HOW SOUND DESIGN MAKES HORROR GAMES TERRIFYING

Turn the volume off on Resident Evil 2 and it becomes a weird third person action game about a cop with bad inventory management. Turn the volume off on Silent Hill 2 and it becomes a man walking through fog. Turn the volume off on Alien Isolation and the Xenomorph just looks kind of goofy clanking around in vents. Horror games are audio products with pictures attached. That's not a diss. That's the entire reason psychological horror games work on a budget that wouldn't cover the catering on a Call of Duty trailer.

I've been thinking about this a lot because we're working on our own horror stuff and the single biggest lesson from every game I respect is that sound is cheap, scalable, and more effective than almost any visual trick you can pull. A jump scare costs you a model, animations, a scripted trigger, and a VFX pass. A creak in the wall costs you a WAV file and a randomizer. One of those scales to a whole game. The other scales to one scene.

Let me break down what actually works.

Binaural audio and why your skull matters

Binaural recording is the thing where audio is captured using a dummy head with microphones inside the ears, so when you play it back through headphones your brain thinks the sound is in the room with you. It's not stereo. Stereo is just left and right. Binaural accounts for the way your skull filters sound, the slight time delay between your ears, and the way the shape of your outer ear changes tone based on direction.

In games this gets simulated with HRTF, head related transfer functions. Hellblade: Senua's Sacrifice is the obvious example because the voices inside Senua's head are placed in 3D space around you. Sometimes they whisper right behind your left ear. Sometimes they argue across the back of your neck. You find yourself physically turning your head even though it does nothing mechanically. Your body reacts before your brain does.

Resident Evil 7 uses binaural in a more subtle way. The Baker house has audio objects placed everywhere, and when Jack is looking for you, you hear his footsteps above or below you with terrifying specificity. Not "somewhere upstairs." Directly above your right shoulder, moving toward the stairs. That precision is what makes hiding in a closet feel like actual hiding, not like waiting for a timer to end.

Here's the dev side of this that nobody talks about. Binaural only works with headphones. Speakers destroy the effect because your left ear hears the right speaker and vice versa. So if you're making a horror game and you care about sound, you're implicitly telling people they need to play with headphones, and you have to design around the fact that some players won't. Most good horror studios just assume headphones and compensate with a secondary stereo mix that's less scary but still functional.

Dynamic music that remembers you

Horror game soundtracks aren't songs. They're systems. The best ones run on what we call vertical remixing, where you have multiple layers of the same piece of music playing simultaneously at different volumes, and the game fades them in and out based on what's happening.

A typical horror track might have four layers:

Sneak layer. Low drones, maybe some distant percussion. Barely there.

Tension layer. Strings that build, off key piano, something with a pulse.

Combat or chase layer. Full drums, dissonant brass, heart rate stuff.

Sting layer. One off cues that fire when something specific happens, like a creature spotting you.

The game watches your state and mixes these in real time. In Dead Space, the music knows when a Necromorph is actively hunting you versus just in the level. The shift from "you are alone in this room" to "something is in the vent about to come out" happens with no loading screen and no interruption. You feel it before you see it. That's the trick. The music spoils the scare by a quarter second, which is enough time for your body to produce adrenaline before your conscious brain catches up, and that gap is where horror lives.

Resident Evil 2 remake does a really smart thing with this. The music in the safe rooms is calming, almost pretty. The music outside is abrasive. Crossing a threshold from a hallway into a save room produces an audible sigh of relief that's scripted into the audio system. You learn to crave that transition. Which means the moment the game denies you a safe room, or puts something inside one, it hits like a betrayal.

Silence as a weapon

This is the hardest trick in the book because it looks like doing nothing. Silence in a horror game is not the absence of sound. It's the deliberate removal of ambient cues that your brain has been using to stay oriented, replaced with a bed of low frequency rumble that you register as wrongness without being able to explain why.

Alien Isolation is the master class here. The Sevastopol station has an ambient bed of hums, creaks, and station noise that's basically constant. When the Xenomorph is nearby, the game pulls certain frequencies out of that bed. Not all of it. Just enough to make you realize the background has thinned. You don't hear the alien. You hear the absence of the station. Your brain does the rest.

SOMA does this with water. The entire game is set underwater, so there's this constant low drone of ocean pressure. In the scariest sections the game strips that away and replaces it with dry room tone, which is wrong because you're still supposed to be underwater. The dissonance between what you see and what you hear tells you something is about to happen.

Silent Hill 2 is even braver about it. Whole sections of that game have no music at all. Just footsteps on wet concrete and occasional distant industrial noises. James's footsteps are actually recorded wrong on purpose. They don't quite match the surface he's walking on in some rooms. You don't consciously notice but it nags at you. That game is full of these tiny audio mistakes that aren't mistakes.

