Mixing with reference tracks: how to compare without fooling yourself
Choosing the right reference, matching loudness before you judge, what to listen for on each pass, and why copying another record's spectrum doesn't work.
What a reference is actually for
Mixing is hundreds of small decisions made with an ear that adapts to whatever you put in front of it. Half an hour in, your mix sounds normal to you — not because it's right, but because it's the only thing you've heard. A reference breaks that habit: it hands you an external starting point against which your decisions look like what they are.
What a reference is not: a destination. The comparison isn't there to make your song end up sounding like someone else's, it's to spot where you've drifted a long way without noticing. The difference shows up in the question you ask while comparing. «How do I get that low end?» leads to copying. «Why does mine weigh twice as much and read half as clearly?» leads to understanding.
Choosing the right reference
Most useless comparisons start with a bad choice. A reference works if it shares context with your song; if it doesn't, any difference you find is information about the genre, not about your mix.
| Criterion | Why it matters |
|---|---|
| Similar instrumentation | A mix with a string section divides the spectrum differently from guitar, bass and drums |
| Similar density | A six-track song and a sixty-track song can't share a balance or a low end |
| Well recorded, not just well mixed | If the reference starts from better takes than yours, you're comparing recording and believing you're comparing mixing |
| One you know by heart | With a track you know you spot the difference in two seconds; with a new one the song itself distracts you |
| Not too distant an era | The standard for brightness and low end has shifted a fair bit: a thirty-year-old reference takes you somewhere else |
How to compare without fooling yourself
- Match loudness first, always. Without it the rest of the method is worthless: louder always sounds better mixed.
- Compare quietly. That's where the real balance shows; at high volume everything sounds big and the differences get papered over.
- Listen for one thing per pass: low end first, then the vocal, then the width. Asking «what sounds different» in general always returns the same useless answer, which is «everything».
- Short passes. Ten or fifteen seconds of the same kind of section: chorus against chorus, never your verse against their chorus.
- Write it down and carry on mixing. Deciding while the reference is still playing is the fastest route to chasing it.
- Come back the next day before calling anything done: half of yesterday's tired-ear comparison usually turns out to be unnecessary.
Matching loudness by ear is enough for almost everything, but if you'd rather have a number than an impression, measuring the LUFS of both tracks settles it in a minute; what they mean is in the LUFS and loudness guide. And to check your impression against a measurement of how the spectrum is divided, the mix analyser does exactly that.
When the reference does harm
There's a point where comparing stops helping and starts damaging. You recognise it because the changes stop solving problems in your song and start moving it towards another one: you add brightness not because it's missing but because the reference has more; you squash the master not because the mix asks for it but because the other one is louder.
The honest test is simple: listen to your mix on its own, with nothing to compare it to, and ask whether you like it. If the answer is yes and you were still about to change something, that change came from the reference and not from the song. Save a version before you make it.
Much of what you learn by comparing is, underneath, ear training: recognising a zone of the spectrum without seeing it and spotting small differences. That trains separately, and the method is in the ear training guide. It's also worth knowing how much of what you hear is your room rather than your mix: the monitoring guide covers that.
Frequently asked questions
- How many references should I use?
- Two or three, chosen for different things: one for the vocal sound, one for low-end weight, one for width. With a single one you end up copying it wholesale, including decisions that don't fit your song. With ten you lose your bearings, because there's always one doing the opposite of the last and the comparison stops deciding anything.
- Do I have to match loudness before comparing?
- Yes, and it isn't a detail: it's the difference between comparing and not comparing. A mastered track will be 6 to 8 dB louder than your mix, and at higher volume the ear hears more lows, more highs and more definition. Without matching, the reference always wins and you learn nothing from the comparison. Turn the reference down until both sound equally loud — by ear is enough, no measurement needed.
- Is it useful to compare against a mastered track when my mix isn't mastered?
- It's useful for balance and tone, which is what you're after, as long as you've matched loudness. It's not useful for judging punch or density: the master put those there and your mix doesn't have them yet. If the comparison leads you to squash your mix to keep up, you've crossed the line where the reference stops helping.
- Can I copy the reference's frequency curve?
- You can look at it, but copying it doesn't work. That curve is the result of one specific arrangement with specific instruments: if your song has two guitars where the reference has a synth and strings, the same curve describes completely different mixes. The measurement is there to prompt questions — «why do I have so much at 200?» — not to hand you targets.
- How often should I go back to the reference?
- Less often than you'll want to. A couple of times per session, always in short ten or fifteen second passes. Comparing every five minutes makes the mix chase the reference instead of finding its own shape, and it tires your ear in exactly the range where you have to make decisions.
- What should I listen to references on?
- The same thing you mix on, on the same day. Changing speaker at the same time as changing song invalidates the comparison: you don't know which of the two changes you're hearing. What is worth doing is repeating the comparison on a second system — phone, car — because low-end and vocal problems show up there before they show up on any monitor.