There's a moment in almost every great horror film where the music stops. Not fades — stops. And in that silence, something is worse than whatever came before it. The sound design in horror isn't about making things loud or startling. It's about manipulating the listener's nervous system — using frequency, timing, texture, and space to create a physical sensation of dread before anything has even happened.
This is some of the most sophisticated and frankly diabolical audio work in any genre. Horror sound designers understand psychoacoustics — the science of how sound affects human perception and emotion — at a level that most music producers never need to think about. But the techniques they use are available to any producer with a DAW and a willingness to be cynical with our hardwired survival instincts.
This is the guide to those techniques.
Why Certain Sounds Scare Us
Before getting into technique, it helps to understand why certain sounds trigger fear responses in humans. This isn't arbitrary — it's evolutionary.
Infrasound — frequencies below 20Hz, below the threshold of human hearing — can't be heard but can be felt. Think the first time we meet a T-Rex in Jurrasic park. Certain infrasound frequencies cause physical sensations including unease, disorientation, and a feeling of being watched. Large animals like tigers produce infrasound as part of their vocalizations. Your nervous system evolved to respond to those frequencies with fear before your conscious brain could identify the source. Infrasound in the 18-19Hz range has been associated with feelings of supernatural presence — which is why it appears in horror film scores and why certain old buildings with infrasound-generating machinery have reputations for being haunted.
Dissonance — the tension created by certain combinations of frequencies — triggers a threat response in the human auditory system. Intervals like the tritone (the diminished fifth, historically called diabolus in musica — the devil in music, not joking) create a sense of unresolved tension that the brain reads as a warning signal. This is why horror music tends to avoid resolution. The tension is the point. Cadence releases, avoid the cadence.
Non-linear sounds — sounds that contain the chaotic, unpredictable qualities of distressed human and animal vocalizations — trigger threat detection systems in the listener. A sound that's partially recognizable as a human voice but wrong in some fundamental way is more disturbing than either a clearly human voice or a clearly non-human sound. This is called the uncanny valley of audio, and horror sound designers exploit it constantly.
Sudden silence after sustained noise — the auditory system treats unexpected silence as a potential threat signal. When sustained sound stops abruptly, the brain interprets the silence as evidence that whatever was making the sound has stopped — possibly because it detected you. This is why horror films use silence as aggressively as they use loud sounds. Jump scares are usually preceded by a brief silence.
The Foundation: Drones and Sustained Tension
The most fundamental tool in horror sound design is the drone — a sustained tone or texture that creates an ambient sense of unease without calling attention to itself.
A well-designed horror drone does several things simultaneously. It occupies the low-mid frequency range where the human voice lives, creating a subliminal sense of presence. It moves slowly — subtle pitch variations, gentle filter sweeps, barely perceptible rhythmic modulation — in a way that keeps the listener's auditory system engaged without giving it anything to lock onto. And it never resolves. A drone that resolves to a stable pitch releases tension (It's not a horror, but Dunkirk does this brilliantly, the resolution in the soundtrack when all of the boats show up is palpable.). A drone that keeps moving, keeps searching, never settling — that maintains it (There's that Dunkirk example again.).
Building a horror drone:
Start with a simple oscillator — a sine wave or a slightly detuned combination of sine waves works well. Keep the frequency in the 80-200Hz range for maximum subliminal weight. Apply very slow LFO modulation to the pitch — a rate of 0.1 to 0.3Hz, depth of just a few cents. The pitch should waver almost imperceptibly, like something breathing.
Layer a second oscillator a tritone above the first. The tritone interval — six semitones — creates the maximum harmonic tension available in Western music. Let it breathe at a slightly different rate than the first oscillator so the two are never quite in sync.
Add a long reverb — room size large, decay time four seconds or more, high damping so the tail darkens quickly. The reverb should blur the edges of the sound, making it feel like it's coming from everywhere and nowhere simultaneously.
Finally, automate a slow filter sweep across the whole texture — a low-pass filter very gradually opening and closing over a cycle of thirty seconds to a minute. The movement keeps the drone alive without making it obviously animated.
Reverse Reverb
Reverse reverb is one of the most recognizable techniques in horror sound design (and hair metal, kind of horror adjacent), and one of the most powerful. It creates the psychologically unsettling sensation of a sound arriving before it should — a reverb tail that leads into the source rather than following it.
The technique:
Take the audio you want to process. Reverse it in your DAW. Apply a large, long reverb to the reversed audio. Then reverse the entire result — audio plus reverb — back to its original direction. What you're left with is the original audio with a reverb tail that swells in before the sound rather than decaying after it.
The effect is deeply unsettling because it violates one of the fundamental rules of how sound works in physical space. Reverb always comes after a sound in nature — it's the room responding to the source. When the reverb arrives first, something is fundamentally wrong with the physics of the space, and the brain recognizes this as a threat signal even if the conscious mind can't articulate why.
Reverse reverb works particularly well on vocals, piano, and any acoustic source where the natural attack and decay are well-defined. The more recognizable the original sound, the more disturbing the reversal.
Pitch Manipulation and the Human Voice
The human voice is the most emotionally loaded sound the human auditory system processes. We are exquisitely sensitive to its nuances — small changes in pitch, timing, and timbre carry enormous emotional information. This sensitivity makes the voice the most powerful raw material in horror sound design.
Pitch shifting down — particularly extreme pitch shifting of one or two octaves — transforms a human voice into something that retains the vocal quality but loses the human scale. The result is something that sounds like a voice but belongs to something much larger than a human being. This is a staple of demon and monster characterization in film, and it works because the voice is still recognizable as a voice — which triggers the uncanny valley response.
