The Science of Not Hearing

Focus sound is engineered so the brain never registers it as an event. That means: all melodic content filtered below 1,200 Hz to stay out of the alarm-zone frequencies your brain cannot habituate to, attack times above 10 ms to avoid triggering the startle reflex, tempo between 68 and 90 BPM to match resting heart rate, and no white noise, no minor seconds, no tritones. Every parameter in Particle's sound engine is derived from a specific psychoacoustic finding — the article below is the full evidence base.

This article is for sound designers, focus-tool builders, and anyone skeptical of "science-backed" claims who wants to see the actual papers.

The best focus sound is the sound you forget is playing.

That's not a tagline. It's a design constraint derived from decades of psychoacoustic research. Every frequency, every filter cutoff, every envelope shape in Particle's sound engine exists because a study told us it should — or warned us it shouldn't.

This article is the full evidence base. Not marketing claims. Not "science-backed" as a buzzword. The actual papers, the actual findings, and the actual parameters we derived from them.

#The alarm zone your brain can't ignore

Human hearing sensitivity peaks between 2,000 and 5,000 Hz. This isn't a preference — it's anatomy. The ear canal resonates at approximately 2,700 Hz, physically amplifying sounds in this range by up to 20 dB.¹

Evolution explains why. This frequency range contains the dominant harmonics of baby cries (fundamental ~400 Hz, but the energy that makes you notice is at 2–4 kHz), predator warning calls, snapping branches, and screams. The brain is hardwired to never habituate to sustained energy in this range.²

Alarm designers know this. Fire alarms operate at 3,100 Hz. Smoke detectors at 3,000–4,000 Hz. Hospital monitors at 1,000–4,000 Hz. Every siren, every alert, every sound designed to be impossible to ignore targets this exact frequency band.³

What we derived: All melodic content in Particle is lowpass-filtered at or below 1,200 Hz. This isn't a taste decision — it's a hard constraint. Even FM synthesis sidebands must stay below this ceiling. We calculate sideband spread for every preset: carrier × (1 + modulationIndex × harmonicity) < 1,000 Hz.⁴

No focus tool should put sustained energy where alarm bells live.

#Low frequencies lower cortisol

A pilot study on simple sound waves — not music, not melodies, just tones — measured cortisol response over one hour of exposure:

• Low-frequency exposure: 86% of subjects showed decreased cortisol
• High-frequency exposure: 65% showed increased cortisol

The critical finding: the cortisol shift correlated with physical frequency, not the listener's emotional response. Both sessions were rated as "unpleasant" by participants. But only high frequencies raised stress hormones.⁵

A separate meta-analysis of natural sounds (waterfalls, ocean waves, rain) confirmed this pattern: sounds with a brown-noise spectrum — heavy in low frequencies, naturally rolling off in the highs — systematically reduce sympathetic nervous system activity.⁶

What we derived: Particle defaults to brown noise for its texture layer. Pink noise is used sparingly for brighter presets, always filtered below 3 kHz. White noise is never used — its equal energy across all frequencies means it delivers exactly as much power in the alarm zone as anywhere else.

#Roughness: why some sounds feel wrong

When two tones sit within the same critical band — roughly 15–20% of the center frequency above 500 Hz — the auditory system cannot separate them. Instead, it perceives what psychoacousticians call roughness: a grating, buzzing quality.⁷

Roughness peaks at approximately 25% of the critical bandwidth. In nature, this quality characterizes screams and aggressive vocalizations. Research shows that rough sounds activate the amygdala more than smooth sounds, triggering a fear and alertness response even at low volumes.⁸

FM synthesis — the backbone of generative sound design — is particularly prone to creating roughness. Higher modulation indices produce dense sidebands that fall within critical bands, creating timbres perceived as "metallic" or "alarm-like."

What we derived: Modulation index is capped at 3. Harmonicity must be an integer (1, 2, or 3) to ensure harmonic rather than inharmonic spectra. Every preset goes through a sideband audit to verify that no adjacent partials sit within the same critical band in the 500–3,000 Hz range.

#The startle reflex bypasses your brain

The acoustic startle reflex is triggered by three conditions:⁹

1. Sudden onset (attack time < 5 milliseconds)
2. Unexpected loudness change (> 6 dB jump)
3. Intermittent on-off patterns

This is a brainstem reflex. It bypasses conscious processing entirely. Even at low volumes, sharp transients cause a micro-startle that prevents the nervous system from fully relaxing.

What we derived: Minimum attack time of 10 ms on all melodic and textural elements. The only exception is sub-bass kicks below 100 Hz, where the transient is felt rather than heard. No sudden volume changes. Activation fades are ≥ 2 seconds, deactivation fades ≥ 1.5 seconds. Every sound enters and exits gently.

#Tempo should match your resting heart rate

Research on music-induced relaxation consistently shows:¹⁰

• Tempo near resting heart rate (60–72 BPM) → decreased heart rate, increased heart rate variability
• Tempo significantly above resting heart rate → increased physiological arousal
• Tempo below 60 BPM → risk of perceived "drag," disconnection from the work

The alpha brain wave range (8–14 Hz), associated with relaxed focus, corresponds to musical subdivisions at tempos of 60–80 BPM.

What we derived: Particle's presets range from 68 to 82 BPM. The upper boundary — 90 BPM for high-energy presets — is only allowed when compensated with extra-gentle timbres. No preset exceeds this. Your nervous system sets the tempo, not the genre.

#Musical intervals carry emotional weight

Not all intervals are neutral. Research on consonance and emotional response shows clear patterns:¹¹

What we derived: Melodies are built from fifths, octaves, and minor thirds. Extended harmony (sevenths, ninths) creates dreaminess without tension. Minor seconds and tritones are formally prohibited — they're on the same avoid list as the alarm zone frequencies.

#Envelopes shape whether you notice

The shape of a sound's volume curve — its envelope — determines whether your brain registers it as an event or as an environment:¹²

Long release times (> 3 seconds) create natural overlap between notes, producing a continuous wash rather than discrete events. The ear perceives texture, not sequence.

What we derived: Primary pads in Particle use 2–5 second attack times and 3–5 second releases. Notes overlap naturally, creating a harmonic field rather than a melody. The listener doesn't hear individual notes — they hear an environment.

#The constraint list

These aren't guidelines. They're rules, derived from the research above. Every preset is validated against them before shipping.

• All melodic content filtered ≤ 1,200 Hz
• FM modulation index ≤ 3, harmonicity integer only
• Sideband audit: no energy above 1,000 Hz from FM synthesis
• No white noise (prefer brown, filtered pink below 3 kHz)
• Attack ≥ 10 ms (sub-kick exception ≥ 3 ms)
• Tempo range: 68–90 BPM
• No minor seconds or tritones
• Activation fade ≥ 2 s, deactivation ≥ 1.5 s
• No intermittent on-off patterns
• No sudden volume changes (> 6 dB)

When in doubt: quieter, softer, warmer. We compete with silence.

#What this means

Most "focus music" products don't publish their constraints. They say "science-backed" and leave it at that. We believe transparency is the stronger position.

Every parameter choice in Particle's sound engine traces back to a finding in this article. If you want to verify our work, the papers are linked below. If you disagree with our interpretation, we want to hear from you.

The goal isn't to claim we've solved focus sound. The goal is to show our work — openly, honestly, with citations — so that every design decision can be questioned, verified, and improved.

That's how research works. That's how Particle works.

#References