Sound & audio design

The register is the auditory equivalent of the visual philosophy: analog-electronic, spectrally restrained, operationally serious. Not cinematic (no BRAAMs, no horn-drops), not consumer app (no iOS tri-tone arpeggios), not arcade (no announcer voice, no achievement jingle), not retro-affectation (no CRT hum, no modem handshake). Think operator console audio - intentional, useful, forgettable when things are fine.

When to use sound

Sound carries information when:

1. The operator’s visual attention is saturated - tactical maps, dense dashboards, multi-screen setups. Audio extends the channel.
2. A time-critical event occurs - a new track enters the scope, a resource crosses a threshold, a permission request requires response.
3. Confirmation of a destructive action is in flight - a subtle tone after a termination command confirms it actually went through.
4. Media plays back - video, presentations, onboarding content.

Sound is not for:

• Routine UI confirmations (field save, menu open, button press). The UI already shows the change visually.
• Arriving at a page, scrolling, hovering. Not events; ambient interaction.
• Decorative reinforcement of branding on non-video surfaces.
• Background ambient in contexts where operators may be in already-loud environments (cockpits, TOCs, field ops) - the product cannot assume a controlled listening context.

The default is silence. Every sound defends its presence.

Frequency character

The system’s audio palette sits deliberately outside the bands where speech intelligibility lives (~2–4 kHz). Operators may be actively listening to comms while the UI plays audio - FDT audio must not compete with speech.

Event class	Frequency center	Character
Critical / low alert	200–500 Hz	Short pulsed tones, filtered. Never sub-bass rumble.
Standard UI event	800–1600 Hz	Transient clicks, soft attack, quick decay
Confirmation / success	1600–2800 Hz (brief)	Short sparkle - never sustained
Spoken voice (VO, narration)	250 Hz – 6 kHz	Standard broadcast voice range

Never occupy the 3–4 kHz band with sustained tones. That’s where vowel-consonant boundaries live; masking it interferes with the operator’s ability to understand radio or team comms while the product plays audio.

Duration

Class	Duration
UI tick / confirmation	40–120 ms
Notification cue	300–600 ms
Alert cue (single shot)	500–800 ms
Alert cue (repeating)	400 ms on, 1200 ms off, repeat until dismissed
Onboarding chime / app launch	800 ms – 1.5 s

Sounds longer than their class budget are over-designed. A 2-second UI confirmation is a consumer-app pattern.

UI sounds

Sparse, functional, event-driven. The sanctioned set:

Event	Signature
Destructive action confirmed	Filtered sub-500 Hz click, 80 ms. Not a beep.
Notification - priority	Dual-tone 800 / 1200 Hz, 400 ms total. Distinguishable from ambient.
Notification - routine	Single 1000 Hz soft tone, 250 ms. Skippable in operator focus mode.
Request denied / error	Descending 600 → 400 Hz tone, 300 ms. Not a buzz.
Connection lost	Two descending 500 → 300 Hz tones, 200 ms apart, 800 ms total
Connection restored	Mirror ascending 300 → 500 Hz, same cadence

Routine field saves, button clicks, and menu opens have no sound. Every additional UI sound has to prove it’s necessary to ship.

Alert tones - tactical products

GRIDWATCH and GHOST GRID carry audio alerts for time-critical tactical events. These are distinct from UI sounds - louder, more distinguishable, designed to penetrate operator attention in dense environments.

Three severity tiers, each with its own signature and repeat behavior:

Tier	Repeat	Muteable?
ADVISORY	Plays once	Yes, per-session
WARNING	Plays twice, 3 s apart	Yes, per-session
EMERGENCY	Repeats every 2 s until dismissed	Not during active missions. Operator can acknowledge (silences the tone and logs the ack) but cannot mute.

The operator must be able to distinguish the three tiers without looking at the screen. Different frequency centers, different envelopes - not just louder versions of the same cue.

Hard rules

• No real-world tactical audio mimicry. FDT alerts must not imitate MEDEVAC tones, chemical-agent alarms, fire alarms, missile-launch warnings, Emergency Broadcast System tones, or any other signal an operator might hear in a real emergency. Confusing a simulation alert with a real-world event is a safety failure.
• No phone ringtones. No musical phrases.
• No voice callouts in alert tones. Voice belongs in dedicated narrative layers - VO for video, guided onboarding - never in tactical alerts.

The specific alert signatures are developed with audio engineers; this section specifies the rules the signatures must satisfy, not the signatures themselves.

Ambient in-product audio

Most FDT products are silent by default. If ambient audio is used, it is optional (off unless toggled on) and serves a functional purpose.

Examples where ambient earns its place:

• GHOST GRID radar scope - a subtle sweep tone timed to the visual sweep rotation, giving an audio sense of rotation rate. Off by default; operators opt in if they prefer the kinesthetic reinforcement.
• GRIDWATCH live-ops mode - a quiet 60 BPM pulse indicating the system is live and processing, distinct from silent paused state. Off by default.
• QRF mission timer - optional audible tick for last-30-seconds timers, at 1 Hz. Off by default.

Ambient never plays in marketing, never plays on first launch, and never plays in authenticated but non-mission contexts (settings, profile, admin). Ambient is a tactical-focus feature, opt-in per operator.

Sonification

Sonification is the audio representation of continuous data - rising pitch for increasing temperature, faster pulse for accelerating rate, stereo pan for direction. Genuinely useful in operator contexts where the visual channel is saturated.

Sanctioned sonification patterns:

• Threshold crossings - pitch rises as a value approaches a limit; crosses to a distinct warning pattern at the limit. Useful for resource burn, fuel remaining, time-to-intercept.
• Stereo bearing - horizontal bearing to a track maps to stereo pan position (-180° full left, 0° center, +180° full right). Useful on headphones in GRIDWATCH and GHOST GRID.
• Pulse rate - rate encodes magnitude or proximity. An approaching track closes the gap between pulses until it becomes a continuous tone at intercept.