Pitch shifting up creates the opposite effect — something smaller, more fragile, more childlike. Used on an adult voice, extreme upward pitch shifting creates a sound that's simultaneously childlike and wrong (or a Scooter sample, you know, childlike and wrong), which triggers protective instincts while simultaneously activating threat responses. This is why so much horror uses children's voices and music boxes — both sit in a register that feels vulnerable and innocent while carrying associations of abandoned spaces and lost things.
Formant manipulation — changing the resonant qualities of a voice without changing its pitch — allows you to shift a voice's perceived size and character without the obvious chipmunk effect of simple pitch shifting. Most modern pitch shifters and vocoders include formant control. Shifting formants down while keeping pitch constant makes a voice sound larger and more resonant, like a voice in a large empty space. Shifting formants up while keeping pitch constant creates an unsettling, almost insectoid quality.
Layering multiple pitch-shifted versions of the same vocal — the original at pitch, a copy shifted down a fifth, another shifted down an octave, all with slightly different timing offsets — creates a choral effect that suggests a crowd or a multitude where there should be one voice. This is extremely effective for suggesting something inhuman wearing the shape of something human.
Noise and Texture
Pure tones and musical intervals are only part of the horror sound design palette. Some of the most effective unsettling sounds are built from noise — broadband, atonal, textural.
White noise shaped by a filter becomes a powerful tool for suggesting environments. Heavily filtered white noise with a resonant peak in the 2-4kHz range suggests wind, breath, or whispered voices just below the threshold of intelligibility. This frequency range is where consonants in human speech live — so the brain keeps trying to resolve the noise into words and failing, which creates a sustained state of alert attention.
Granular synthesis breaks audio into tiny fragments — grains — and reassembles them in ways that preserve some characteristics of the original while fundamentally transforming others. Running a human voice through a granular synthesizer with a low grain density and random pitch scatter creates something that retains the vocal texture while losing all intelligibility and most humanity. It sounds like a voice that has been scattered — which is exactly the psychological effect it creates.
Convolution with unusual impulse responses — using a convolution reverb with an impulse response recorded from a non-traditional space, like a drainage pipe, a metal tank, or a concrete bunker — creates reverbs with unusual resonant characteristics that don't correspond to any natural listening environment. The result sounds wrong in a way that's difficult to identify consciously but immediately obvious emotionally.
Silence and Dynamics
Horror sound design uses dynamics — the relationship between loud and quiet — more aggressively than almost any other genre.
The jump scare is the cheap version of this principle. A sudden loud sound after relative quiet triggers the startle reflex — an involuntary physical response that has nothing to do with actual fear. It's effective but blunt.
The sustained tension build is the sophisticated version. Gradually increasing the density and intensity of a sound texture over thirty seconds to several minutes — adding layers, widening frequency content, increasing reverb, pushing the low end — creates a sense of mounting dread that is physiologically different from a startle. The listener feels the threat approaching before it arrives, which activates the anticipatory fear response rather than the startle response. Anticipatory fear is more sustained, more exhausting, and more emotionally affecting than the startle reflex (Still beating this horse, but the first half of Dunkirk is so stressful because of the soundtrack, remove the soundtrack, and it's just people waiting on a beach with explosions in the background.).
Silence as punctuation — ending a sustained build with abrupt silence rather than a loud sound — is often more effective than the expected loud impact. The brain, primed for threat, interprets the silence as evidence that the threat has arrived and stopped announcing itself. The imagination fills in what the sound design leaves out. Nothing a sound designer creates is as frightening as what a listener's brain creates in silence.
Practical Horror Sound Design Exercises
The heartbeat — record or synthesize a low-frequency pulse at around 60-80 BPM. Process it with a gentle distortion or saturation to add harmonic richness. Automate the tempo to gradually increase to 120 BPM or above over sixty seconds. The listener's heartbeat will follow — a phenomenon called entrainment — creating a physical anxiety response that precedes any conscious evaluation of the sound.
The whisper layer — take any melodic element in your track and add a layer of heavily filtered white noise just below the audible threshold. Set a resonant peak in the 3kHz range and automate it to subtly swell and recede. The listener will perceive something almost being said without being able to identify what, which keeps the auditory system on alert.
The detuned music box — sample or synthesize a music box melody. Pitch shift each note slightly differently — some sharp, some flat, no two exactly the same — so the melody is recognizable but fundamentally out of tune. Add a convolution reverb with an unusual impulse response. The familiarity of the melody creates an expectation of safety that the detuning immediately violates.
The room that breathes — using automation, set a long reverb decay time to cycle very slowly between two seconds and eight seconds over a period of thirty to forty seconds. The room itself appears to inhale and exhale, suggesting an organic presence in what should be an inanimate space.
Putting It Together
Horror sound design is architecture. Individual techniques — drones, reverse reverb, pitch manipulation, noise texture, dynamic control — are building materials. The craft is in how they're assembled.
A sustained drone establishes presence. A slowly building texture creates anticipation. A voice element processed beyond recognition suggests something inhuman. A reverse reverb swelling in before a source suggests a violation of natural law. A sudden silence suggests arrival.
None of these techniques are complicated. All of them require intention — a clear sense of what emotional state you're trying to create and why each element serves that state. The difference between horror sound design that genuinely disturbs and horror sound design that just sounds dark is intentionality with a bit of that aforementioned cynicism. Every element should be doing specific psychological work.
The tools are in your DAW and this isn't rocket science. Hack the lizard brain with sound, and make people who are paying to be uncomfortable experience it.
