Spotify and Amazon audio ads share one loudness target: -14 RMS

Spotify's video ad page and Amazon's audio ad page both ask for RMS at -14 dBFS and peak at -0.2 dBFS. Where Sume's gain_db helps and where you must measure.

5 min readSume
All posts

One mix can serve both: Spotify's in-stream video ad page and Amazon's audio ad page each ask for audio with RMS normalized to -14 dBFS and peak normalized to -0.2 dBFS. Sume's timeline has a gain_db field on the audio spine to move the level, but nothing in the docs measures loudness, so check the result with a meter you trust before you upload.

Spotify's spec page (read 2026-10-04) states both figures and says it accepts creative up to 30 seconds. Amazon's audio page (read 2026-10-04) states the same two numbers, plus at least 192 kbps. Neither page says how they measure RMS, so use the same method on every file.

Where do the two specs differ?

The loudness lines match; almost everything else does not.

Spotify video and Amazon audio ad specs compared, read 2026-10-04
ItemSpotify in-stream videoAmazon audio
RMS-14 dBFS-14 dBFS
Peak-0.2 dBFS-0.2 dBFS
LengthUp to 30 seconds10 to 30 seconds
File size500 MB maximum3 MB maximum
FormatsMOV, MP4WAV, MP3, OGG
Bitrate198 kbps ideal, 320 kbps maximumAt least 192 kbps

How do I move the level in Sume?

Timeline's audio spine accepts an url or parts, a mode, a source_in offset and a gain_db value. Use a negative gain to pull a hot mix down, and re-render. A timeline render costs $0.10 per output minute, rounded up, and needs an Idempotency-Key. The loudness numbers themselves are yours to measure outside Sume, because the docs list no normalization step.

For the audio-only Amazon deliverable, use audio-detach to pull a WAV from the finished video, as the post on Amazon's 3 MB cap shows. Detach is $0.01 per job and does not change the level.

What is a safe workflow?

Build the mix once, render the video, detach the WAV, and measure both files. If the peak sits above -0.2 dBFS, lower the gain by the overshoot plus a small margin and render again. Lowering gain also lowers RMS, so check both numbers after each change instead of one.

Peak and RMS interact: a dynamic mix can hit the peak ceiling long before its RMS reaches -14 dBFS. That is a mastering problem, not a Sume setting, and a limiter in your audio tool is the usual fix.

Is it worth building one mix for both?

Yes if the same voice and music run on both services, since you pass the check once. If the video has loud effects that Amazon's audio-only cut does not need, give each its own mix. The stored posts on Spotify video ad specs and audio ad specs list the other numbers.

Sources

Related posts

More in Media tools

All Media tools posts

Written by Sume