atmosphere, make a hidden process audible, follow the rhythm of visitors, or give an installation the feeling of being alive.
Sound scenography defines how this behaviour relates to architecture, movement, media, attention and time.
Related practices include sound scenography, interactive sound, data sonification and spatial audio.
Examples of generative sound in our work include:
A generative soundscape follows patterns of animal migration. Low frequencies trace terrestrial movement, high tones follow birds and fluid harmonies suggest marine life, creating a living composition about biodiversity in motion.
Machine learning translates river behaviour into an audiovisual composition. Generative music follows shifting flow, erosion and sediment patterns across the visual system.
An artificial voice develops from chaotic frequencies into structured speech-like sound. Generative sound makes a learning process audible as the system begins to form language.
More than 300 individual objects create rhythm and light through vibration. Synchronisation is heard directly from the objects, turning behaviour into sound without loudspeakers.
A generative sound engine transforms white noise into a three-state thunderstorm. Sound follows the movement and tension of the kinetic sculpture, shifting between accumulation, intensity and release.
Visitors explore different forms of collaboration with artificial intelligence. Generative and AI-based systems become playable instruments for composing, conducting and improvising.
The first step is to define what should stay stable and what should be allowed to change.
From there, a generative sound system can be designed through musical rules, probability, data, live input, machine learning or interaction.
This work can include composition, sound design, programming, data mapping, interaction design, spatial audio and on-site tuning.
The goal is not to make sound endlessly variable. The goal is to create a system with a clear musical identity and a meaningful behaviour in space.
What is generative sound?
Generative sound is sound created through rules, systems or algorithms. Instead of playing back one fixed composition, the system produces variation over time.
Where is generative sound useful?
It is useful in installations, exhibitions, public spaces and adaptive environments where sound needs to continue for long periods without exact repetition.
Is generative sound the same as interactive sound?
No. Generative sound can evolve independently through internal rules. Interactive sound responds directly to input such as movement, gestures, sensors or visitor actions. Many projects combine both.
Does generative sound mean AI-generated sound?
Not necessarily. Generative sound can use simple rules, probability, data, algorithms or AI. AI is one possible method, not the definition of the field.
How does generative sound relate to sound scenography?
Generative sound defines how a sonic system evolves. Sound scenography defines how that evolving behaviour belongs to space, dramaturgy, movement and visitor experience.
Planning a long-duration installation or adaptive sound environment?
We design generative sound systems that evolve over time and respond to context.
