REELS AUDIO BLOG / 8 MIN READ

What loudness should Instagram Reels be?

Instagram turns loud audio down so the feed stays even, which means hot masters buy you nothing. Master your Reel near −14 LUFS with peaks under −1 dBTP, and here is exactly how to hit that and keep the voice on top.

−14 LUFSloudness target
−1 dBTPpeak ceiling
9:16the Reels frame
85%+watch on a phone

By Thomas, founder of CutScore · Updated June 2026

LOUDNESS CHECK · reel_export.mp4
A mixing console glowing with channel faders and level meters, the kind of desk where a Reel's loudness gets set before it is exported and uploaded.
CRAFT SCORE
FIXES ADVISED
measuring the Reel before upload
Loudness too quiet · −19.4 LUFS, raise toward −14whole
True peak in spec · −1.3 dBTP
Music over voice · duck −4 dB00:06
The 30-second answer Master your Instagram Reel around −14 LUFS integrated, with the true peak held at or below −1 dBTP. Instagram normalises loud audio downward so every clip in the feed sits at a similar level, which means a hotter master gains you nothing: the app simply turns it back down, often while adding distortion. Aim near −14 LUFS for a full, confident sound, keep a decibel of peak headroom so it survives the upload re-encode, and make sure your voice clearly beats the music. If checking that by hand sounds tedious, that is exactly the job CutScore does in one pass.
WHY THE QUESTION IS TRICKY

Instagram has never published an official loudness number for Reels, and that silence is why the question keeps coming up. So people guess. Some master as hot as a music single, then wonder why their Reel sounds crushed and brittle. Others master far too quiet, then wonder why their clip feels weak the instant the next one autoplays. Both are reacting to the same hidden mechanic: the feed evens everybody out.

Here is the part that trips people up. Instagram applies playback loudness normalisation, the same broad idea YouTube and TikTok use. It measures how loud your clip is and nudges the playback level so your video does not blast or whisper relative to the one before it. That is great for viewers and slightly annoying for creators, because it means the loudness you exported is not necessarily the loudness people hear.

So the real question is not "how loud can I make it." It is "what target survives normalisation cleanly and still sounds full." For short vertical video in 2026, that target lives right around −14 LUFS, with a strict peak ceiling underneath it. Let me show you why that number, and how to actually land on it.

THE TARGET

What loudness should Instagram Reels be?

Two numbers carry almost all the weight: an integrated loudness target and a true peak ceiling. Hit both and your Reel sounds at home in the feed instead of fighting it.

What to setTarget to hitWhy it matters for Reels
Integrated loudness≈ −14 LUFSFull and confident, and it survives Instagram's downward normalisation without sounding crushed.
True peak≤ −1 dBTPLeaves headroom so the upload re-encode does not push a peak over and add crackle.
Voice vs musicvoice clearly on topOn phone speakers, buried speech is the fastest way a Reel reads as amateur.
Short-term spikescontrolled, no slammingA drop or a laugh that jumps 10 dB above the rest is jarring on a small speaker.
Silence at the startnoneA quiet first half-second reads as a dead Reel before the audio even arrives.
Why −14 and not louderBecause louder does not reach the listener. Push your master to −9 LUFS and Instagram normalises it back toward its target anyway, so all that extra limiting earns you is a flatter, more fatiguing sound for no loudness gain. The same −14 LUFS target works as a sane home base for LUFS on YouTube too, which makes life easier if you post the same clip in more than one place.
DON'T GUESS THE NUMBER

Reading a loudness meter on every Reel gets old fast. CutScore measures your integrated loudness and true peak, compares them to the target, and tells you the exact gain change to make.

Join the waitlist
HOW TO ACTUALLY HIT IT

Hitting −14 LUFS without guessing.

1. Measure what you already have

You cannot fix a number you have not read. Drop your finished Reel onto a loudness meter that reports integrated LUFS, the one that measures the whole clip rather than the moment. Most editors have one built in now, and there are free meters if yours does not. If it reads −19 LUFS, you are quiet and the feed will make you feel quieter. If it reads −9 LUFS, you are slamming and Instagram is about to turn you down. Read first, then act.

2. Move the whole mix, not just the peaks

Integrated loudness is about the average energy of the clip, not the single loudest spike. So if you are too quiet, raise the gain of the entire mix, then re-measure. A loudness normalisation tool does this for you: tell it −14 LUFS and it shifts the level to land there. The part people forget is the ceiling. After you raise the level, confirm your true peak still sits at or below −1 dBTP. If it crept over, a true peak limiter on the master pulls it back without making the whole thing quieter.

