PRESS·HIT
Prompt craft · Guide

Describing a mix in a prompt

Updated May 21, 2026

Two prompts with identical instruments can come out sounding like different songs — the difference is the mix. Mix language tells the model how the parts sit together, and most people skip it entirely, inheriting a loud, bright, modern default. A few well-chosen words take that back.

Warmth versus brightness

This is the single biggest tonal lever. "Warm, mid-forward, treble rolled off above 8 kHz" reads as intimate and vintage; "bright, airy, crisp high end" reads as modern and polished. Naming tape or vinyl character ("tape-saturated", "vinyl crackle") pushes warm even further without you ever saying the genre.

Space and depth

Dry or reverberant? Close or roomy? "Close-miked, dry, headphone-focused" sounds present and intimate; "plate reverb, roomy, elements set back" sounds spacious and cinematic. You can also place elements front-to-back: "vocal upfront, pads sitting behind".

The low end

Say what the bass should do: "clean sub, rounded kick" versus "punchy, aggressive low end". For club material, "sub clean and centred, kick prominent" keeps the bottom usable on a system. Left unspecified, the low end is where generators most often get muddy.

Presence and balance

Decide what's loud. "Vocals upfront" versus "vocals tucked into the mix" changes the whole feel of a track. Scooped mids read as scooped and aggressive; mid-prominent reads as warm and full. These balance words quietly do a lot of the emotional work.

Roll your own

2,300+ deterministic prompts. Free to try.

Pick your decisions — mood, groove, instruments, energy — and the engine fills the rest in under a second. No card needed.

Try the engine →

Frequently asked

Do mix terms actually change the output?

Yes — mix language is one of the strongest signals a model reads. "Warm, mid-forward, treble rolled off" and "bright, airy, crisp" applied to the same instruments produce noticeably different results.

What's a safe default mix to ask for?

If you're unsure, "warm, balanced, vocals present, clean low end" is a forgiving starting point that works across most styles. Adjust toward warm or bright once you hear it.