/// Six Modules, One Engine

Abilities

Every module below runs inside Flux Engine V5, sharing one analysis pass. Expand each one to see what it's actually listening for.

01 Clarity Engine
Room tone & noise, gone without the smear. +

Most noise reduction works by subtracting a fixed noise profile from the whole signal, which is why it leaves that underwater, phasey smear behind. Clarity Engine rebuilds the noise floor continuously, re-profiling every few hundred milliseconds so it never has to over-subtract to stay safe.

It's tuned specifically for vocal recordings: computer fan hum, room reflections, mic self-noise, and bleed from headphones are all recognized as separate problems instead of one generic "noise" category.

Listens For

  • Steady-state hiss and hum
  • Room reflections and reverb tails
  • Headphone bleed on double-tracked takes
02 PitchLock
Correction that keeps the performance. +

PitchLock tracks pitch in real time and nudges wandering notes back toward the intended target — but it distinguishes between a genuine pitch error and an intentional slide, scoop, or rasp, and leaves the latter alone.

The correction curve is adaptive, not a fixed snap-to-grid: a note that's slightly under target is nudged gently, while a note that's clearly off is corrected more firmly.

Listens For

  • Sustained pitch drift within a phrase
  • Intentional slides, bends, and scoops
  • Vibrato rate, so it isn't flattened
03 Harmonic Bloom
Presence without turning up the volume. +

Harmonic Bloom adds analog-style harmonic saturation — modeled after tape and tube coloration — to bring a vocal forward in a busy mix. Because it works on harmonic content rather than raw gain, a vocal can feel louder and closer without actually being louder on a meter.

The amount of saturation is scaled to the thinness of the original recording, so a already-warm vocal chain gets touched lightly.

Listens For

  • Thin or brittle upper midrange
  • Existing harmonic richness in the source
  • Transient sharpness on consonants
04 AutoLevel
Gain riding, word by word. +

AutoLevel replaces manual gain automation by riding level on a phrase-by-phrase basis, the way an engineer's hand would move a fader through a take. Quiet verses and belted choruses land at the same finished, consistent volume without audible pumping.

It tracks perceived loudness rather than raw amplitude, so sustained loud notes aren't punished the way a standard compressor would punish them.

Listens For

  • Phrase-to-phrase volume inconsistency
  • Perceived loudness vs. peak level
  • Breath and pause gaps, left untouched
05 DeEss Core
Sibilance tamed, top end intact. +

Rather than ducking a fixed frequency band whenever volume spikes — the classic de-esser approach that also dulls cymbals and air — DeEss Core identifies the specific harmonic signature of sibilance frequency by frequency, note by note.

Only the offending frequencies are pulled down, so brightness and "air" elsewhere in the vocal are left completely alone.

Listens For

  • S, Sh, T, and Ch sibilant peaks
  • Microphone-specific resonant harshness
  • The line between "air" and "harsh"
06 SpaceCraft
A room that matches the track. +

SpaceCraft analyzes the instrumental context around the vocal and builds a sense of space — early reflections, a short tail, subtle width — that's proportioned to match it, rather than applying one fixed reverb preset regardless of genre or arrangement.

A dry, tight vocal stays close when the track calls for intimacy, and opens up when the arrangement is bigger.

Listens For

  • Instrumental density and arrangement size
  • Existing room character in the source
  • Genre-typical vocal proximity

All six, running on one engine.

See Flux Engine V5