Sound in games often feels invisible when it is working well. Players notice visual worlds immediately. Landscapes stretch into the distance, characters move through environments, and stories unfold through action and dialogue. Sound tends to arrive more quietly. Footsteps simply seem to belong beneath a character, background ambiences appear to exist naturally around us, and a creature’s voice feels inseparable from its personality. Everything seems to fit together so naturally that the work behind these experiences often disappears from view.
Yet creating sound for games involves a challenge that differs fundamentally from many other forms of media. Film and television unfold through fixed sequences of events. A sound designer working on a film knows exactly when a door opens, when dialogue occurs, when music begins, and when tension rises. Audiences experience those moments in the same order every time. Games behave rather differently. Players stop unexpectedly, move in different directions, repeat actions endlessly, ignore objectives, or spend long periods interacting with things designers never anticipated would receive much attention. Some players rush directly through environments while others investigate every possible corner of a world. A sound designer may know what can happen inside a game, though cannot always know what will happen, when it will happen, or how often particular experiences will occur. Sound therefore cannot simply be attached permanently to images and left alone. It must continue adapting long after the designer has stepped away.
During an online guest lecture, Malin Arvidsson explored this challenge through reflections on her own experiences working across game audio. Throughout projects involving children’s games, procedural systems, and large-scale interactive worlds, a recurring idea gradually emerged. Game audio frequently involves building systems rather than constructing isolated sounds. Designers create frameworks, relationships, and behaviours that continue operating within worlds that remain unpredictable.
Arvidsson described discovering games somewhat unexpectedly. Having decided at an early age that she wanted to work with sound, she initially pursued sound engineering and recording work before later encountering opportunities in game production. Games had not necessarily appeared to be an obvious destination at the time. Film and television perhaps felt more visible as career directions, while game audio remained relatively unfamiliar. Yet after joining Audio Interactive and working on early projects, games gradually became something much larger than a temporary opportunity. Part of this attraction appeared to emerge from constant change. Technologies evolve rapidly, development processes shift, while projects rarely require exactly the same approaches twice. Many creative fields involve continual learning, though games introduce an additional layer of complexity through their combination of artistic decisions and technical systems. Sound designers are often required to think simultaneously about recording, editing, implementation, behaviour, memory, interaction, and player experience.
Some of the earliest examples discussed during the lecture illustrated how dramatically workflows have changed over time. While working on Action Man: Jungle Storm, implementation tools remained extremely limited compared with contemporary systems. There were no dedicated audio middleware environments, no simple methods for previewing sounds directly within gameplay, and no convenient ways of rapidly testing ideas. Implementation frequently involved manually replaying sections of gameplay while attempting to synchronise sounds externally. Looking back, the process appears cumbersome and time-consuming. Yet despite those limitations, hearing newly created sounds finally appearing inside the game still produced a strong sense of satisfaction.
Later projects introduced another challenge as assumptions taken from linear media no longer translated effectively into interactive environments. Arvidsson described work on Republic: The Revolution, where large numbers of character animations required accompanying sounds. Initial approaches appeared straightforward enough. Individual animations were paired with carefully designed sounds in much the same way they might be within film production. Footsteps, movements, and interactions each received specific audio elements designed to support visual actions. Problems quickly appeared once these sounds entered gameplay. Memory limitations immediately became one issue, with thousands of individual files consuming valuable resources. Yet another issue proved equally important. Players repeatedly encountered exactly the same actions throughout long periods of gameplay. A movement animation viewed once might feel entirely convincing, though hearing precisely the same sound attached to the same movement hundreds of times gradually became distracting rather than believable.
This problem reveals something broader about realism itself. Human beings often tolerate variation without noticing it consciously, while exact repetition becomes highly noticeable. Everyday experiences rarely unfold identically from one moment to another. Footsteps change subtly according to movement, surfaces, speed, and context. Someone walking across gravel rarely produces exactly the same sound twice. Objects interact slightly differently each time they collide, while environmental sounds fluctuate continuously. We generally ignore these small differences, though their absence can become surprisingly noticeable. Once a sound begins repeating with complete consistency, attention gradually shifts away from the world itself and towards the system generating it. Perfect consistency can therefore begin feeling less realistic than controlled variation.
