In a bed under speech
Most mixing decisions for video are really about keeping the 1–4 kHz speech band clear.
Read more about PodcastsWhat Is Mixing in Music?
Balancing the parts of a song against each other — levels, panning, EQ, effects — so everything is audible and sits in its own space. Mixing decides whether a good arrangement actually reads.
In short
Mixing is the stage where the separate parts of a song are balanced against each other — levels, panning, EQ and effects — so that every element is audible and occupies its own space in the stereo field and the frequency spectrum.
- Works on
- Individual parts, before mastering
- Three dimensions
- Level · frequency · stereo position
- Core skill
- Cutting what competes, not boosting
- Check on
- The worst speakers you own
A mix is mostly a problem of space, and there are only three kinds of space to work with. Loudness decides what the ear notices first. Frequency decides whether two parts fight: a bass guitar and a kick drum occupying the same 80 Hz will mask each other no matter how the faders sit. Stereo position decides how wide the record feels and gives you somewhere to put things that would otherwise collide.
Most of the craft is subtraction. Beginners reach for a boost when the vocal is unclear; experienced engineers usually cut whatever is competing with it. That is why a good mix often has less going on than you would expect — every element that stays has to earn the room it takes.
The reference point matters more than any plugin. Mixes drift, and after two hours your ears have adapted to whatever you have been doing. Comparing against a commercial record in the same genre, at the same loudness, resets that. Checking on the worst speakers you own resets it further: if it works on a phone, it works.
For AI-generated music, mixing decisions are made before generation rather than after. You cannot rebalance a stereo bounce, so a prompt that specifies an uncluttered midrange for a voiceover, or a prominent low end for a workout track, is doing the work a mix engineer would otherwise do later. Generate stems and the usual mixing options come back.
Music and AI terms related to mixing
Where mixing decides something
Three places mixing decides something.
In the browser
Level and dynamics are the two mix decisions you can make without a DAW.
Read more about Volume changer
In a genre
How dense a mix can get before it collapses is genre-specific — metal and lo-fi sit at opposite ends.
Read more about MetalExplore
More music and AI terms explained
The terms that carry enough confusion to need a full page.
Mixing — questions people ask
What is mixing in music production?
Balancing the recorded parts against each other — levels, panning, EQ, compression and effects — so the arrangement reads clearly as one piece of music.
What is the difference between mixing and mastering?
Mixing balances the parts within a song; mastering prepares the finished mix for release across different playback systems. Mixing is many faders, mastering is one stereo file.
How do I make a mix sound professional?
Usually by removing things rather than adding them. Amateur mixes are too busy — every part competing for the same frequency range — and carving space is what makes each element audible.
What frequency range does the human voice need?
Intelligibility lives roughly between 1 and 4 kHz. A music bed with its energy concentrated there will fight a voiceover no matter how far you turn it down.
Do I need to mix AI-generated music?
Usually not — it arrives mixed. What it often needs is level matching against the rest of your project and a little EQ if it sits under speech.
What is panning for?
Placing sounds across the stereo field so they occupy different space. Double-tracked guitars panned hard left and right is why a rock mix sounds wide.
Put the vocabulary to work
Describe the track you want and hear it in under a minute.