Why louder sounds better, and how to A/B fairly

Every engineer knows the feeling: you nudge a fader or a limiter, it gets a touch louder, and suddenly it sounds better. Not just louder, better. The trap is that "better" arrived with the level, not with the change you made. Learn to see past it and your decisions get more honest overnight, whether you're picking a plugin setting, choosing between two mixes, or getting a client to sign off on a master.

Why louder sounds better

This isn't imagination, it's how hearing works. Our ears are not equally sensitive across the frequency range, and that sensitivity changes with level. The equal-loudness contours (the Fletcher-Munson curves) show that as things get louder, we perceive relatively more bass and treble. So a louder version sounds fuller at the bottom and brighter on top, and more of its low-level detail rises above the room and the noise floor. The result is a version that sounds richer and more finished, instantly, with nothing else changed. A difference of about a decibel is enough to tilt a quick comparison, and around three reads as clearly better on first impression.

Where the bias costs you

It's not a curiosity, it's a decision-corrupter, and it shows up everywhere:

• Comparing processing. A plugin that adds a little output level seems to "improve" the sound. Bypass it and the track seems to get worse, when all that changed was volume.
• Choosing between mixes or masters. Play two versions back to back at different levels and you'll pick the louder one, then rationalize why.
• Chasing loudness. The loudness war was largely this bias at scale: hotter masters won A/B comparisons on the radio and in the shop, so everyone pushed louder.
• Client approval. Send a client a version and a slightly hotter revision and they'll approve the hotter one for a reason that has nothing to do with the music. That's the single most avoidable way a revision gets approved wrongly.

The fix: gain-match before you judge

The cure is simple to state and easy to skip: match the loudness of the two things you're comparing before you decide anything. Then the only difference you hear is the work, not the level.

• Comparing a plugin, in vs out. Trim the output so bypassed and engaged are the same loudness, then judge. Many plugins have an output gain for exactly this.
• Comparing versions. Measure the loudness of each and bring the louder one down to the quieter. Attenuate, never boost, so nothing clips.
• Against a reference track. Match your master's loudness to the reference before you compare tone and dynamics, or the reference's level does the deciding.

By ear, trim until they sound equally loud. For a decision that matters, use a meter: the free loudness meter reads integrated LUFS and true peak, and getting two versions within about a decibel is close enough that the level stops voting.

Level-matched A/B, without touching the audio

Matching loudness for a comparison should not change the file. The clean way is to attenuate the louder version's playback level toward the quieter one, so both play at the same loudness while the audio stays exactly as rendered. Drop two files into the free loudness-matched A/B tool and it measures both and plays them level-matched in your browser, with a blind mode so your own expectation stops voting too. It's the same idea as gain-matching a plugin, applied to whole masters.

Loudness is still a creative choice

Gain-matching is for judging, not for the final level. Once you've decided which version is better on its own merit, you set the release loudness for the song and the genre, which is a creative call, not a number to hit. Streaming platforms normalize playback, so a reference like -14 LUFS is a reference and not a ceiling, and a denser or more dynamic master can each be the right answer. The one hard line is true peak: keep it under -1 dBTP (-2 for Amazon) whatever loudness you land on. If the metrics are new, this guide explains LUFS and true peak.

When the client is the one comparing

The bias is at its worst when the person comparing isn't an engineer. A client flipping between two masters at different levels will almost always prefer the louder one, and then it's approved. The fix is to let them compare level-matched, so their approval is about the master and not the volume. That's exactly what Fair Loudness does inside a review link, and it's covered in how to get client sign-off on a master.

The short version

Louder sounds better because our ears hear more bass and treble as level rises, and about a decibel is enough to swing a comparison. So match loudness before you judge anything: trim a plugin's output, or attenuate the louder version down to the quieter, and only then decide on merit. Set the release loudness afterward, as a creative choice, with true peak under the ceiling.

Frequently asked questions

Why does louder music sound better?

Partly physics. At higher listening levels our ears perceive more bass and treble (the equal-loudness contours), so a louder version sounds fuller and more detailed even when nothing else changed. The effect is real and immediate, which is why a small level difference can decide a comparison before you've judged anything about the actual sound.

What is gain matching?

Gain matching means setting two things you're comparing to the same perceived loudness before you decide which is better, so the only difference you hear is the processing or the mix, not the volume. Engineers gain-match a plugin's input and output before judging it, and match two mixes or masters to the same loudness before choosing between them.

How do I A/B two mixes or masters fairly?

Match their loudness first. Measure the integrated LUFS of each and attenuate the louder one down to the quieter one (attenuate, don't boost, so nothing clips), then compare. By ear, trim until they sound equally loud; for a precise match, use a LUFS meter and get them within about a dB. Only after they're matched does the comparison reflect the master rather than the level.

How much louder does something have to be to sound better?

Not much. A difference of around 1 dB is enough to be noticeable and to tilt a quick A/B, and around 3 dB reads as clearly louder and usually 'better' on first impression. That's why matching within roughly a dB matters when you're making a decision on merit.

Does gain matching mean my master should be quiet?

No. Gain matching is only for judging, not for the final level. Once you've decided which version is better on its own merit, you set the release loudness for the song and genre. Loudness is a creative choice; keep true peak under the ceiling (around -1 dBTP, -2 for Amazon), and remember that streaming normalizes playback so a reference like -14 LUFS is a reference, not a target to chase.