How to normalize audio (and why your tracks sound uneven)

You have felt the problem even if you have never named it: one song plays at a comfortable level, the next one blasts, and you lunge for the volume. The same happens across podcast episodes, voice memos and clips stitched from different sources. Normalizing is the cure — it brings everything to a consistent, deliberate loudness so your listener never has to touch the dial. But it only works if you normalize the right way.

What does normalizing audio actually do?

Normalizing measures the level of a file and applies a single, uniform gain change to move it to a target. Crucially, it does not squash the dynamics or change the mix — a quiet passage stays proportionally quiet, a loud chorus stays loud. The whole track simply moves up or down together until it hits the target you asked for.

That makes normalizing different from compression or limiting, which reshape the relationship between loud and quiet parts. Normalizing is the gentlest possible loudness tool: one number applied evenly. Because of that, it is safe, reversible in spirit, and exactly what you want when several files just need to agree on how loud they are.

Why do your tracks sound uneven?

Uneven loudness comes from the files being mastered or recorded to different levels. A song from a modern loud master, an old quiet transfer, and a phone recording all carry different amounts of energy, and nothing lines them up automatically unless you tell it to.

• Music from different eras and labels was mastered to wildly different loudness.
• Voice recordings depend on mic distance and gain, which vary every session.
• Clips merged from several sources each bring their own level.
• Peak-based exports can measure "full" while sounding quiet, because peaks and loudness are not the same thing.

If you are stitching clips together, normalizing each one first is the difference between a seamless result and a rollercoaster. Do it before you merge audio files so the joined track holds one consistent level throughout.

Peak vs loudness normalization: what's the difference?

This is the single most important distinction, and getting it wrong is why so much "normalized" audio still sounds uneven. There are two methods that measure completely different things.

Peak normalization only looks at the highest single peak and scales the file so that peak sits just below the ceiling. Two tracks peak-normalized to the same level can still feel very different in loudness, because a dense, compressed track carries far more energy than a sparse one with the same peak. Loudness normalization, which measures perceived loudness across the whole file, is what actually makes tracks feel even.

What is LUFS and what target should you use?

LUFS stands for Loudness Units relative to Full Scale — a measurement built to match how humans perceive loudness, standardised in the EBU R128 and ITU-R BS.1770 specs. It is the number the streaming world runs on. Lower (more negative) LUFS means quieter; higher (closer to zero) means louder. Pick a target based on where the audio will live.

How do you normalize a file?

In practice it is a few clicks. Load the file, choose loudness normalization, set your LUFS target, and export.

1. 1Open the normalize audio tool and drop in your file — it processes on your device, nothing is uploaded.
2. 2Choose loudness (LUFS) normalization rather than peak, unless you specifically only want to prevent clipping.
3. 3Set the target: -14 LUFS for music, -16 for a podcast, -23 for broadcast.
4. 4Export, and repeat with the same target for every file in the set so they all agree.

If a file just needs to be a bit louder or quieter without a formal loudness target, adjust audio volume applies a simple gain change instead. And when the audio comes from a clip, you can extract audio from a video first, then normalize it so it matches the rest of your project. Checking the result with audio file info confirms it landed where you meant.

Frequently asked questions

Will normalizing make my audio distort?

Loudness normalization applies even gain and includes a peak safeguard, so it will not push the file into clipping. If a target would require raising peaks past the ceiling, a true-peak limiter or a slightly lower target keeps it clean. Distortion comes from careless peak boosting, not from sensible loudness normalization.

Should I normalize before or after other edits?

Do your edits, trims and mixing first, then normalize as one of the last steps, so you set the final loudness on the finished audio. If you normalize early and then keep editing, the level will drift again.

What LUFS target should a podcast use?

Around -16 LUFS is the widely used target for stereo podcasts, with -19 sometimes used for mono. It sits a little louder than music streaming because speech benefits from being clearly present, while still leaving headroom.

Is peak normalization ever the right choice?

Yes — when your only goal is to maximise level without clipping, such as bringing a very quiet recording up to near full scale. For matching how loud several tracks feel, though, loudness (LUFS) normalization is the correct tool.