Environmental audio and the language of old buildings

Every horror game has to decide what its building sounds like. Not the monsters, not the music. The walls themselves. This is where you get your groaning metal, your dripping water, your wind through broken windows, your pipes settling, your distant indecipherable mechanical noises.

The reason these work is that they read as evidence of scale and age. A building that groans is a building too big to see all of. A pipe that drips is a system that's broken. Wind through a window tells you there's an outside that's also wrong. These sounds create a sense of place that extends past the polygon count. You can make a three room horror game feel like a sprawling complex if the audio implies rooms you're never going to see.

Resident Evil always does this well. The Spencer Mansion has taxidermy creaking, grandfather clocks ticking in distant rooms, rain on windows you can't find. The Baker house in RE7 has structural groans and insects in the walls. The police station in RE2 has flickering fluorescents and distant gunfire from fights you missed. None of those sounds are tied to gameplay events. They just exist to make the building feel real, which makes the horror feel real.

One of my favorite examples is Amnesia: The Bunker. The Grunt, which is the thing in the walls, has footsteps that travel through the walls. You hear it thump along behind plaster. You hear it stop. You hear it breathe. Then you hear it move on. It's sound without a visible source, which is horror's favorite flavor. Your brain has to fill in the creature, and whatever you imagine is worse than whatever they could render.

Unreliable audio, or the fake out economy

Here's a technique horror games abuse and usually to great effect. You train the player to associate a specific sound with a specific threat, and then you play that sound without the threat. Or you play a sound that doesn't belong and don't follow up on it. The player waits for the scare. The scare doesn't come. The player waits longer. Eventually the player moves on. And then, twenty minutes later when the player has forgotten about it, the scare happens, disconnected from the setup.

Silent Hill's radio is the canonical version of this. The radio crackles when monsters are near. You learn this in the first hour. For the rest of the game the radio is your proximity sensor. Except sometimes it crackles and nothing is there. Or it crackles for something you can't see because it's in a wall. Or it stays silent when it shouldn't and you learn too late. The tool you trusted is compromised, and now you can't trust any sensory input.

Dead Space messes with this via the vacuum sections. Most of the game has normal audio. Then you enter a zero atmosphere section and the sound design cuts everything except your own breathing and muffled impacts conducted through your suit. Necromorphs in vacuum are silent. You can't hear them coming. The thing that's been keeping you alive the whole game, your ears, is gone. It's a mechanical change that produces a psychological effect, and it works because the game trained you to rely on audio first.

Resident Evil 2's Mr. X uses fake outs too. His footsteps are loud and consistent. Thump. Thump. Thump. You learn to track him by sound. But the police station is echoey and there are rooms where the acoustics throw his footsteps around, so you think he's upstairs when he's actually on your floor. Or you think he's close when he's fading away. The ambiguity is the fear.

Why sound is the cheapest horror tool

Here's the production argument, from a studio perspective. A good composer plus a good sound designer plus a decent middleware setup like Wwise or FMOD is maybe two to four people. The assets are small. You can iterate fast. You can A/B test scares by swapping WAV files. You can ship a patch that changes the feel of a whole level by tweaking a couple of reverb busses.

Compare that to visual horror. To scare someone with a model, you need the model, the textures, the rig, the animation set, the scripted triggers, the VFX, the lighting, the post processing. It's a massive pipeline. And once you ship it, any player who watches a YouTube video has seen the scare, so it stops working.

Audio is harder to spoil. You can watch a YouTube video of SOMA's underwater sections and still get chills when you play them yourself, because the binaural positioning puts the sound inside your head in a way video can't capture. You can read a text description of Silent Hill's radio and still flinch when you hear it for real. Audio is personal in a way visuals aren't. It bypasses the part of your brain that can analyze an image and goes straight to the fight or flight circuits.

The smartest horror devs I know spend more than half their development time on audio. They record foley in weird places. They process voices through broken amps. They build custom reverb profiles for every room. They tune the dynamic music for hours to get a single transition right. And the results are games that feel bigger, scarier, and more expensive than they actually are.

What this means for our own work

We're not making a horror game right now, but Kaiju Protocol has plenty of moments where scale and dread matter, and the audio pass is where those come alive. The sound of a building collapsing three blocks away is a bigger world than any vista shot can produce. The groan of steel before a skyscraper goes is more menacing than the collapse itself.

If you're working on horror, or anything that needs to feel dangerous, invest in audio before you invest in anything else. Get headphones. Get middleware. Get a sound designer who cares. Learn what HRTF does to a player's nervous system. Study Alien Isolation and SOMA and Silent Hill 2 with the video off.

The reason psychological horror games work isn't the monsters. It's what you hear before the monster arrives, and the silence after it leaves.

If you want the bigger picture on why all this works in the first place, I wrote more about what actually makes horror games scary.