How Vanta works
Pick a service, drop your audio, and Vanta's engine does the rest — analysing every stem, choosing the right processing, and narrating what it's doing as it goes. Here's exactly what happens in each mode.
Preparing your files — for the best mix
A great mix starts with clean stems. Two minutes of prep makes a real difference:
- Export separate stems — one file per instrument or group (drums, bass, vocals, synths…), not a single mixdown. (For Master or Remix a Song, a single finished track is exactly right.)
- Same start & length — bounce every stem from bar 1 / time zero to the same end point, so they line up perfectly when stacked.
- Leave headroom — aim for peaks around −6 dBFS and never let a stem clip (hit 0 dB). Take any limiter or compressor off your master bus before exporting — Vanta needs clean, un-squashed stems to balance properly.
- Name them clearly — kick.wav bass.wav lead_vocal.wav pad.wav. Vanta reads the filenames to auto-detect each instrument (and you can still correct any of them in the UI).
- Use a lossless format — WAV or AIFF, 24-bit, and the same sample rate for every stem (44.1 or 48 kHz). Avoid MP3 for mixing — it's lossy.
- Keep stems mostly dry — print a signature creative effect if it defines the sound (a special delay or reverb), but leave heavy processing off so Vanta can shape and balance from a clean starting point.
- Mono or stereo, as recorded — export mono sources as mono, and stereo parts (wide synths, drum overheads) as stereo.
The engine — every stem-based mode
For each stem you upload, Vanta runs the same pipeline:
- Analyse — loudness, transients (crest), spectral shape and band energy.
- Detect the instrument — vocal, drums, bass, synth, guitar… which selects that stem's effect chain.
- Apply the chain — e.g. vocals get de-ess → presence EQ → compression → reverb; drums get EQ → saturation → parallel compression; bass gets a high-pass → EQ → compression → drive.
- Balance & pan — loudness-matches each stem to a musical target and places it in the stereo field.
- Glue — light bus compression pulls the whole mix together.
- Master (mastering modes) — tonal balance, analog saturation, mid/side EQ, multiband compression, stereo width and true-peak-safe loudness.
The live narration shows this as it happens — “Compressing the drums”, “Top end's hot — easing 4k for clarity”.
The five modes
Mix
🎚 Mix
In — your stemsOut — a balanced mix (no master)
Upload your multitrack stems. Vanta runs steps 1–5 above and returns a mixed stereo track — balanced, panned and glued, but left unmastered so you keep headroom to master it yourself.
Master
✨ Master
In — one finished stereo trackOut — a release-loud master
Upload a single finished mix. Vanta skips the stem stage and runs the mastering chain only — glue, tonal balance, saturation, multiband and loudness with true-peak safety — returning a mastered track at your target loudness.
Full pipeline
⭐ Mix + Master
In — your stemsOut — a finished master
Upload your stems and Vanta runs the whole engine — mixing the stems and mastering the result into a release-ready track. This is the flagship path.
Vocals
🎙️ Vocal Mix
In — your stemsOut — polished vocals only
Upload your stems. Vanta keeps only the vocal stems (lead + backing), then mixes and masters just those into a clean, upfront vocal bus — ideal for topline and a cappella work.
New
🎤 Remix a Song
In — one finished songOut — freshly mixed & mastered
No stems? No problem. Upload one finished track and Vanta splits it into stems (drums, bass, vocals, other) with AI stem separation, then runs the full mix + master pipeline. The separated stems are cached, so tweaking and re-rendering is fast.
At a glance
| Mode | You upload | You get |
|---|
| Mix | Stems | Balanced mix (unmastered) |
| Master | One finished track | Mastered track |
| Mix + Master | Stems | Release-ready master |
| Vocal Mix | Stems | Polished vocals only |
| Remix a Song | One finished song | Split → mixed & mastered |
After processing — make it yours
Every result can be revised. Use the plain-language guided-revision chips per instrument (“too muddy”, “too loud”, “not punchy”) and Vanta translates them into the right DSP moves — or dive into the manual controls (instrument, gain, harmonies, auto-tune, saturation character and more) and re-render. Your plan sets how many revisions each track gets.