AUDIO · LOUDNESS BLOG / 8 MIN READ

How loud should my YouTube video be?

YouTube normalises loudness toward −14 LUFS and turns loud uploads down to match. Master to that target, keep peaks under −1 dBTP, and your audio plays at full strength instead of sounding weak or distorted.

−14 LUFSYouTube target
−1 dBTPtrue peak ceiling
voiceon top of music
0 to 100craft score

By Thomas, founder of CutScore · Updated June 2026

LOUDNESS CHECK · final_mix.mp4
An audio engineer leaning over a mixing console, watching the meters, the moment a video's loudness gets set before it is exported and uploaded.
CRAFT SCORE
FIXES ADVISED
where loudness gets decided
Too quiet · −19.4 LUFS, lift +5 dBwhole file
True peak safe · −1.6 dBTP
Music over voice · duck −5 dB00:42
The 30-second answer Master your YouTube video to an integrated loudness of about −14 LUFS, with a true peak no higher than −1 dBTP. YouTube measures every upload and turns loud videos down to that level, but it never turns a quiet one up. So if you sit much below −14 LUFS your audio plays weak next to everything else; if you go far above it and let peaks run hot, the platform's re-encode can crackle. Hit the target, keep the voice clearly on top of any music, and you are done. If reading a loudness meter is not how you want to spend your evening, that is exactly the measurement CutScore makes for you.
WHY THIS QUESTION IS CONFUSING

The reason "how loud should it be" feels slippery is that loudness has two different units, and most people only know the wrong one. Your editor probably shows you decibels on a meter, peaking somewhere near zero. That tells you how close a single instant is to digital clipping. It tells you almost nothing about how loud the video feels over its whole length, which is the thing YouTube actually cares about.

The unit that matters is LUFS, short for Loudness Units relative to Full Scale. It is a perceptual measure, averaged across the entire file, weighted to match how human ears hear. YouTube computes the LUFS of everything you upload and uses it to even out the feed, so a quiet lo-fi clip and a loud trailer play at roughly the same level back to back. I have shipped videos that sounded fine in my editor and arrived on the platform sounding like a whisper, because I was reading peaks and ignoring LUFS.

So there are really two numbers to hit, not one. Integrated loudness sets how strong the whole thing feels: target −14 LUFS. True peak sets how close the loudest split-second gets to distortion: keep it at or below −1 dBTP. Get both right and the rest is just keeping your voice above the music. Here is what each number means and how to land it.

THE TWO NUMBERS

The loudness targets for YouTube, plainly.

Two measurements decide whether your audio sounds professional or weak. Both have a target you can hit, and a cost if you miss it.

MeasurementTarget to hitWhat it costs you if you miss it
Integrated loudness≈ −14 LUFSToo quiet and YouTube leaves it quiet, so your video sounds weak next to the next one.
True peak≤ −1 dBTPPeaks above this can clip and crackle once the platform re-encodes the file.
Loudness rangecontrolledWild swings make viewers ride the volume knob, loud then inaudible.
Voice vs musicvoice on topMusic sitting over speech is the most common loudness complaint there is.
Noise floorlow, steadyRaising a quiet mix also raises the hiss, so clean the floor before you lift.
The one trap to rememberYouTube only ever turns loud videos down to −14 LUFS. It will not turn a quiet video up. So there is no prize for mastering louder than the target, and a real penalty for going quieter. When in doubt, sit right on −14 LUFS, not above or below it.
SKIP THE METER

CutScore measures integrated loudness, true peak and the voice-to-music balance on your export, then tells you the exact gain change to make. No meter to read, no guessing.

Join the waitlist
HOW TO ACTUALLY HIT IT

Four passes to the right loudness.

1. Measure where you are now, in LUFS

You cannot fix a number you have not read. Open a loudness meter on your master bus and play the whole video, start to finish, because integrated loudness is an average over the entire file. The reading you want is −14 LUFS for YouTube. If your meter says −19, you are five decibels too quiet. If it says −10, you are four too loud and YouTube will pull you back anyway. Note the gap, that is your gain change for the next step.

2. Set the level, then guard the peak

Raise or lower the whole mix by the gap you just measured to land on −14 LUFS. Then put a true-peak limiter as the last thing in the chain with its ceiling at −1 dBTP. That ceiling is not optional. When YouTube re-encodes your file to its own format, the conversion can push samples slightly higher than your meter showed, and a clip that peaked at 0 dB on your timeline can arrive distorted. Leaving a decibel of headroom is what keeps the loud bits clean.

Two people working over a mixing board together, setting levels and balancing voice against music before a video is exported.
Set the overall level first, then guard the peak at −1 dBTP. Photo: cottonbro studio / Pexels.

