Guides

How to master an ElevenLabs Music track for Spotify

Updated 2026-09-16

ElevenLabs Music v2.5 became the default model on September 11, 2026. The change that matters most for anyone releasing tracks is the download, not the songwriting: every plan now gets lossless WAV, five tracks a day on Free, four hundred a month on Pro. That is a real file to work with, not an MP3 that already threw its top end away.

What an ElevenLabs export actually measures as

Unmastered. Typical integrated loudness sits between -16 and -20 LUFS, quiet by streaming standards, and Spotify's normalization will turn it up and take any density problems along with it. True peak is not controlled at the source, so a file that looks fine at -18 LUFS can still carry peaks that clip once a platform's codec rebuilds the waveform. Like Suno and Udio, ElevenLabs tracks tend toward a soft high-frequency rolloff and a low-mid buildup, though usually milder than a hard 16 kHz cliff.

Step 1: check before you touch anything

Run a spectrum and loudness analysis first. An ElevenLabs file is closer to a finished mix than a Suno export, so the repair it needs is usually lighter, sometimes none at all. Do not run a generic AI-repair chain by default. Measure, then fix only what is actually there.

Step 2: repair what the measurement finds

• If the top end rolls off early, rebuild gently from what survives rather than boosting a shelf over nothing.
• If there is a low-mid buildup, check 250 to 400 Hz first. That is the same range that causes boxiness in other generators.
• Check stereo width rather than assume it. It usually starts closer to normal than a Suno export, but confirm it on this file.

Step 3: master to -14 LUFS, -1 dBTP for Spotify

Since the file starts quiet, getting it to a competitive streaming loudness, roughly -9 to -12 LUFS integrated for pop and adjusted for genre, takes real gain. That gain is exactly where an uncontrolled true peak turns into clipping if nothing is watching it. Hold true peak at or under -1 dBTP with a true-peak limiter, not a sample-peak one, and check that crest factor did not collapse on the way there.

Why this is not the same job as a Suno repair

A Suno or Udio export usually needs the repair stage first: a top-end rebuild across a hard cutoff, a real mud cut, real widening, before loudness means anything. An ElevenLabs export is more often just quiet and unmastered, with milder versions of the same tendencies. Running the full repair chain on a file that does not need it can do more harm than good, thinning a top end that was only a little soft to begin with. Measure first and let the numbers decide how much to touch.

Provenance and disclosure

ElevenLabs embeds provenance information in what it generates. Burnish preserves that metadata and never strips a watermark, on any generator's output. Distributors do not treat AI disclosure the same way: DistroKid has an AI credits field, TuneCore currently accepts AI tracks only from its list of partner models, and CD Baby rejects AI-generated music outright. Check your distributor's current policy before you submit.

Run the free analysis on your ElevenLabs export first. It shows the actual loudness, true peak and spectrum on your file, so you know how much a mastering pass actually needs to fix.