YouTube Loudness: Why Your Video Plays Back at -14 LUFS

Upload a loud video and YouTube turns it down. Upload a quiet one and — oddly — it stays quiet. Understanding that asymmetry is the whole game of YouTube loudness. This guide explains the -14 LUFS playback target, what YouTube does and does not fix for you, and how to master so the platform leaves your audio alone.

For podcasts and audio-first uploads, the Streaming & YouTube preset in the loudness normalizer applies the same -14 target.

What YouTube Does With Your Audio

YouTube measures the integrated loudness of every upload and turns loud content down to about -14 LUFS. It does not turn quiet content up — turning up would raise noise and break the peaks of quiet masters, so quiet uploads stay quiet. The volume slider your viewer sees is then relative to that normalized level.

Your upload measuresYouTube's moveResult
-8 LUFS (loud master)About -6 dB gain reductionYour master, quieter; dynamics preserved
-14 LUFSNoneExactly your master
-22 LUFS (quiet master)NoneViewers reach for the volume — and the next video blasts them

That last row is the one that hurts channels. The viewer turns up for your quiet video, forgets, and the next video on the platform — mastered at -14 like everything else — arrives 8 dB louder than expected. Viewers do not file a report about your loudness. They flinch, turn down, and associate the jolt with your video. Meanwhile two otherwise identical videos, one at -14 and one at -22, get exactly the same treatment from the algorithm — the -22 one simply sounds like a smaller, weaker channel to every viewer who finds it.

Why YouTube Only Turns Down

The asymmetry is deliberate, and once you see the reasoning it is obviously right. Turning a loud video down is always safe: gain reduction cannot clip, cannot raise noise, cannot damage anything. Turning a quiet video up is a gamble: if the quiet master has peaks near full scale — and quiet average with loud peaks is the signature of most amateur audio — a straight turn-up drives those peaks into clipping for every viewer. YouTube would be adding distortion to millions of uploads it cannot audition individually. So the platform corrects the safe direction and leaves the dangerous one to you.

The lesson for creators is plain: loudness on YouTube is a one-way ratchet, and the only way to be loud enough is to arrive loud enough — at exactly -14, with peaks under control, so the ratchet never touches you.

How to Master for YouTube

  1. Mix for speech clarity first. Loudness cannot rescue a muffled voice; it only sets the overall level (balance problems are covered in the loudness range guide).
  2. Measure the final export's audio. Video editors show peaks; you need integrated LUFS — the LUFS explainer defines it.
  3. Normalize the audio to -14 LUFS with true peak at or below -1 dBTP (why the ceiling), then re-attach it to the video, or export the video's audio track, normalize it, and re-upload the finished file.
  4. Verify before upload. The report should read -14 ±1 LUFS. That file passes through YouTube's system untouched.

The same -14 target serves Spotify, which is why podcasters who publish video versions can share one master — the Spotify guide covers the audio platforms' side.

In practice, step 3 has two common shapes. If your editor lets you export the finished video, extract its audio, normalize the audio file, and set it back onto the timeline before the final render, the loudness you measured is the loudness you ship. If your workflow exports audio separately anyway (podcast-first channels often do), normalize that audio master to -14 and use it for both the video render and the podcast feed — remembering that the podcast feed itself prefers -16, as the guide linked above explains. Either way, the rule is the same: measure the artifact that actually leaves your machine.

A Worked Example: Two Uploads, One Channel

A creator uploads two videos a week apart, both exported from the same editing template:

  • Video A was mixed by ear with the editor's meter peaking near 0. It measures -9.5 LUFS, true peak -0.2 dBTP. YouTube applies about -4.5 dB. The video plays at platform loudness — the loud master bought nothing, and its heavy limiting is still audible as harshness.
  • Video B used a guest recording that was quiet, and the creator "did not want to touch anything." It measures -23 LUFS. YouTube applies nothing. Viewers turn up 9 dB to follow the conversation; the sponsor's pre-roll, mastered at -14, arrives like a shout, and the comments mention the audio within the hour.

The fix for both is the same ten minutes: normalize the audio of each to -14 LUFS with a -1 dBTP ceiling, verify the report, then render and upload. Video A's turn-down disappears because there is nothing to turn down. Video B arrives at full platform loudness with its peaks intact, because the gain was applied by a tool watching the peaks, not by a playback slider flying blind.

Music vs Speech on YouTube

Music can sit happily at -14 because platforms treat the whole catalogue the same way. Speech needs the same target but more care with range: a whispered aside at -14 overall still vanishes under a viewer's room noise. Keep speech LRA moderate (roughly 5–10 LU) and the loudness number will do the rest.

