Loudness Matching
Loudness matching means comparing two versions at the same perceived volume. It is one of the simplest ways to make better mastering decisions.
Without loudness matching, the louder version usually wins the first impression. That can be useful when checking final delivery level, but it is a poor way to judge whether EQ, compression, stereo width, saturation, or limiting actually improved the track.
Why Louder Feels Better
When two similar sounds are played back at different levels, the louder one often seems clearer, brighter, wider, and more exciting. This happens even when the only difference is gain.
That bias is especially dangerous in mastering because many processors change loudness as a side effect:
| Processor | How it can bias comparison |
|---|---|
| EQ | Boosts can make the processed version louder and more impressive. |
| Compression | Average level rises even if peaks are controlled. |
| Saturation | Added harmonics can increase perceived loudness. |
| Limiting | Peaks are reduced, then makeup level makes the master denser. |
| Stereo widening | A wider image can feel larger and more exciting even when it weakens the centered, mono-safe core of the mix. |
Matching loudness does not make the comparison perfect, but it removes the most obvious trap: choosing the louder version because it is louder.
How To Use It
Use loudness matching while making creative or technical judgments:
- Render the processed version.
- Enable loudness-matched A/B playback.
- Switch between before and after on the same musical phrase.
- Listen for tone, punch, vocal stability, low-end control, stereo image, and artifacts.
- Turn matching off only when checking the actual final delivery level.
If the processed version only sounds better when matching is off, the processing may mainly be a gain increase. In that case, reduce the effect and compare again.
What It Does Not Tell You
Loudness matching is not a delivery target. A file can compare fairly in A/B mode and still be too loud, too quiet, or unsafe for release.
It also does not replace metering. Use LUFS (Loudness Units relative to Full Scale, the broadcast/streaming unit for perceived loudness) and true-peak readings to check delivery constraints, then use matched listening to decide whether the processing is musically useful.
In The libsonare Demo
The mastering demo enables Match loudness for A/B playback by default. Whichever version you switch to is attenuated down to the quietest of the versions you have loaded, so nothing in the comparison is favoured by raw level. With a reference track loaded, that anchor can be the reference, in which case both the source and the rendered master play back attenuated.
The exported audio is not changed by this playback setting. The WAV download keeps the rendered level. Loudness matching only affects how the browser player compares the two versions.
Implementation notes
The demo takes the integrated LUFS of the source, the rendered master, and the reference track when one is loaded, and uses the lowest of the three as an anchor. Whichever side is playing gets a playback-only gain offset that brings it down to that anchor; the offset is never positive, so matching only ever attenuates.
This avoids modifying either buffer and keeps the exported WAV authoritative. It also means any report or CLI/API reproduction should use the rendered LUFS values, not the temporary comparison gain used by the browser player.
Related: A/B comparison, LUFS, What Is Mastering?