Two people leaning over an audio mixing board adjusting faders together, the moment where a Reel's voice and music get balanced before the loudness is locked in.
Loudness is the last step, after the voice already beats the music. Photo: cottonbro studio / Pexels

3. Win the voice-versus-music fight first

Loudness is the last thing you set, not the first. If your music is sitting on top of your voice, raising the whole mix to −14 LUFS just makes the problem louder. So before you touch the master level, pull the music down four or five decibels under the speech, and duck it further whenever you are actually talking. This is the same trap I describe in why your music ends up louder than your voice, and it is the single most common reason a perfectly loud Reel still sounds amateur.

4. Check it on the worst speaker you own

Studio headphones lie to you in a flattering direction. Most of your audience is on a phone speaker the size of a fingernail, sometimes on a bus, sometimes with a thumb half over the grille. So play your final Reel through exactly that: your phone, no headphones, at a normal volume. Can you understand every word over the music? Does the loud moment make you flinch? If yes, fix it now. For the wider picture, see how to check your audio levels before uploading.

5. Export, upload, then listen to the published version

This is the step everyone skips and then regrets. Instagram re-encodes your audio into a lossy format on upload, which is the moment a peak sitting right at 0 dBFS can tip over into crackle. That is the whole reason for the −1 dBTP ceiling. Export your Reel, post it, then play the live version back in the app, not the file on your desktop. If the published version distorts on loud moments, your peaks were too hot, and you can read more in how to know if your audio is clipping.

RATHER SEE IT THAN READ IT?

Here is a real CutScore coaching report: loudness and true peak measured, the voice-versus-music balance flagged, with timestamps and the exact gain changes.

See a sample report
SHORT ON TIME

If you only fix three things.

Most of the jump from "thin and quiet" to "this sounds professional" comes from these three moves. Do them in this order.

1
EDITAUDIO
Get the voice clearly above the music
Loudness will not save a mix where the music is winning. Pull the bed down four or five decibels under the speech and duck it whenever you talk. On a phone speaker, intelligible voice beats loud music every single time.
How Solo the music, then bring it up under the voice until it supports rather than competes. See music louder than voice.
2
2-MIN FIXAUDIO
Normalise the whole Reel to about −14 LUFS
Measure your integrated loudness, then shift the entire mix to land near −14 LUFS. Quiet audio is the fastest way to look amateur and it has nothing to do with your microphone. This single move puts you at the same level as everything around you.
How Use a loudness normalisation tool set to −14 LUFS, or let CutScore measure it and tell you the exact gain change.
3
QUICKAUDIO
Cap your true peak at −1 dBTP
After raising the level, put a true peak limiter on the master set to −1 dBTP. That one decibel of headroom is what stops the upload re-encode from tipping a loud moment into crackle. It is the cheapest insurance in audio.
How Check the true peak readout on your meter, and limit anything that pokes above −1 dBTP. More on true peak.
How CutScore checks your Reel's loudness CutScore is an AI video quality coach for pre-publish QC. It measures loudness deterministically with an EBU R128 meter, reads your integrated LUFS and true peak, and compares them to the right target for short vertical video, so you are not eyeballing a number. It also flags when the music is louder than the voice and points at the timestamp. You get one score, the evidence behind it, and the exact gain change to make, before the Reel goes live. It judges the craft of the video, not its tags or thumbnail, so it sits next to a growth tool rather than replacing one. More on the method and the standards.
QUESTIONS

Frequently asked.

Master your Reel around −14 LUFS integrated, with true peaks held at or below −1 dBTP. Instagram normalises loud audio downward to keep the feed even, so anything much hotter just gets turned down anyway. Aiming near −14 LUFS gives you a confident, full sound that survives that process without distorting after the app re-encodes your file.
Yes. Instagram applies playback loudness normalisation so videos in the feed sit at a similar level, the same idea YouTube and TikTok use. That means mastering far above the target buys you nothing, because the platform pulls it back down. It also means a video mastered far too quiet stays quiet, since normalisation will not aggressively push weak audio up.
Keep your true peak at or below −1 dBTP. Reels get re-encoded into a lossy format on upload, and that compression can push a sample that sat right at 0 dBFS slightly over, which is where crackle and distortion come from. Leaving one decibel of headroom gives the encoder room to work without clipping your loudest moments.
Usually the integrated loudness is too low, often well below −14 LUFS, so even after normalisation it sits timid in the feed. The other common cause is the voice fighting the music. Pull the music down four or five decibels under the speech, raise the whole mix toward −14 LUFS, and check the true peak stays under −1 dBTP.
EARLY ACCESS

Stop guessing your loudness.

CutScore measures your Reel's loudness and true peak, flags the voice-versus-music balance, and tells you the exact fix, with the evidence to back it up. Join the waitlist for early access.

Join the waitlist