why your video sounds quieter than everyone else's
short answer: the platform turned your audio down, or it left your audio alone while it turned everyone else's down to the same place. either way, exporting louder will not fix it. what fixes it is raising the average level of your audio, not the peaks, and there is one meter in your editor that shows you the difference.
the one rule that explains all of it
tiktok, youtube, instagram and spotify all run loudness normalization. the software listens to your whole video, measures how loud it is on average, and adjusts playback so it sits in the same neighbourhood as everything else in the feed.
the rule is asymmetric, and that asymmetry is the entire problem:
- too loud → turned down. the platform applies negative gain. your careful loud master gets quieted on the way in.
- too quiet → left where it is. most platforms will not add gain to a quiet upload. they leave it, and it plays back thin next to the video above it.
so a creator who exports at a sensible average sounds normal. a creator who exports quiet sounds quiet forever. and a creator who slams everything into the ceiling gets turned down to the same place as the sensible one, minus all the dynamics they crushed to get there.
peaks are not loudness (this is the part nobody explains)
your editor probably shows you a peak meter: the bouncing bar that hits red when a sound clips. peak is the single loudest instant in your audio. loudness is the average over time, which is what a listener perceives and what a platform measures.
picture two clips. one is a whisper track with a door slam in it. the peak meter hits the top on the slam, so the clip looks loud. the other is steady speech at a moderate level with no slam at all. the second one is far louder to a human, and to the platform's measurement, even though its peak is lower.
that is why "turn it up until the meter is nearly red" fails. you are optimising the one number nobody hears.
do this check on a video you already posted
before you change anything, find out what the platform did to a video of yours that felt quiet.
- open that video on youtube on a desktop browser.
- right-click on the player and pick stats for nerds.
- read the line labelled volume / normalized.
if the normalized figure is well under 100%, youtube turned you down and you were mastering too hot. if it reads at or near 100%, youtube left you where you were, and if it still sounded thin next to other videos, your average level was low. two opposite problems, one readout that tells you which one you have.
tiktok and instagram do not expose this, but the behaviour is close enough that a fix aimed at one works for the others.
the fix, in the order that works
1. get the dialogue right first. in short form, speech is the anchor. set your voice track so it sits at a comfortable, consistent level with the music pulled well under it. if you are unsure how far under, mute the music: your voice alone should already feel like a finished video.
2. compress the voice before you raise it. a compressor narrows the gap between your loudest and quietest words so the whole track can sit higher without the loud bits clipping. a gentle setting on a voice track does more for perceived loudness than any amount of fader.
3. put a limiter last. a limiter is a hard ceiling. set it a touch below full scale so the encoder has room, then raise the input into it until the average level comes up. this is the step that raises loudness without adding distortion.
4. leave headroom, do not chase the ceiling. anything you gain by pushing past a sensible average gets taken back by normalization anyway, and you paid for it with a flat, fatiguing mix.
5. check on a phone speaker. a laptop or a monitor flatters a mix. a phone speaker is where your video will be watched, and it exposes a voice buried under a bassy track immediately.
if your editor has no loudness meter
capcut and most in-app editors give you a volume slider and nothing else. you can still get this right by ear, using a reference.
open a video from a creator whose audio always sounds good, in the same app, at the same system volume. play a few seconds. then play yours. do not adjust anything else, only listen for whether the voice sits at the same level and the music sits the same distance underneath. keep nudging your music down and your voice up until the two feel interchangeable.
that reference habit is the cheapest audio training there is, and it is the same instinct behind building a sound library you own. when the sounds you work with live in one place, you start hearing your own catalogue as a standard instead of guessing every time.
the three-line version
peaks are not loudness. platforms turn loud audio down and leave quiet audio alone. raise your average with compression and a limiter, keep the voice on top, and check it on a phone.
