Play these two finished songs. Don't analyse them. Notice: after a few seconds of each, what are you following? If someone stopped the music and asked you to hum it back, what would you hum?
In each song, which part are you following? Is it the loudest thing, or something else?
Where does your attention go?
RenderedMatched by integrated loudness over each whole clip: louder versions are turned down to the quietest one.
You probably followed one thing in each: a voice, a hook, a groove. Now listen again and count everything else that's playing. There's a lot of it, and you barely noticed most of it. That isn't an accident. Someone decided what you'd follow, and turned everything else into the room around it.
That's what mixing is for. Not making every track sound amazing. Deciding what the listener notices, and when.
Why the rough demo sounds better
Almost every producer has lived this: the rough bounce you made in ten minutes feels alive, and the mix you spent a week on feels worse. Studios even have a name for it, demo-itis. Bruce Springsteen's Nebraska is the famous extreme: he recorded the songs alone on a four-track cassette as demos, tried them with a band, and in the end released the cassette.
Why does it happen? Usually because the rough balance was made in one go, listening to the whole song, with one thing in mind. The "proper" mix was made track by track, in solo, making each part as good as it could be on its own. Every part got bigger, brighter and more exciting. And when they all came back together, they were all shouting for your attention at once.
A part can be great alone and wrong together
Solo is a wonderful tool for finding a click, a noisy edit or a ringing resonance. But it hides the one thing you're actually mixing: the relationships.
A guitar sounds gorgeous and full on its own, then covers the singer's words. A bass with a bright, growly edge sounds exciting, then competes with the lead. A dry, plain snare sounds boring alone, and in context gives the groove exactly the snap it needs.
So the question to ask of every part isn't "does this sound good?" It's "is this doing its job, with everything else playing?" Its job might be to lead, to carry the groove, to add weight, to fill the harmony, or to make the space feel bigger. Different jobs need different sounds.
Start with one sentence
Before you touch a fader, write one sentence:
"In this section, I want the listener to follow ___."
The lyric. The bass groove. The conversation between the drums and the synth. A slowly changing texture. Then every move has something to serve.
The answer often changes by section. The vocal leads the verse. The guitar takes over for the solo. A drum fill steals the spotlight for one bar before the chorus.
An attention map
That's an attention map. It doesn't need to be drawn. But if you can't say who leads each section, the listener won't be able to tell either.
Faders first
The quickest experiment in mixing is still a level move. Pull the competing part down a couple of dB, with the whole section playing. Does the thing you want to follow come forward? Does the section lose energy? Push it back and listen again.
Loudness isn't the same as attention, though. A quiet, unexpected sound can grab your ear instantly. A loud bass can sit underneath everything and still feel like support. Timbre, register, rhythm, lyrics and surprise all pull focus. But a small fader move is so fast to try that it should almost always come before a plugin.
A good order to build a balance: start with the part you want followed, bring in the groove and low end around it, then the harmony, then the decoration. If some part needs a strange level to work, that's a clue its sound or its arrangement is fighting the others.
Pan opens a second dimension
Pan moves a part left or right. Two guitars playing similar parts, one each side, stop tripping over each other. A shaker drifting off to one side gives the centre more room for the voice. It's like seating a band on a stage.
Most mixes keep the lowest and most important parts (kick, bass, lead vocal) in the middle, where they're strongest on every system, and spread supporting parts around them. That's a habit, not a law. Some of the best-loved records of the 1960s put the drums hard on one side.
Always check in mono once. Plenty of phones and smart speakers play mono, and if your separation was holding up the clarity, you'll hear it collapse.
Your volume knob changes the mix
Here's the same song at three playback levels. Switch between them and listen for which parts survive.
As the song gets quieter, which part disappears first? Which one survives best?
The mix at three volumes
RenderedNot matched: each version plays at its own level. On purpose, 12 dB quieter is set −12.0 dB and 24 dB quieter is set −24.0 dB.
As it gets quieter, the bass and the weight of the kick fade away first, while the vocal holds on. That's the equal-loudness effect from Sound: at low volume, your hearing loses the lows (and the very highs) first.
So a balance made very loud will sound thin when played quietly, and a balance made very quiet will sound boomy when played loud. Mix at a comfortable, moderate level, and check at one quieter level. The question at low volume isn't "does it sound the same?" It's "can I still follow what I'm supposed to follow?"
When the mixer isn't the answer
Some problems begin before the mix. Two busy parts in the same register may need an arrangement change, not EQ: move one up an octave, drop a few notes, leave a gap. A pad with a long release may need a shorter envelope. A clashing take may need a different take. Recognising that a problem isn't a mixing problem is part of mixing well.
Same idea, different place: Form and arrangement decides when parts enter and leave. A good arrangement gives the mix clear relationships to show off; a good balance makes the section changes land.
Make it yours
Open a project of your own and pick eight bars. Write your one sentence: "Here, the listener should follow ___."
Now try this: bypass every optional effect, and rebuild the balance using only faders and pan, with the whole section playing. Listen at your normal level, then quieter, then in mono. Can you follow your chosen part in all three?
Compare it to the version with all the processing. Be honest about which one tells the story better. Save the faders-only version alongside your sentence, even if you go back to the processed one. It's the version that shows you what the song actually needs.
Quick check
- You EQ a guitar in solo until it sounds massive. Back in the mix, the vocal is harder to understand. What went wrong? The guitar got better at a job it didn't have. "Massive" in solo often means it now fills the space where the vocal lives. Judge parts in context.
- Your mix sounds great loud, but when you play it quietly for a friend the bass line disappears. Should you turn the bass up? Not straight away. Quiet playback always loses low end first. Check whether you can still follow the line, and see whether the bass has enough upper detail to survive. Mix at a moderate level so you don't overcorrect either way.
- In the chorus, a synth hook is much quieter than the vocal, yet it's what everyone remembers. How? Attention isn't just level. A catchy melody, a distinctive sound or a well-placed gap can pull focus even at a low fader.
Next steps
- Explore the chapter: Effects. Explore individual processors once you can name the change the mix needs.
- Go deeper: Gain staging. You pushed every fader up and the master started clipping, though no single track did. Where did the room go?
- Explanation: Listening and comparing. Make sure the version you prefer isn't just the louder one.









