Fluid Audio

  • Area: LudoNarrative (WP5)
  • Contributors: University of Southampton
  • Key Contact: David Millard (dem@soton.ac.uk)
  • Date: March 2026

1. Overview

Name

  • Popular Name: Fluid Audio
  • Alternative Names: Concertina Audio, Fractal Audio, Stretch Audio

Intent

To vary the length of an audio experience that is listened to while walking a route, such that the audio is longer the slower the listener walks.

Classification/Mapping

Operation/Type: Narrative

Underlying Principle: Personalisation

Strength of evidence

Implemented on site

2. Target

Problem

Locative audio experiences at cultural heritage sites (for example, audio tours at museums) usually leave queuing and playing of clips to listeners at trigger points, or automatically play when walking over thresholds. This can mean waiting around for audio to end before moving on. Adaptive audio allows for a similar audio experience with less friction.

Context

  • Situation: For this pattern, the user has to have a device that can sense location and also play back dynamically selected audio.
  • Environment: The play environment needs to be in a large enough to allow for audio playing between points.

Forces

The success of the pattern depends on:

  • Whether the application that queues up tracks based on walking speed adequately tracks that speed.
  • The sophistication of the writing so that the audio experience doesn’t present non sequiturs.

Consequences

Weaknesses:

  • Possible to have narrative vary quite considerably between users, which only works for certain kinds of narrative.
  • Narrative complexity in authoring.
  • Works best for a lengthy route with clearly differentiated paths.
  • Requires either a simple route or a sophisticated route identification system.
  • Not all users may be walking and may move faster than the design assumes (e.g. faster users on a mobility scooter).

Strengths:

  • The experience can feel complete regardless of walking speed.
  • Factual information and narrative can be prioritised in order of salience without needing to deliver it in that order.
  • There is no need for “dead air”.
  • The experience can be continuous rather than contiguous, without manual user intervention or intentional interactions.
  • The variety of possible experience can open up new narrative affordances.

3. Application

Audio clips are arranged in a hierarchical model of priority. Clips are dynamically selected based on the distance to waypoints.

Sensitizer:

Participants: Single user or shared group.

Breakdown:

  • Hierarchical audio model
  • Algorithm for converting current location into current depth
  • Delivery mechanism for audio

Variations:

  • Non-Linear: Instead of one route, multiple possible routes with a mechanism for playing the correct sequence based on which choice of route taken.
  • Decoupled Location: There could be multiple location nodes that can be travelled to at any given moment, with audio sequences playing regardless of which next location is chosen.

Rationale

This works because the audio playback is adaptive to the user rather than requiring the user to adapt to the system. It keeps attention on environment rather than device, contributing to a sense of place.

Implementation Details

Suggestions:

  • This can be achieved by tracking the user location and speed, and queuing up audio tracks based on a calculation of how much longer they have until they reach the end of the route or section of a route.

Issues:

  • If a user moves quickly and then slowly they can hear less than someone moving at a steady pace, as earlier potential audio in the chronology would have been skipped over.

Pitfalls:

  • Lots of additional content may be needed to fill out the audio model. The device needs to be switched on for the whole time, which can be tricky in some phone implementations. It is complex to test due to the wide variety of possible variations.

Impact on Immersion

Sensory
The focus of the user is away from their device, and they can be immersed in the landscape though not the wider soundscape.
Narrative
The user can be immersed in a story as they travel without stopping and starting of audio.

Ludic Considerations

  • Writing could add more playful elements, surfacing the choice outcomes of speed of movement (e.g. “if you want to do x then stay here a while, otherwise hurry on to the next turning”).

Example

This technique was developed in the Tales of Flint audio app. Listeners to the experience in the app walk a route across the countryside around the Neolithic site of Avebury, taking between 45 minutes to an hour. They hear six historical monologues which are lengthened or abridged accordingly to match the pace of the listener.

More information is available here:

Code: Ages of Avebury is available from Github:
https://github.com/LoGaCulture/Ages-of-Avebury

4. Supplementary Information

Biography

v.1.0 20/03/26
v.2.0 23/03/26

Discussion

Success depends strongly on the device knowing where the user is going and correctly judging their pace. The ideal experience is a continual stream of audio, but this requires a great deal of potential content to draw from and in any case can always be exhausted if someone chooses to wait in place. The fluid nature of the audio can either be made clear to the user or remain a hidden part of the underlying experience.

Related Patterns:


  • Skeuomorphism

Team


  • Dr Jack Brett – Lead Engineer
  • Dr Charlie Hargood – Academic Investigator and Architect
  • Dr David Millard – Academic Investigator and Architect
  • Dr Yoan Malinov – Engineer
  • Dr Bob Rimmington – Qualitative Researcher

Partners


University of Southampton

Related Resources