How to Use Reference Tracks: Pick, Loudness-Match, Compare (2026)

The short answer: pick two or three finished songs in your genre with the sound you're after, get them as lossless files, and put them in your session where your master bus processing can't touch them. Then turn the references down until they match your mix's loudness, because a mastered song is much louder than a mix and louder always sounds better. Compare the same section (drop to drop, verse to verse) one element at a time: low end, vocal level, top end, width. Loudness-matching is the step most people skip, and it's the one that decides whether a reference helps you or misleads you.

I reference every mix and master I finish, on my own releases (Sony, Universal, Ultra, Dim Mak, 50M+ streams) and on client work, and my mastering intake asks every client for one to three references. This is the process I use, start to finish.

What to compare, and what to use

How to use a reference track, step by step

1. Pick two or three references, not ten. Same genre, similar arrangement density, and a sound you'd actually want your record to have. A song you love with a completely different instrumentation teaches you little. It helps to pick each one for a reason: one for the low end, one for the vocal.
2. Get lossless files. Buy the WAV or another lossless version. Lossy files and anything recorded off a stream can mislead you about the top end and the transients.
3. Route the reference around your master bus. Put it on its own track and send it straight to your interface outputs, or use a reference plugin placed last on the master bus, which switches between your mix and the reference after your processing. If your master bus EQ and limiter are also processing the reference, the comparison is broken.
4. Loudness-match. Measure the loudest section of both with a LUFS meter and turn the reference down until they read the same. A finished master can be 6 to 10 dB louder than an unmastered mix, and the louder one wins every comparison for reasons that have nothing to do with the mix. What LUFS is covers the measurement.
5. Compare the same section, briefly. Drop to drop, chorus to chorus. Switch every few seconds; after about ten seconds your ears adjust to whatever is playing and the difference fades.
6. Listen for one element per pass. One pass for the low end, one for the vocal level, one for the top, one for width. Trying to hear everything at once means hearing nothing specific.
7. Write the gap down, change the mix, compare again. "My kick is quieter than the reference's, my vocal is louder" is a note you can act on. Stop when the remaining differences are taste, not problems.

The same rule applies to any A/B, not only references. It's why the before/after player on my vocal chain page is loudness-matched by default: without that, the louder version wins and the test tells you nothing.

Mixing vs mastering: reference for different things

When you're mixing, reference the balance: the vocal level, the kick and bass relationship, how bright, how wide. Ignore loudness entirely and keep your mix peaking well below 0 dBFS; that's covered in how to mix a song and what headroom is.

When you're mastering, loudness becomes one of the things you compare, along with tone and dynamics. Even then, match by ear at equal loudness first, then decide how loud the master should actually be. How loud your master should be has the numbers by genre, and how to master a song covers the rest of the stage.

Reference plugins

You can do all of this with a track, a LUFS meter, and a mute button. Reference plugins make it faster: they sit last on your master bus, switch instantly between your mix and up to several references, and many of them level-match automatically. ADPTR Metric AB and Mastering The Mix REFERENCE are two widely used ones; check current versions, pricing, and DAW support before buying. Metric AB is compared with free and stock meters in the best metering plugins. Whatever you use, confirm the level-matching is on before you trust a single comparison.

The mistakes that make references hurt

• Comparing at different loudness. The most common one, and the reason people over-compress and over-brighten their mixes chasing a reference.
• Referencing your mix against a master and pushing the mix to match. A mix isn't supposed to be as loud or as dense as a master. Match the balance, not the loudness.
• Too many references. Ten references means ten different targets. Two or three agree on something; that's the thing to chase.
• Referencing late in a long session. Tired ears hear the top end wrong. Do your reference passes early, or after a break.
• Copying the reference's flaws. A reference shows you where your mix is; it doesn't make every choice on that record right for yours.

Sending references to an engineer

If someone else is mixing or mastering, references are the clearest brief you can give, as long as each one comes with a specific cue: "the low end of this one at 1:05," "the vocal level of this one in the chorus." One sentence per reference beats a paragraph about vibe. How to prep your mix for mastering covers the full handoff, and how to hire a mixing engineer covers what to send for a mix. My Stereo Mastering intake asks for one to three references for exactly this reason.

If you'd rather measure the gap than guess at it, that's part of why I built Masterproof. The free check reads your master the way a mastering engineer would: loudness in context, true peak, and the issues worth fixing before you distribute. Comparing your track against a reference is part of the paid tier.

Quick answers

How do you use a reference track? Put two or three lossless references in your session, routed around your master bus, turn them down to match your mix's loudness, and compare the same section one element at a time.

Why should reference tracks be loudness-matched? Because louder always sounds better at first. Without matching, the mastered reference wins every comparison and you end up over-compressing and over-brightening your mix to catch up.

How many reference tracks should I use? Two or three. Enough to see what they agree on, few enough to keep one clear target.

Can I use Spotify or YouTube as a reference? For a quick feel, yes, but not for detailed comparison. Buy the lossless file, so the top end and transients you compare against are real.