Sonification is always supplementary, never the sole signal. A sonified value also has a visible numeric readout.

Music

Music appears in video and presentations only. Never in product UI, never in onboarding, never as decorative background.

• Register - ambient, electronic, tonal. Think Brian Eno’s Music for Airports or Ryuichi Sakamoto’s late work, not Hans Zimmer, not trailer music.
• No BRAAMs, no horn drops, no “epic” crescendos. The visual philosophy rejects theater; the audio philosophy rejects the score that tells the viewer what to feel.
• Tempo - 60–90 BPM. Faster reads as marketing-energy.
• Loudness - mixed so voiceover sits clearly above the music. The music is context, not presence.
• Licensing - licensed library (Musicbed, Artlist, Marmoset) or commissioned original. No stock-music cinematic-epic playlists.

Voiceover and narration

When used in video, onboarding, or guided content:

• Register - conversational, technically literate, unstentorian. Never “in a world where…” Never broadcast-news gravitas. A thoughtful colleague explaining something clearly.
• Pace - approximately 150–170 words per minute. Faster reads as impatient; slower reads as condescending.
• Voice diversity - rotate across productions. No single “voice of FDT.” Different content pieces use different voices; the brand is not anchored to one narrator.
• Recording quality - studio-grade (Shure SM7B or equivalent, treated room, professional interface). No laptop microphones, no room reverb, no detectable noise floor.
• No voice callouts in UI or alerts. Voice is high channel load; tactical alerts and notifications don’t use it.

Sonic branding - the FDT signature

The system reserves one short audio signature for:

• FDT corporate video intros
• Product launch videos
• First-run onboarding moment (if the product has one)
• Presentation deck openers (optional, sparingly)

Character

• Duration - 1.0–1.5 seconds
• Structure - three-note phrase, ascending or descending, ending on a stable tone
• Timbre - warm synthesis (filtered sawtooth or sine-triangle blend). Not percussive, not orchestral.
• Frequency center - 400–1200 Hz, avoiding the speech-intelligibility band
• Tail - short decay, ≤400 ms. No long reverb, no stadium-size space.

The signature itself is developed with audio designers under these constraints - the design system specifies the rules, not the melody.

Mute and volume control

Every product with audio provides three levels of control:

1. System mute - one keyboard shortcut or single toggle, mutes all product audio instantly. ⌘ + ⇧ + M on desktop.
2. Category mute - separate toggles in settings for UI sounds, Alerts, Ambient, Voice. Operators can disable UI sounds while keeping tactical alerts.
3. Master volume - single 0–100 slider, persisted per-user per-product.

EMERGENCY-tier tactical alerts ignore system mute during active missions (see Alert tones). This is the only exception; all other audio respects mute absolutely.

Accessibility

• All video and presentation content carries captions - SDH (captions for the deaf and hard-of-hearing, including non-dialogue audio cues). Never auto-generated; human-reviewed before publication.
• Audio descriptions available for video content where visual information carries meaning not conveyed by the audio track. An AD track describes on-screen action during natural pauses.
• Visual equivalents for all audio cues. No information is delivered by audio alone. Notification sounds are always paired with visible notifications; alert tones with visible alerts. Operators who can’t hear the audio never miss information.
• Mute does not hide information. Setting master volume to 0 silences the audio channel; the system continues to deliver information visually at full fidelity.
• Transcripts for long-form audio/video content, linked as searchable text.

Spatial and stereo audio

For tactical products used on headphones:

• Bearing sonification - horizontal bearing to a track maps to stereo pan. Elevation is not sonified (vertical spatial audio is inconsistent across headphone types).
• Mono fallback - all audio must sound correct on mono systems. Spatial cues degrade gracefully; no information is delivered by stereo position alone.
• No 3D / binaural without explicit operator opt-in and calibration - HRTF varies per listener, and uncalibrated 3D audio produces confusing perception.

Delivery specs

• Sample rate: 48 kHz, 24-bit
• Source format: WAV (uncompressed)
• Delivery format: AAC 256 kbps for web audio assets
• Loudness targets: −23 LUFS for broadcast video, −14 LUFS for social, −16 LUFS for podcast / narrative content
• Dynamic range: preserved. No aggressive compression that flattens quiet dynamics. UI audio specifically retains transient character - clipping or limiting destroys the functional cue.

Anti-patterns

• Cinematic stings on product reveals, section transitions, dashboard loads - never.
• Hollywood radio affectation - static, bleeps, squelch-tail in video. FDT products are clean audio; stylized broken-comms audio is a consumer-entertainment trope.
• iOS-style notification tones - FDT sounds are not tri-tone arpeggios, bell chimes, or glassy sparkles.
• Mimicking real tactical audio - MEDEVAC, missile launch, chemical alarm, fire alarm, air-raid siren. Safety-critical real-world signals must never be confused with FDT product events.
• Infinite loops without user control - ambient tracks, “hold music,” startup loops. Every audio has an exit condition or an operator-accessible control.
• Musical phrases in UI sounds - a melodic sequence in a UI tone reads as “cute app,” not operator tool.
• Voice callouts (“Notification received,” “Track detected”) outside of explicit guided narration.
• Sub-bass rumble - impressive on studio monitors, inaudible on laptop speakers and field-hardened tablets. Not a reliable signal.
• Scoring that tells the viewer what to feel - when the music swells to tell you the moment is Important, that’s consumer-marketing energy. FDT doesn’t telegraph.
• Keypad clicks on every keystroke. UI clicks are for meaningful actions, not ambient confirmation.