Genre notes beyond that rule of thumb:

  • Gaming and commentary. Game audio and voice fight for the same loudness budget. Duck the game under speech in the mix — a file that integrates to -14 with the game on top of the voice fails its viewers even though it passes the meter.
  • Vlogs and field recording. Changing environments make wild short-term swings. Moderate range matters more than the last half-LU of integrated precision; level the scenes before the loudness pass.
  • Music performances and covers. You are graded against produced tracks at -14. A live-room recording at -14 integrated with wide dynamics will feel quieter than a dense studio track at the same number; a touch of bus compression in the mix is the honest fix.
  • Shorts and clips. Short content is judged fast and loud by scrolling viewers. The same -14 reference applies; do not master clips hotter hoping to grab attention — the ratchet removes the advantage and keeps the harshness.

Build Loudness Into Your Upload Routine

One-off fixes fade; routines stick. A workable per-video routine looks like this: export the audio of the final cut, measure it, normalize to -14 with the ceiling, re-attach or render, and glance at the report before the upload starts. On LoudNorm that is one pass per video — upload the audio, pick the Streaming & YouTube preset, download the corrected file with its report. The free plan's 60 minutes a month covers a typical weekly channel's audio; heavier schedules sit comfortably inside the paid tiers, and no plan watermarks or stamps your audio.

Keep the reports in the video's project folder. Three months from now, when a viewer says a video "sounded quiet," the report settles in seconds whether the file was on target — and if it was, you know to investigate range or the mix instead of re-uploading blind. Channels that sound professional for years are not using secret tools. They are running this boring routine on every single upload, including the rushed ones — especially the rushed ones, because those are the uploads where loudness quietly goes wrong.

Common Mistakes

Mastering to 0 dB peak because the editor's meter allows it. Peak is not loudness. A video can peak at 0 and integrate to -20 (mostly quiet, a few bangs) or to -9 (loud throughout). YouTube reads the integrated number; master to it.

Fixing quiet audio with the editor's "normalize" button. In most video editors that button is peak normalization — it raises the loudest moment to a ceiling and leaves the average wherever it was. A -22 video can come out of "normalization" still measuring -20. Use a LUFS measurement, not the button's promise.

Letting the platform's volume be your quality control. If you preview your upload at full laptop volume and it sounds "fine," you have learned nothing — the viewer's normal volume, against the previous video in their session, is the real test. The report's -14 ±1 is a better preview than any speaker in your room.

Re-uploading a corrected version without checking the new file. Measure the replacement artifact, not the project you fixed. Export settings, an old render in the upload folder, a soundtrack re-attached from the wrong take — the measurement on the final file catches all of them in seconds.

Assuming subscribers will not notice. They notice constantly; they just describe it as "audio seems off lately" or stop finishing videos. Loudness inconsistency is one of the few technical flaws viewers reliably feel without having words for it.

Troubleshooting

My video plays quieter than others even though it measures -14. Compare loudness range, not just the integrated number. Wide-range speech at -14 spends much of its time well under -14, and viewers judge by the moment they are hearing. Narrow the range in the mix, then re-normalize.

YouTube turned my loud video down and now it sounds dull. The turn-down is clean gain; dullness was in the master. A master limited hard to reach -9 sounds flat at any playback level. Re-master with dynamics intact at -14 — the platform cannot give back what the limiter took.

My quiet video has been up for a year. Re-upload or leave it? YouTube offers no way to replace a video's audio in place for loudness alone. A re-upload loses the URL, views and comments. Unless the video is a flagship evergreen piece, leave the archive and fix the workflow going forward; new viewers judge the channel by its current sound.

Podcast videos: one master or two? If the video is the product and the podcast feed is secondary, master at -14 for both and accept the slight heat in podcast apps. If the audio feed is the product, keep two masters — -16 for the feed, -14 for the video — and let each be native where it lives.

Frequently Asked Questions

Does YouTube publish an official loudness spec? YouTube's help pages cover audio quality for uploads rather than a formal delivery spec, but the -14 LUFS playback reference is widely documented and directly observable: upload a file at a known loudness and the applied adjustment matches the distance to about -14. Master to it and the platform leaves you alone.

What true-peak ceiling should I use for YouTube? -1 dBTP, the same as Spotify. YouTube re-encodes every upload, and the ceiling is the headroom that encode needs to avoid clipping your loudest moments.

Does loudness differ for live streams? Live audio is normalized less predictably, and you cannot re-measure a stream after the fact. Aim your live chain at the same -14 average with peaks under control, using a loudness meter on the output — the habit transfers directly from uploads.

Will a -14 master hurt my podcast feed? It plays about 2 LU hot against the -16 podcast convention — noticeable to careful listeners, forgiven by most. It will not hurt the video at all, which is the destination this guide is about. Choose per your main audience, and keep the choice consistent across the catalogue.

Can I check a competitor's loudness? Not exactly — the platform does not publish per-video loudness data. What you can do is the queue test: their video and yours, back to back, one volume setting. If you reach for the volume, one of you is off the reference — and the report tells you whether it is you.

References

← All guides