LUFS for TikTok
Published · updated · by Loopin
Short-form audio is almost always heard on a phone speaker, in a room with other noise, at partial volume. That changes what a good master sounds like more than the platform does.
Target around −12 to −14 LUFS
Short-form platforms normalise, so a hyper-loud upload gains nothing. Landing between −12 and −14 LUFS integrated puts you at parity with everything else in the feed while keeping the transients that make a hook land.
What matters more is that your clip is not noticeably quieter than the clips around it. Being quieter than the feed is the one loudness mistake that genuinely costs you attention, because people do not turn a phone up for a stranger.
The mids carry everything on a phone speaker
A phone speaker produces almost nothing below about 500 Hz. Whatever makes your hook recognisable has to exist between roughly 500 Hz and 5 kHz, or on short-form it does not exist at all.
Check by playing the clip on your own phone at half volume with a fan or a tap running. If the vocal or the lead melody disappears, that is what a viewer hears — and no amount of mastering loudness fixes a hook that lives entirely in the bass.
Do not master a separate louder version
People bounce a squashed version specifically for short-form. It does not help, because the platform normalises anyway, and it means the eight seconds people hear are not the eight seconds on the streaming release — which is a strange thing to discover after somebody looks the song up.
Master once, properly, and use the same file for the clip. Consistency between the snippet and the record is worth more than a decibel nobody keeps.
Ceiling at −1 dBTP, same as everywhere
Short-form audio is re-encoded, often more than once, and on some pipelines fairly aggressively. A file at 0 dBFS routinely comes back distorted, which on a phone speaker sounds like a bad recording rather than like a loud one.
One decibel of headroom prevents it and costs nothing perceptually. It is the same true peak rule that applies to every lossy delivery.
Check the clip, not the song
The eight seconds you post have their own loudness, and a chorus pulled out of a dynamic track can measure very differently from the track as a whole. A clip taken from a quiet verse will sit noticeably below the feed.
Pick the clip from a loud section, and if you must use a quiet one, accept that it needs to earn attention visually instead. That is a content decision, not a mastering one.
Frequently asked questions
What LUFS should I use for TikTok?
Around -12 to -14 LUFS integrated. The platform normalises, so louder gains nothing — but being noticeably quieter than the feed does cost you, because nobody turns a phone up for a stranger.
Should I make a louder version for short-form?
No. It gets normalised anyway, and it means the snippet people hear is not the record they find afterwards. Master once and use the same file.
Why does my song sound thin on TikTok?
Phone speakers produce almost nothing below 500 Hz. If your hook lives in the bass it does not exist on short-form. Test on your own phone at half volume with background noise.