What happens to your track.

First, it listens.

Drop in a premaster and Acetate measures it before touching anything: integrated loudness (LUFS), true peak, dynamic range, spectral balance, noise floor, stereo correlation, and clipping - actual flat-topped samples, not just peaks near 0 dB.

It also flags specific problems: rumble that eats headroom, low-mid buildup that makes vocals sound muddy, harshness in the upper mids, and mixes that collapse when summed to mono.

Then it picks a target.

Your measurements get compared against a target curve - either a genre profile (techno, deep house, minimal and friends, each with its own sonic character) or the destination preset you picked: streaming, club, or critical listening. The gap between where your track sits and where it should be is what drives every decision downstream.

The chain responds to what it heard.

Nothing in the chain is fixed. Each stage has a trigger and a reason:

• Corrective EQ. Cuts what measured hot - mud in the low mids, harsh resonances, fizz up top - and lifts what measured thin. This is why muddy vocals come out clearer: the buildup swallowing them gets removed.
• Saturation. Adds gentle harmonics so the mix translates on laptops and phones, not just big speakers.
• Compression and density. Quiet details get lifted while transients stay intact, and loudness is added without crushing the dynamics. Character first - the limiter backs off rather than crush your transients to hit a number.
• Peak limiting. A true-peak ceiling keeps the master safe for streaming platforms.

After every master, the app shows you this reasoning for your specific track - what it found, and what each stage did about it.

A real example.

A vocal track comes in quiet with low-mid buildup swallowing the vocals. Acetate cuts the buildup, lifts presence so consonants read clearly, then adds density to bring the track up to target. The vocals did not get "enhanced" - the thing burying them got removed.

What it doesn't do.

Acetate masters; it doesn't mix. It can't un-mute a buried vocal, fix a kick that fights the bass, or repair a clipped source - when it detects one, it tells you to re-bounce with headroom instead of pretending. If the mix has a problem, the report card says so in plain words.