How to A/B Compare Mix Versions Without Fooling Yourself

Comparing two mixes and picking the better one sounds simple. It isn't, because your ears are easy to fool. The mix that's even half a dB louder will almost always sound "better" to you, even if it's objectively worse. If you want an honest A/B comparison, you need to match loudness first, limit how many options you're judging at once, and control the conditions you're listening under.

Here's the short version: match both versions to the same perceived loudness (within about 0.3-0.5 dB), listen blind or at least avoid knowing which file is which while you judge, compare only two versions at a time, and don't trust a decision made after more than 20-30 minutes of critical listening. Everything below explains why that works and how to do it in practice.

Why loudness wrecks your judgment

This isn't a willpower problem. It's basic psychoacoustics. A slightly louder mix sounds like it has more punch, more low end, more clarity — even when the only thing that changed is the fader level. This is called the Fletcher-Munson effect in casual studio talk: our ears perceive frequency balance differently at different volumes, and louder almost always reads as "better produced."

This matters a lot when you're comparing:

• Two mix revisions from an engineer
• A mix before and after a plugin change
• Your mix against a reference track
• A mastered version against the pre-master

In every one of those cases, if the levels aren't matched, you're not comparing mixes. You're comparing volumes.

Step 1: Match loudness before you judge anything

Before you listen to two versions back to back, get them to roughly the same integrated loudness. You don't need studio-grade precision, but get close.

Simple ways to do this without specialized gear:

• Use a loudness meter plugin (many DAWs include one, and free options exist) to check LUFS on each file, then trim the quieter one up or the louder one down with a gain plugin until they're within half a dB.
• If you don't have a meter, use your ears with pink noise or a familiar reference track as a loudness anchor, switching between the two mixes at matched perceived volume using your monitor controller, not the DAW fader (small moves on a monitor knob are easier to judge consistently).
• When comparing a mix to a master, remember mastering almost always adds loudness. Turn the master down to match the mix's loudness before deciding whether the tonal and dynamic changes are actually better.

Skipping this step is the single biggest reason people pick the wrong version in a revision round.

Step 2: Compare two at a time, never five

If you've got four mix revisions sitting in a folder, don't try to rank them all in one sitting. Ears fatigue fast, and holding more than two options in short-term auditory memory is genuinely hard.

Instead:

1. Pick the two most likely candidates.
2. Loudness-match them.
3. A/B them on a specific section — usually the chorus or the densest part of the song — switching every 10-15 seconds.
4. Pick a winner.
5. If there's a third contender, compare it against that winner, not against the original two together.

This is a simple elimination bracket, and it keeps every comparison binary, which your brain handles far better than ranking a pile of options at once.

Step 3: Blind yourself when you can

Knowing which version is the "new" one biases you — you'll want to hear improvement, so you often will, whether it's there or not. A few low-effort ways to blind the test:

• Have a bandmate or collaborator load both versions and label them "A" and "B" without telling you which is which, then reveal the answer after you've chosen.
• Rename files to meaningless labels (session1, session2) instead of "v3_FINAL" or "client_notes_applied" before you sit down to compare.
• If you're comparing your own work, wait a day between making a change and judging it. Fresh ears the next morning are more reliable than tired, biased ears five minutes after a tweak.

Step 4: Control the listening environment

The comparison is only as good as the system you're listening on. Keep these consistent across both versions:

• Same speakers or headphones for both files, not monitors for one and earbuds for the other.
• Same playback volume, set by ear-matched loudness, not by eyeballing a fader.
• Same room position — don't lean in for one and sit back for the other.
• Same time of day if possible. Ears genuinely behave differently after hours of listening fatigue versus first thing in the morning.

If the decision matters (final mix approval, choosing a master), repeat the comparison on at least two systems — studio monitors and one consumer reference like earbuds or a car stereo — since a mix that wins on one system might lose on another.

Step 5: Know when to stop listening

Critical listening has a shelf life. After 20-30 minutes of focused A/B work, your perception of "better" starts drifting toward "whatever I've heard most recently" rather than any real quality judgment. If you're still undecided after that window, stop, take a 15-minute break away from speakers, and come back. Don't force a ninth comparison pass in one sitting — you'll just end up flip-flopping.

A quick pre-comparison checklist

Common mistakes

• Judging a loud master against a quiet pre-master and assuming the master is simply better-sounding.
• Comparing a mix on headphones against a reference heard on monitors last week, from memory.
• Letting a client or collaborator know which file is "the new one" before they listen.
• Trying to compare three or more revisions in one long session instead of eliminating in pairs.
• Deciding at 1am after six hours of mixing, when every version starts sounding the same anyway.

How this fits into a mix revision workflow

Once you're doing proper A/B comparisons, you also need a clean way to keep versions straight so you're not accidentally comparing an old bounce against a new one. In Waves, every mix revision uploaded to a song stays organized with the rest of the project's files, and stems or mixes play back in sync in the browser so you can switch between versions on the same system without re-importing anything. Timestamped comments make it easy to note exactly where in a section a version won the comparison, so that feedback doesn't get lost by the next revision round.