Why is my Suno song so quiet? What 248 measured exports say

October 5, 2026 · All posts

"My Suno song sounds quiet next to everything else" is one of the most common complaints from people releasing AI music. The usual answer is "master it louder". Our measurements say that is often the wrong fix. Before you push the level, find out which of four reasons applies to your track.

First, measure it

Ears are bad at judging loudness when you switch between apps, headphones and speakers. Measure the file you are going to release: integrated loudness (LUFS), true peak (dBTP) and loudness range (LU). The free LUFS checker does it in a few seconds, without an account.

Then compare your numbers with what Suno exports usually look like.

What Suno exports actually measure

We measured 248 unique Suno exports that people uploaded for mastering between July and October 2026 (method and full table in the Suno loudness report):

• median integrated loudness: −14.3 LUFS, and half of the tracks sit between −15.0 and −13.5 LUFS;
• only 8% are quieter than −16 LUFS;
• 36% are louder than Spotify's −14 LUFS reference;
• 40% have a loudness range under 5 LU, so verses and choruses play at nearly the same level.

So a typical Suno export is not quiet by streaming standards. If yours measures around −14 LUFS and still sounds quiet, the cause is one of the first three reasons below.

Reason 1: you are comparing it with unnormalized playback

In a DAW, a media player or a plain audio file on your phone there is no loudness normalization. Many commercial masters are made much louder than −14 LUFS, so next to them a −14 LUFS track sounds smaller.

Streaming services level this out. Spotify's default "Normal" setting plays tracks at −14 LUFS and turns louder masters down to that level, so a −14 LUFS Suno track and a much louder commercial master end up at about the same volume. Compare your song where it will actually be heard, or with normalization switched on.

Reason 2: peaks leave no room to turn it up

Normalization also works upwards, but only within limits. Spotify lifts a quieter track only as far as its true peak allows. In Spotify's own example, a track at −20 LUFS with a true peak of −5 dBFS is lifted only to −16 LUFS, not to −14.

A quiet Suno track with loud, spiky peaks therefore stays quiet on Spotify. The fix is not "louder" but controlled peaks: a limiter that brings the peaks down so the whole track can reach about −14 LUFS, with true peak at or under −1 dBTP. Spotify recommends below −1 dBTP, and below −2 dBTP if the master is louder than −14 LUFS. More in Spotify loudness and true peak explained.

Reason 3: it is loud but flat, so it sounds weak

Four in ten Suno exports in our sample have a loudness range under 5 LU. When everything plays at one level, the chorus does not lift, drums lose their punch and the song feels smaller than its loudness reading.

Pushing a flat track louder makes this worse: the limiter squeezes the little contrast that is left and brings the metallic AI shimmer forward. What helps is the opposite: clean the AI artifacts first, then master at a sensible target without crushing the dynamics. See AI artifact cleanup.

Reason 4: it really is quiet

About one Suno export in twelve in our sample measured quieter than −16 LUFS. If yours is one of them, master it up to about −14 LUFS for Spotify and YouTube, keeping true peak at or under −1 dBTP. For Apple Music, which plays at about −16 LUFS, −14 LUFS is still fine: it will simply be turned down a little. Targets for every service are in streaming loudness targets.

A quick checklist

1. Measure integrated loudness, true peak and loudness range of the exact file you will release.
2. Around −14 LUFS and still sounds quiet? Compare with normalization on before changing anything.
3. True peak above −1 dBTP? Control the peaks so the track can be played at full level.
4. Loudness range under 5 LU? Clean the artifacts and keep the dynamics instead of pushing louder.
5. Below −16 LUFS? Master it to about −14 LUFS, true peak at or under −1 dBTP.

Echoteka does steps 3–5 in one pass and shows the numbers before and after, so you can see what changed. One full master is free every day, no account needed.