Your podcast is mastered to the ceiling, peaks at 0 dBFS, and still sounds limp next to everything else on Spotify. The problem is not peak level. It is loudness — and the platform is going to change yours anyway.
dBFS is not loudness
dBFS measures sample peaks, not perceived level. RMS gives an average but no frequency weighting. LUFS (Loudness Units Full Scale), defined in ITU-R BS.1770, filters the signal with a K-weighting curve that approximates human hearing, then integrates over the whole program with gating, so silence does not drag the average down. One LU equals 1 dB; 0 LUFS is full scale, which is why normal values are all negative.
| Measure | Answers | Typical target |
|---|---|---|
| Peak (dBFS / dBTP) | highest sample / true peak | below -1 dBTP for lossy delivery |
| RMS | average power, unweighted | no universal standard |
| Integrated LUFS | perceived loudness of the whole program | -14 streaming, -16 Apple Music, -23 broadcast |
| Short-term LUFS | loudness over roughly 3 seconds | checking whether a chorus or climax overloads |
What platforms do to your master
| Platform | Normalization target |
|---|---|
| Spotify | -14 LUFS |
| YouTube | -14 LUFS |
| Apple Music | -16 LUFS |
| Tidal | -14 LUFS |
| European broadcast (EBU R128) | -23 LUFS |
| US broadcast (ATSC A/85) | -24 LKFS |
LKFS and LUFS are effectively the same unit. Master at -8 LUFS and Spotify turns you down about 6 dB: you win no loudness, while the limiter artifacts stay exactly where they were. The loudness war is a straight loss in a normalized world.
Master for the target, not the meter
For streaming, aim for -14 LUFS integrated with true peak at -1 dBTP, or -2 dBTP if you are limiting hard. Judge the integrated value over the full track, not a short-term reading. A quiet track might get lifted a little; a hot one is guaranteed to come down, so keeping dynamic range costs you nothing. Measure LUFS and true peak before uploading — watching the waveform hit the top of the screen tells you almost nothing.