Disruptive Audio: when sound works against the game design goals
How do we know that the sounds we create support the game design and improve the player experience? For many sound designers, the answer is “Intuitively.” Even though I don’t want to undervalue the power or professional intuition, I think it has its limits. In this post, I want to explore cases where intuitive judgment is not enough. I’ll be looking at the disruptive potential of game audio from the audio UX perspective, focusing on sounds that distract, mislead, or irritate the player.
Our hearing is very good at driving bottom-up attention for evolutionary reasons. By hearing, we detect distant movement and vibration. It covers 360 degrees and works at long distances. We can hear what happens in the dark or behind an obstacle. Our ears don’t blink, ensuring a stable stream of information about our surroundings. Shutting our ears is much harder than turning away from something we don’t like. We always process a massive amount of auditory information. If we were actively following everything we hear, we would constantly suffer from sensory overload. That’s why evolution taught us to quickly redirect attention to potential threats and other survival-related signals or important sounds we are conditioned to.
Human active auditory attention is selective, which makes it disruptable. A sudden sound effect may distract the player from the current goal, causing a feeling of irritation. Since irritation is the most obvious case of the game audio being disruptive, I started my exploration by simply googling for “the most annoying sounds in video games.”
Common causes of irritation
I chose to ignore the trivial cases where sound effects were irritating because of bugs or poor quality of the assets; luckily, there weren’t many mentions of those. I also tried to prioritize the sounds reported by several people over those that appeared only once or twice. In the end, I had a good collection of examples that share a few common patterns.
Some sounds became irritating simply because of too many repetitions and little importance. This group includes repetitive “idle” phrases that play when you approach the NPC, noticeable ambiance elements, less informative character sounds with little variation, etc. Logically, most of those sounds didn’t annoy the players from the very beginning. The players found them to be excessive after getting familiar with the game and figuring out which sonic clues were helpful to them and which were or became redundant.
It was rather evident that players mention voice-overs much more frequently than sound effects. I assume this happens because we are hard-wired to pay more attention to human-made sounds. Since human vocalizations naturally drag our attention, sequential repetitions of the same pre-recorded phrase are more likely to become particularly annoying. On top of that, spoken messages are often longer and more cognitively demanding than other sounds because listening to them involves the language processing mechanisms that are required to extract the meaning from speech.
Then, there were many examples where the sounds exaggerated the message, generating false alarms. For instance, one RTS game informed the player that their unit was under attack in the offscreen space. The voice alert sounded like the unit is under a serious threat even at low-risk encounters, forcing the player to react, disregarding the severity of the problem. Another example came from an open-world RPG that started an overwhelmingly epic music theme every time the player approached an enemy on the map. If the player fought a generic, low-level foe, the emotionally charged soundtrack felt like a mismatch that degraded the gameplay experience.
Third, some players mentioned intense sounds that communicate a gameplay-critical message but keep playing while a particular gameplay state is active. Many games have a low-health alarm that doesn’t stop unless the player finds a way to restore health. Another typical example is a siren that plays when the enemies are chasing the player’s character. Both exist in the game for a good reason, but not necessarily help the player in an already stressful situation. Even though the players probably wouldn’t want these sounds to disappear entirely, they drag too much attention and make it harder to concentrate on the current goal.
Irritation control
Luckily, once we have identified these problems, we can try to prevent them.
The first step is to understand whether the sound has a significant ludic function. Is it there to inform the player about some gameplay event or provide necessary feedback, or does it act as a mood or a narrative element? If the latter is correct, we simply need to get some control over the repetition. Creating more variations is an obvious way to go, but on top of that, we can dynamically control the trigger chance through the player’s progression.
Once the player has heard the sound several times, we can make it less frequent or even turn it off for a while. This system can react to overall playtime, real-time, or a combination of both. If every time you pass by some non-player character they sing the same song, you will quickly get tired of it. But if you only hear it once in a couple of weeks, you may not even remember that you have heard it before.
If the sound is there to communicate a gameplay-relevant message, reducing the repetition is not an option. We want the player to rely on informative sounds, so we can not allow them to play occasionally. Here we should start by evaluating the message itself and understanding what information does it communicate. It is especially useful to think about the edge cases when the sound will misinform the player or appear without bringing any significant value. Once we found such cases, we could address them with separate messages or with adding more dynamicity to the existing one. For instance, we could strengthen a current alert with an extra layer to communicate a higher threat.
Of course, it would be too costly to design a dynamic system for every type of alert or similar informative sound in the game. So it is our job to look for cases where dynamicity will be the most effective. The rest could be simply “neutralized” to communicate less specific information that will fit into various contexts. Even though being ambiguous is rarely good when we design informative elements, I think it is better to be less specific than to occasionally misinform the player.
