Skip to main content

Tips for Writing AI Mixing Prompts

AudioNova's AI mixer responds to natural language prompts to shape your sound. The better your prompt, the better your mix.

🧠 How Prompts Work​

You type mixing instructions like:

Warm analog sound with soft reverb and tight low end:

The AI interprets and applies processing based on cues like:

  • 🎚️ EQ
  • 🧵 Compression
  • 🌫 Reverb / Delay
  • 🎛 Stereo width
  • 🎧 Genre & vibe

✅ Good Prompt Examples​

PromptEffect
Clean pop vocal with light reverb and subtle compressionPolished vocal chain
Deep house kick with boosted low mids, no reverbFocused kick tone
Lo-fi guitar with tape saturation and mono widthVintage texture
Bright EDM drop with sidechained bassModern loudness & movement

❌ Common Mistakes​

  • Too vague
    Make it better → instead try:
    More aggressive drums and wider stereo image

  • Too long
    Break it up!
    ❌ I want soft vocals with airy highs, low mids cut, and stereo enhancement with no reverb
    ✅

    • Soften vocals and boost highs
    • Cut low mids
    • Widen the stereo

  • Only genre
    Make it techno is okay, but better:
    Punchy kick with dark atmosphere and minimal reverb

✍️ Prompt Formula​

Use this structure:

[Tone or Genre] + [Processing] + [Optional Reference]

Examples:

  • Warm acoustic mix like Bon Iver
  • Trap vocal with autotune and slap delay
  • Ambient pads with shimmer reverb and slow attack

💡 Bonus Tips​

  • Iterate:
    → Add brightness to vocals
    → Now compress slightly

  • Use musical or emotional language:

    • Dreamy
    • Harsh and crunchy
    • Club-ready

  • Think like a producer, not an engineer. Speak to the AI like it’s a collaborator. 🎛🧠

The more musical your language, the more musical your mix.

🧠 How to Speak to the AI — Prompt Grammar for AudioNova​

AudioNova is powered by a language model, not a traditional knob-based plugin — so how you phrase your request matters.

Here are some special syntax tricks and prompt structures that can help guide the AI more precisely:

1. Use Clear Signal Words​

Instead of vague ideas like "make it cool," use effect-oriented words:

  • ✅ brighten the vocals → triggers EQ boost in high frequencies
  • ✅ tighten the kick → invokes transient shaping or compression
  • ✅ add atmosphere → usually means reverb or stereo widening

2. Use Brackets for Instructions​

You can include extra context inside square brackets that helps the AI understand your intent, without it being processed as literal text:

Give the drums a gritty texture [like old vinyl or cassette]

Add shimmer reverb to the synth [emulate Valhalla-style tone]

This helps the AI narrow the stylistic reference or aesthetic.

3. Use Parentheses for Variations or Emphasis​

You can use () to repeat or emphasize traits or layers in your mix:

Add deep, booming low end (sub-heavy)

Delay the vocals slightly (1/4 note) for movement

The model sometimes interprets these as secondary weighting — especially useful when layering.

4. Chaining with Arrows →​

You can guide sequences step by step:

Compress the vocals → brighten the top end → add slap delay

This helps the AI follow a logical mixing order.

5. Role-Based Prompting​

Try speaking to the AI like you're instructing a real engineer:

You're the mastering engineer. Make this sound polished and commercial-ready.

Or:

Mix this as if it were for an ambient film score.

Framing the role helps condition the AI toward your goal.

6. Comparisons Work!​

Make the drums hit like Travis Scott production

Vocal tone similar to Billie Eilish — intimate and breathy

Names, styles, or analogies are often well-understood by the model.

💡 Final Thought​

You're not turning knobs — you're speaking in ideas. The clearer your intention, the better the AI will translate it into sound.

This is not just prompting — it's the new language of mixing. 🧠🎚✨