Solutions required a different form of thinking. Rather than attaching one sound permanently to one action, sounds became collections of possibilities. Footsteps could exist within larger groups of variations, different surfaces could trigger different responses, and small adjustments in pitch, timing, and volume could introduce subtle differences between repetitions. Players no longer heard identical events replaying endlessly. Instead, they experienced systems capable of producing varied outcomes.
Arvidsson reflected on this through an observation extending beyond the immediate technical problem. She noted that changing sounds can sometimes create the impression that animations themselves are changing. Sound was therefore no longer simply accompanying visual information. It had begun influencing how visual information itself was interpreted.
Repetition emerged again through examples involving dialogue. While working on Evil Genius, background conversations between characters introduced similar difficulties. Real dialogue becomes recognisable very quickly once repeated frequently, though replacing speech with meaningless placeholder sounds created worlds that felt strangely artificial. The eventual solution involved constructing thousands of vocal recordings using invented forms of structured nonsense speech. Colleagues recorded large collections of vocal performances resembling language without becoming meaningful dialogue. The purpose was not literal realism. Players were not expected to understand these conversations or extract semantic meaning from them. Instead, the objective involved creating evidence that activity continued occurring around the player. Worlds rarely feel alive merely through visual detail alone. People often listen for small signals suggesting that environments continue existing independently of their own actions. Background conversations, distant movement, as well as changing environmental activity all contribute to the impression that spaces continue functioning whether or not the player directly observes them.
Memory constraints returned in a different form during discussion of LittleBigPlanet. Storage restrictions within the PSP version introduced significant constraints compared with larger console releases. Some reductions remained relatively straightforward. Numbers of variations could be lowered and certain content could be simplified, though environmental soundscapes proved more difficult. Long ambient recordings consumed considerable amounts of memory, while straightforward looping solutions introduced repetition problems of their own. Instead, Arvidsson described constructing simpler environmental foundations combined with shorter sound fragments including birds, insects, and environmental details. Individual elements could then appear according to changing probabilities and timings while introducing subtle variation. Rather than hearing static recordings replaying continuously, players experienced environments appearing more dynamic and less predictable.
Examples such as these suggested that technical limitations did not merely reduce possibilities. Constraints frequently redirected attention towards different forms of design thinking. Rather than storing larger quantities of material, systems could generate richer experiences from fewer resources.
Increasingly interactive systems introduced another layer of complexity. Physics systems created situations where players themselves generated outcomes that designers could not fully predict beforehand. Within LittleBigPlanet, players could construct objects using different combinations of materials and structures. Objects then collided using changing amounts of force under varying conditions. Questions that initially appeared simple quickly became more complicated. Which material should dominate when metal collides with sponge? Should paper dominate plastic? What happens when multiple materials contribute simultaneously? Questions such as these reveal how game sound often shifts away from designing isolated sounds towards establishing behaviours and rules. Designers create relationships and systems, allowing games themselves to determine outcomes dynamically.
Broader reflections on working within the industry also appeared near the end of the lecture. Networking, persistence, and long-term relationships emerged repeatedly throughout these discussions. Freelancing across games, film, and television introduced uncertainty alongside flexibility, requiring continual adaptation as projects, collaborators, and opportunities changed over time. One comment near the conclusion captured this relationship clearly. Arvidsson described game sound design as roughly forty percent creativity and sixty percent technical implementation and problem solving.
Initially this ratio may appear unexpected. Sound design often seems associated primarily with creativity and artistic expression. The examples discussed throughout the lecture suggested something slightly different. Creativity within games frequently emerges through solving problems. Memory restrictions, implementation systems, player unpredictability, and technical limitations all shape the final experience.
Players rarely notice these systems directly. They simply hear worlds that feel alive. Background conversations seem to continue without them, environments appear to change naturally, as movement feels connected to the spaces around it. Much of the underlying complexity disappears beneath the experience itself.
Perhaps that invisibility forms part of the achievement. Successful game audio may involve more than creating individual sounds. It may involve building worlds capable of continuing to surprise players long after the designer has stepped away. Rather than asking whether a sound works in isolation, a broader question may involve whether an entire system continues behaving convincingly once players begin doing things nobody predicted.
