A track can seem better simply because it plays louder. Loudness measurements give you a reference when comparing exports and preparing a listening decision.

Three complementary measurements

Integrated LUFS describe overall loudness across a track. True peak estimates peaks, including those between samples. Loudness range describes level variation. Demorphus uses FFmpeg’s ebur128 scanner on the complete file. FFmpeg method ↗

Spotify normalization happens during playback

Spotify describes a normal target of −14 LUFS and playback gain compensation. This reference is not a promise of track acceptance. Behavior also varies with settings and devices. Spotify documentation ↗

Measure before changing the audio

Start with the original export and save its report. Then compare the processed version. If levels differ, adjust playback volume to judge tone and vocals more fairly. Moving a measurement closer to a target is not enough to conclude that the sound is better.

What the diagnostic cannot tell you

A high peak invites inspection; it does not prove audible distortion on its own. A narrow loudness range may be an artistic choice. Silence and signals below the scanner’s gate do not receive a usable LUFS reading.

This diagnostic does not master the track: it changes neither equalization, dynamics nor the file. It does not calculate an AI probability.

Prepared with AI assistance. Sources below.