3. Put the voice clearly on top of the music

A correct −14 LUFS mix can still sound wrong if the music is fighting the voice. The whole mix averages to the target, but inside that average the speech has to win. If you ever find yourself raising the music because the voice feels thin, you are solving it backwards. Pull the music down a few decibels instead, or duck it automatically under speech. People will forgive quiet music. They will not forgive missing every third word you say.

4. Clean the floor before you lift a quiet clip

If your recording came in at −20 LUFS, you might be tempted to just crank everything up to target. Careful: turning up a quiet mix turns up its hiss, hum and room tone in equal measure, and now you have a loud video that sounds cheap. Reduce the background noise first, then raise the level. The order matters. A clean floor at −14 LUFS sounds confident; a noisy floor at −14 LUFS sounds like a stairwell.

RATHER SEE IT THAN READ IT?

Here is a real CutScore report for an everyday vlog: loudness, true peak and voice balance, all measured, with timestamps and the exact gain changes.

See a sample report
PLATFORMS DIFFER

Is −14 LUFS the same everywhere?

YouTube settled on roughly −14 LUFS, and most desktop and TV listening is built around it. Short, phone-first feeds tend to run hotter, so the target shifts a little.

1
YOUTUBEAUDIO
YouTube: aim for −14 LUFS
This is the safe home target for long-form and most uploads. The platform turns louder videos down to match it and leaves quieter ones alone, so −14 LUFS is the level where your audio plays at full strength without being attenuated.
How Master the whole mix to −14 LUFS, ceiling at −1 dBTP. See LUFS for YouTube.
2
SHORTS & TIKTOKAUDIO
Vertical feeds tend to run a little louder
Phone-first, sound-on, scrolling feeds tolerate a hotter mix than a TV does, so short-form often sits closer to −14 down to around the level the platform allows. The voice-on-top rule does not change, and the −1 dBTP ceiling does not change either.
How Check the specific targets for TikTok and Instagram Reels.
3
ONE MASTERWORKFLOW
When in doubt, master to −14 and let the platform manage it
If you post the same video to several places, mastering once at −14 LUFS with a −1 dBTP ceiling is the sane default. Loud platforms turn it down predictably; nobody turns it up. You avoid the trap of mastering hot for one feed and arriving distorted on another.
How Pick the target for your main platform, export once, verify the upload. See how to check audio levels.
THREE WAYS TO CHECK IT

By ear, by meter, or in one pass.

OPTION 01

By ear, against a reference

Free, and a useful gut check. Play a YouTube video you think sounds right, then play yours at the same system volume. If yours is noticeably weaker, you are below target. The catch is your ears adapt fast, and laptop speakers flatter your mix, so this only catches the big misses.

OPTION 02

With a loudness meter

Accurate and honest. An EBU R128 or LUFS meter on your master bus gives you the real integrated loudness and true peak. The cost is knowing the targets, running the whole file through, and reading two numbers correctly on every single video you make. Fine if you enjoy it.

OPTION 03

With a coach in one pass

Hand the file or a link to CutScore. It measures integrated loudness, true peak and the voice-to-music balance, tells you the exact gain change, and rolls it into a 0 to 100 craft score with timestamps. No meter to read. See a sample report.

How CutScore measures loudness for you CutScore is an AI video quality coach for pre-publish QC. It computes the measurable craft deterministically: integrated loudness with an EBU R128 meter, true peak in dBTP, the loudness range, and where the music is winning over the voice. You get one score, the evidence behind each number, and the exact gain change to make, before anyone else hears the video. It judges the craft of the audio itself, not your tags or thumbnail, so it sits next to a growth tool rather than competing with one. More on the method and the standards.
QUESTIONS

Frequently asked.

Aim for an integrated loudness of about −14 LUFS with a true peak no higher than −1 dBTP. YouTube measures every upload and turns loud videos down to that level anyway, so mastering near −14 LUFS means your audio plays at full strength instead of getting quietly attenuated. Voice should sit clearly on top of any music.
YouTube turns it down to roughly −14 LUFS during playback, but it never turns a quiet video up. So going louder gains you nothing and can cost you: if your true peak is above −1 dBTP, the platform's re-encode can push it into clipping, which is the crackle you hear on hot audio. Master to the target, do not exceed it.
Almost always because your integrated loudness is well below −14 LUFS. YouTube turns loud uploads down to match, but it leaves quiet ones quiet, so a video mastered at −20 LUFS plays noticeably weaker than the one before it. Raise the overall level toward −14 LUFS and keep peaks under −1 dBTP.
No. The voice should sit clearly on top, with background music usually sitting several decibels under it. The whole mix together targets about −14 LUFS, but within that mix speech has to win. Music drowning the voice is the single most common audio mistake on the platform, and it is fixed by pulling the music down, not the voice up.
EARLY ACCESS

Stop guessing your loudness.

CutScore measures integrated loudness, true peak and the voice-to-music balance, then tells you the exact change to make, before you publish. Join the waitlist for early access.

Join the waitlist