Let’s return to the RTS example from the above to see what I mean here. The game reported a high-danger situation every time the player’s unit was attacked in the offscreen space. The alert played even when the player’s army encountered a lonely enemy scout, which posed no threat. Since the player didn’t know the real situation, they had to interrupt their current activity and react. If it was nothing significant, they could feel deceived and frustrated. But if the message was more ambiguous, merely telling that the unit is under attack, the player could use their current knowledge and best guess to decide whether to react or not. Even though they could feel frustrated if they underreacted, they would know that it is a consequence of their decision, not of a reaction to a misleading message from the game.
Deliberately repetitive sounds like low-health alarms are harder to manage because they are often critically important for the gameplay or the narrative. Repetitive signals work great when we can quickly act to resolve the problem, like fasten the seat belt in the car. But when there is no immediate action to take, such alerts raise the stress level, nudging the players into reckless behavior by disrupting their concentration, and eventually making them more likely to fail their goal. You may argue that this is an efficient way to communicate the feeling of stress that the player’s character experiences. And you will be right if your design goal is to create a positive feedback loop, pushing an already failing player to fail faster. In other cases, the repetitive alarm works against the intention to stabilize the game.
A good tactic for irritation control here would be decreasing the intensity of the sound over time. When the player enters a critical state, they need to be informed about it. And even though they may want some reminder that the state is still active, the reminders don’t have to be equally intense for the entire duration. We could slowly replace the repetitive low-health alarm with less distracting signals, followed by a noticeable message, telling that the critical state is over. Police sirens during the car chase could become a little quieter in the mix after a minute of gameplay.
This technique applies far beyond irritation control. The idea of highlighting the novel sounds in the mix leads to a more vibrant sonic experience. We can emphasize the first moments of a thunderstorm to set the mood for the scene, turning it down after some time to give room to other elements of the soundscape. Or we can use more articulated sounds for the first few events in the sequence without being afraid to overwhelm the player.
Subtle disruptions
As I said before, irritating audio is a clear example of sonic disruption in games, but it doesn’t end here. We react to sounds at a pre-conscious, pre-attentive level. It is safe to imply that we sometimes get affected by them without noticing and understanding what exactly has happened. And if we don’t notice, we can only hypothesize.
The best example that comes to my mind is a poorly timed UI sound alert that pushes you to briefly check a HUD element on the edge of the screen. Redirecting your gaze to the side, analyzing the message, and then looking back to the center takes a few hundred milliseconds, which may be just enough for you to fail a challenging action sequence. Saying that such tiny details may affect player performance sounds far-fetched, but I believe that I have experienced this effect myself while working on a game. Anyway, I would be happy to see this hypothesis confirmed or disproved in an experiment.
Sonic disruptions can simply emerge from the interplay of different systems and elements within the game. They are more likely to appear in systemic games with emergent gameplay, but not limited to those.
Hellblade: Senua’s Sacrifice has received a lot of well-deserved praise for its audio design. It is indeed exceptional, but here is one observation about this game. When an enemy attempts to strike our character from the back, we hear an obvious audio clue that warns us about the dangerous situation. Listening to this signal is more efficient than trying to visually interpret more ambiguous attack animations from multiple enemies that move simultaneously. This sonic warning creates a non-trivial gameplay dynamic that makes it beneficial to have some enemies behind the characters back in the offscreen space. Such dynamic makes little sense from the narrative perspective, so I assume that it emerged unintentionally, in contradiction with the battle system design goals.
I don’t think this audio clue worsens the experience of playing Hellblade, but it demonstrates how a single sonic message may alter it. And even though it is relatively harmless in this case, it could be worse. Think of a combat music theme that starts once it detects the nearby enemies and ruins the surprise of an ambush on your character.
After reading this post, you may have come up with a reasonable question. Does this all mean that to avoid disruption, we’d better have fewer sounds in our games? Well, no. Our brain always searches for patterns because they ensure predictability. Unexpected silence has the same effect on us as a sudden noise. Imagine walking in the forest and accidentally not hearing the sound of your footstep. Your attention instantly gets redirected to your feet: “Did I step into something?”. In the same way, the absence of expected sonic feedback is equally disruptive for the experience.
Audio can affect the player experience and sometimes does it in rather obscure ways. In this post, I described a few common patterns behind the disruptive sounds in games, along with some techniques for irritation control. It goes without saying that the good sonic experience is a result of countless iterations, which are only possible when audio is treated as an integral part of the game development process.
Being no expert in auditory psychology and related fields, I feel like I barely scratched the surface of this topic. I’d be super interested in deconstructing different cases of emergent sonic disruptions, especially the less obvious ones, like in the Hellblade example from above. If you have experienced anything like that, please tell me about it! I would be happy to analyze similar cases in a separate post.
If you have enjoyed reading this, you may want to check my post on Ludic and Narrative sound in games, that has more examples of Disruptive audio.
