‹ Journal

Artificial identities · Lyfera · 2026-10-26

An artificial character needs a world you can hear

Sound gives an artificial character a place beyond the frame. A room, a corridor and an unseen presence can turn a sequence of images into a particular encounter.

Donna alla stazione centrale футuristicale

Imagine a short film in which an artificial talent finishes wrapping a parcel at a worktable. The frame shows paper, string and a closed door. Beyond it, someone moves a chair. A bicycle passes in the street. The figure pauses before tightening the knot. Nothing in the picture establishes who else occupies the building, but the sound has begun to give the room neighbours, an outside and a reason to listen.

For a mentor developing screen work around an artificial identity, those relationships deserve a place in the creative brief. A recognisable face can move through several attractive images while the surrounding world remains difficult to imagine. Sound offers another way to make that world specific: it can establish what continues beyond the frame and what the character's actions disturb.

A recent historical essay provides a useful occasion to reconsider that work. In James Wierzbicki's study of Gil Mellé's score for The Andromeda Strain, published on 21 October 2026, director Robert Wise's brief emerges as central to the score's unusual character. Wise wanted music that approached sound effects rather than familiar orchestral accompaniment. The subject is a film from 1971, not a new AI production technique. Its relevance lies in the relationship between a sonic choice and the world a film asks its audience to enter.

Give the room an existence beyond the picture

Return to the hypothetical worktable. If the soundtrack contains only the scrape of paper and the tightening string, the scene could feel enclosed around one task. Add a muffled chair moving upstairs, and the same task takes place within a shared building. Add conversation just beyond the door, and wrapping the parcel might begin to look like preparation for an encounter. Each treatment supplies a different situation for an otherwise unchanged figure.

Sébastien Balanger's HEAR module on DARIAH-Campus, published on 30 September 2026, describes five sound treatments of a stable visual scene, ranging from familiar ambience to intrusive noises and amplified bodily sounds. It introduces the position from which a viewer seems to hear, alongside sounds originating outside the frame. This is teaching material for audiovisual analysis, rather than an experiment establishing how every audience responds. It offers a useful vocabulary for discussing what a soundtrack contributes.

In our invented film, the creative team could begin by deciding what kind of building surrounds the worktable. Is it an isolated studio, a workshop with other people still working, or a room being quietly emptied? Those alternatives would give the same parcel different circumstances. Selecting an appealing ambient track before deciding between them could allow the sound library to supply a setting the story never chose.

The brief need not contain a complete floor plan or a noise for every object. It needs enough relationships to guide a collaborator: the corridor is occupied, the street is distant, and the closed door separates the talent from someone expected to arrive. These choices give the sound designer something to interpret. They also leave room for a quieter or more surprising solution than a literal illustration of every visible movement.

Decide where the audience is listening

Suppose the invented sequence cuts from the whole room to the talent's hands. Should the paper suddenly sound much closer? That change could bring the audience into the concentration of the task. Keeping the surrounding room prominent might instead suggest that the figure is listening beyond it. The picture establishes a new distance; the soundtrack can follow that distance or develop another relationship deliberately.

An accidental change would be harder to interpret. If the corridor disappears at every close-up and returns at every wider shot, the room may seem to start and stop with the camera. A continuing sound can help the shots belong to one place, while a chosen interruption can make an event consequential. The useful distinction concerns what the change means in the sequence, rather than a rule that every cut must sound seamless.

This develops a different question from giving an artificial character time to react. Timing determines when a pause or gesture can be read. Sound can establish what there is to react to. A held expression acquires another possible meaning when a chair stops moving beyond the door. The hypothetical character has not gained a new facial performance, but the scene has given the existing performance a more exact circumstance.

Let a sound make a promise it can keep

An audible event can introduce something the rest of the film must accommodate. In the parcel scene, a key turning in a lock would suggest a particular kind of access. A knock would suggest someone asking to enter. A repeated mechanical pulse might be understood as equipment in the room, or as music supplied for the audience. These alternatives can sound equally polished while placing different demands on the story.

Wierzbicki's account of Mellé's score is helpful here because it treats the music within the film separately from its presentation on an album. A sound's interest in isolation does not settle what it does beside an image. His interpretation of the score's recurring pulsation concerns how it contributes to the film's developing threat. Applying that attention to an artificial talent's short film means asking what a selected sound makes available in this particular scene.

The team might deliberately leave the pulse's origin uncertain. It could begin as something apparently mechanical and gradually become part of the music. That ambiguity needs a reason in the work. If the scene instead requires an ordinary workshop, an ominous pulse may create an expectation the remaining images cannot sustain. A compelling sound can be the wrong choice without being a poor piece of sound design.

Leave something for the character to change

In a final hypothetical version, the corridor remains audible as the parcel is wrapped. The figure pulls the knot tight, sets down the scissors and waits. The chair outside stops. A knock follows. The surrounding sound has allowed a small action to become part of a relationship: someone is preparing something, and someone else is about to receive it.

That progression could work without a spoken line. Our discussion of performance in a convincing artificial voice concerns how delivery gives words a purpose. Here, purpose develops through the relation between a task and a surrounding presence. A mentor can consider both forms without making speech responsible for explaining everything the scene has left unspecified.

These are editorial proposals for commissioning audiovisual work, not a claim that Lyfera supplies a sound production service. The film would still need its soundtrack made, reviewed and heard with the finished images. The next useful brief could be as concrete as the parcel itself: keep the corridor present, let the wrapping interrupt it, and give the knock time to arrive. The artificial talent then occupies more than a succession of frames. There is a world around the figure, and something in that world can answer.

Afvisning af valgfri værktøjer begrænser ikke din adgang til Lyfera. Google Tag Manager indlæses først efter valgfrit samtykke.

Dit valg gemmes i seks kalendermåneder. Nye formål eller væsentlige ændringer kræver et nyt valg. Læs mere.