Normalize audio online with a free LUFS meter.
To normalize audio, choose your file and see its loudness in LUFS and its true peak right away. Then pick a target: YouTube, Spotify, podcast or broadcast. We set the gain, hold the peaks at -1 dBTP and you download MP3 or WAV with a before and after report. Free, no sign-up.
Free, no sign-upLarge file: reading it can take a few minutes. Keep this tab open until it finishes.
Measured loudness
- Integrated loudness
- - LUFS
- True peak
- - dBTP
- Loudness range
- - LU
- Duration
- -
| Measure | Before | After |
|---|---|---|
| Integrated loudness (LUFS) | ||
| True peak (dBTP) | ||
| Loudness range (LU) |
Loudness sorted. What about the transcript? We transcribe it with speakers and chapters.
How to normalize audio online
- 01
Choose the file
Drop an MP3, WAV, M4A or a video on the area above. Integrated loudness in LUFS, true peak and loudness range show up right after.
- 02
Pick a target
YouTube and Spotify at -14 LUFS, Apple Podcasts at -16, mono podcasts at -19, broadcast at -23 or your own value. The peak ceiling stays at -1 dBTP.
- 03
Download the result
Click normalize audio, check the before and after table and download MP3 or WAV.
What loudness should I use for YouTube, Spotify and podcasts?
Platforms measure loudness in LUFS, a scale that follows how the ear perceives sound, and adjust playback toward a target. If your audio is above it, they turn it down; if it is well below, it sounds weak next to everything else. That is why it pays to publish close to each target, with the true peak under -1 dBTP.
The mono podcast target is 3 LU lower because one channel played in both earbuds sounds like a -16 stereo file. The numbers in the table are the reference values published by the platforms and by the EBU.
| Destination | Target | Max peak |
|---|---|---|
| YouTube | -14 LUFS | -1 dBTP |
| Spotify | -14 LUFS | -1 dBTP |
| Apple Podcasts | -16 LUFS | -1 dBTP |
| Podcast, stereo / mono | -16 / -19 LUFS | -1 dBTP |
| Broadcast (EBU R128) | -23 LUFS | -1 dBTP |
Making audio louder without distortion: gain, true peak and the limiter
Making audio louder means adding gain. The trouble is the peaks: a clap, a laugh or a hard consonant hits the top before everything else and clips. So we measure the true peak, which also accounts for what happens between samples, and never let the result go past the ceiling.
With the limiter on, the few peaks that would cross the ceiling are held down for a few milliseconds and the rest of the audio comes up untouched. With it off, the level stops rising once a peak touches the ceiling. When the target would take more than 12 dB of peak limiting, we stop short and the report tells you by how much.
- The gain applied and how hard the limiter worked appear in the result.
- If the audio is too loud, it gets turned down, with no limiting.
- Background noise rises with the voice: normalizing does not clean audio.
Online LUFS meter: what each number means
As soon as the file opens, you see three measurements. Integrated loudness is the average level of the whole recording, in LUFS, computed the way ITU-R BS.1770-4 and EBU R128 define it: K-weighting filter, 400 ms blocks and gates that ignore silence and very quiet passages. That is the number platforms compare with their target.
True peak, in dBTP, is the highest point of the signal, estimated with 4x oversampling. Loudness range, in LU, shows the spread between quiet and loud parts: a well recorded conversation usually stays under 8 LU, while a church service with worship and a sermon sits well above that.
Using normalized audio in Premiere, DaVinci, CapCut and on YouTube
Normalize the final audio, after editing and mixing, not each loose clip. If the sound goes back into a video, download WAV and import it in Premiere Pro, DaVinci Resolve, Final Cut Pro or CapCut in place of the original track. Then export the video without touching the level again.
On YouTube, publishing louder than -14 LUFS does not make your video louder: the platform turns it down. For podcasts, upload the MP3 straight to your host. If you edited after normalizing, run the final file through here once more just to measure it.
More than an audio normalizer. Once the level is right, the content.
See also
Frequently asked questions
How do I normalize audio online?
Choose the file on this page, check the measured loudness and pick a target, like YouTube at -14 LUFS. Click normalize audio: we apply the gain, hold the peaks at -1 dBTP and you download the result as MP3 or WAV. It is free, with no sign-up.
What LUFS should I use for YouTube?
About -14 LUFS, with the true peak below -1 dBTP. YouTube turns down videos louder than that, so pushing the level only adds distortion. A video well below the target is not turned up and sounds weaker than the rest, so it is worth adjusting before you upload.
What LUFS should I use for podcasts and Spotify?
Spotify uses -14 LUFS as its reference. Apple Podcasts recommends -16 LUFS, give or take 1 dB, with the true peak at or under -1 dBTP. For a mono podcast, the usual target is -19 LUFS, which sounds the same as -16 stereo in earbuds.
What is true peak (dBTP)?
It is the real peak of the signal, including what happens between samples when the audio is turned back into sound. It can exceed the sample peak and clip on a phone or in a platform's encoder. That is why the standard ceiling is -1 dBTP rather than 0.
Why didn't my audio reach the loudness I chose?
Because reaching the target would cross the peak ceiling. Without the limiter, the level stops when a peak touches the ceiling. With the limiter, we only stop when it would take more than 12 dB of peak limiting, which distorts. The report shows the final value and the reason.
Does normalizing improve audio quality?
No. Normalizing changes the level; it does not clean the sound. Background noise, echo and hiss rise with the voice. What improves is the listening experience: your audio sits at the same level as other videos and episodes, so nobody has to reach for the volume button.
Is there a watermark or a usage limit?
No watermark, no sign-up and no usage cap. Measure and normalize as many files as you like. Long recordings, one or two hours, work too, with the progress shown as a percentage.
The level is right. Now the transcript. With speakers, chapters and snippets.
Upload the episode, the service or the lecture. We transcribe it word by word, tell the speakers apart, create chapters and you download TXT, PDF, SRT or MP3.
See the result on your own recording before adding a card.