The Loudness War in 2026: Why Louder Masters Lose

October 17, 2026 · All posts

The loudness war is the long push to make masters louder than the next record by squeezing out their dynamics. On streaming in 2026, that push mostly costs you. Platforms measure each track and adjust playback to a target, so a master pushed to −9 LUFS is turned down to the target and keeps only the damage. A louder master loses because the extra loudness it gained is taken back, while the squeeze stays.

This article covers how the war started, what normalisation does, why louder masters lose, what Echoteka's measurements of Suno exports show, and how to check the loudness penalty on your own track.

How the war started

For decades a louder track stood out on the radio and on the CD, and a louder song could win the room in a club. Engineers answered with stronger compression and limiting, and peaks were pushed up to the digital ceiling of 0 dBFS. Digital masters have a hard ceiling, so there was no soft edge to hide behind. The result, which the industry called the loudness war, was music with less dynamic range.

Streaming changed the rules. Services now normalise playback, so a louder file does not sound louder on the platform. The competition moved from how loud a file could go to how much of its dynamics survived.

What normalisation does to a loud master

Take a master at −9 LUFS integrated, on a platform with a −14 LUFS reference in Normal mode. The platform turns it down by about 5 dB. The listener hears roughly the same level as a track at −14 LUFS, so the loudness gained by squeezing has been taken back.

What the listener keeps is the squeeze. Transients are flattened, the gap between quiet and loud sections narrows, and peaks sit closer to the ceiling. Echoteka's Spotify loudness figures note that tracks louder than −14 LUFS should stay below −2 dBTP, because encoding can push peaks up. A loud master has less room for that.

Apple Music is usually measured around −16 LUFS, and YouTube and Amazon sit near −14, so the same logic applies across services. On each of them, a master above the reference is turned down.

A worked example

This is an illustration, not a measurement. Imagine two masters of the same song. Master A reads −9 LUFS integrated, 4 LU of loudness range and −0.3 dBTP true peak. Master B reads −13 LUFS, 7 LU and −1.2 dBTP. On a platform with a −14 LUFS reference in Normal mode, A is turned down by about 5 dB and B by about 1 dB. Both then play at about −14 LUFS. A still has 4 LU of range, and B still has 7 LU. The turn-down lowers the peaks too, but it does not return the range. At the same playback level, B simply sounds less squeezed.

What the loudness penalty costs

There are three costs, and they add up:

• Dynamics. The loudness range shrinks, so the chorus has less lift over the verse, and the small level changes that carry groove get flattened.
• Headroom. A loud master has less room to survive encoding, so overs can appear after conversion.
• Distortion. Heavy limiting of a dense mix adds audible distortion, and the platform's turn-down does not remove it.

What Echoteka's measurements show

Echoteka measured 248 unique Suno exports that users uploaded for mastering, from 26 July to 2 October 2026, before any processing. Most exports already sit near streaming loudness. The weak spot is dynamics:

• The median integrated loudness is −14.3 LUFS, with half of the tracks between −15.0 and −13.5 LUFS.
• 36% of exports are louder than −14 LUFS, and 6% are louder than −10 LUFS.
• Among exports louder than −14 LUFS, the median reduction Spotify would apply is 0.9 dB.
• The median loudness range is 5.8 LU, and 40% of exports are below 5 LU.
• The median crest factor is 14.6 dB.

So among the loud exports, the platform's cut is 0.9 dB at the median, and a platform cannot give back range that was already squeezed out. The dynamics problem sits in the master, not in the playback chain. Cite these numbers with a link to the Suno loudness report, which is published under CC BY 4.0.

Where a louder target fits

Not every loud target is a mistake. Club and DJ playback is a different setting, and the calculator's club target is about −9 LUFS. There, loudness is a choice for the room. For streaming, the reference belongs to the platform, and a louder file does not earn a louder playback level.

How to check your own loudness penalty

1. Measure the master. Note integrated loudness, true peak and LRA with the LUFS checker.
2. Find the gap. Subtract the target: about −14 for Spotify and YouTube, about −16 for Apple. The difference is roughly the turn-down to expect.
3. Make a matched copy. Lower the master by that gap so it sits on the target. This is roughly what the platform would play at its normal level.
4. Compare at matched loudness. Play the loud master and the matched copy on the same speakers, then on a phone and earbuds, at the same integrated level. A louder file always wins the opening bars, so judge after a full verse and chorus.
5. Check the range. If the louder version has lost more than 1 to 2 LU of loudness range without a clear gain in clarity, you have paid the penalty.

If the penalty shows up, the fix is in the master: choose a target the mix can hold without squeezing, and let the platform handle the rest.

Common mistakes

• Expecting the loud file to win on streaming. Normalised playback turns it down.
• Judging a level match by ear. A 1 dB difference is enough to make the louder version win for the wrong reason, so match loudness before you compare.
• Chasing a number for the competition. Other people's loud masters are turned down too.
• Copying a club reference. A reference heard in a DJ set is a room target, not a streaming target.
• Stacking a maximiser on a limited master. Each extra stage squeezes more and leaves less room for encoding.
• Forgetting true peak. A master can sit at the right loudness and still clip after encoding. Check the true peak too, as true peak explains.

How to do it in Echoteka

• Echoteka masters every track through an oversampling true-peak limiter with a −1 dBTP ceiling and reports true peak before and after.
• The Universal sound profile targets −14 LUFS, and the EDM profile aims at −9 LUFS. If a target is not reached, the report says so.
• Every master comes with a before-and-after report on loudness, true peak, loudness range, crest factor and stereo.
• Everyone gets 1 free master every day, no account needed. For the platform targets, see the streaming loudness page.

Loudness is one setting among many. The numbers show whether level, peaks and range suit a platform. They do not show whether the mix is good, and that still needs listening.

FAQ

Why does a louder master sound worse on streaming? The platform turns it down to its target, which removes the loudness you gained. The dynamics you gave up to get there stay lost.

Is the loudness war over? On streaming, the reward for loudness is gone, because playback is normalised. Many masters are still pushed hard, so the cost is still worth checking.

What LUFS should a streaming master be? About −14 LUFS integrated for Spotify and YouTube. Apple does not publish a target, and third-party measurements put its playback near −16 LUFS.

What is the loudness penalty? The turn-down a platform applies to a master that is louder than its reference, together with the dynamics the master gave up to reach that loudness.

Does Echoteka aim for a particular loudness? Each sound profile has its own target. Universal aims for −14 LUFS and EDM for −9 LUFS. The report shows whether the target was reached.