What Does Audio Normalization Do — and Should You Use It for Voice-Over?
Nov 26, 2023Audio normalization is one of those features that looks simple but is often misunderstood.
At its most basic, normalization changes the gain of an audio file so its highest peak reaches a target level. It does not compress the audio, change its tone, or automatically make a recording sound better.
That makes normalization useful in some situations—but it can also create problems when it is used as a substitute for setting the recording level correctly in the first place.
In the video below, I explain:
- what audio normalization actually does
- how peak normalization works
- how it differs from changing your recording input gain
- when normalization can be useful
- when it may be the wrong tool
- how dynamic voice-over performances can complicate normalization
- why proper recording level should come first
I also demonstrate the process in both Adobe Audition and Logic Pro so you can see exactly what is happening to the audio.
What Does Normalizing Audio Actually Do?
Normalization looks for the highest peak in the audio you select and raises or lowers the entire selection until that peak reaches the target level you choose.
The important part is that everything moves together.
If you normalize a recording upward by 3 dB, the entire selected recording increases by 3 dB. The relationship between the loud and quiet parts does not change.
Normalization also does not change the tone of the recording. It is fundamentally a gain adjustment, not EQ, compression, or mastering.
That is exactly how you explain it in the video: normalization identifies the highest peak and moves the entire selection to the chosen target without altering the tone of the audio.
Should You Normalize Voice-Over Recordings?
Sometimes—but I would not use normalization as a replacement for recording at the proper level.
Normalization can be useful for making relatively small gain adjustments after recording. If your recording only needs to move a few decibels up or down, normalization can be a convenient tool.
But if you recorded far too quietly and need to raise the file by 6, 8, or 10 dB afterward, the better solution is usually to correct the input gain and record it properly.
A good recording starts before the audio reaches your software.
This closely matches the position you take in the video: a few dB of fine-tuning can be reasonable, while large corrections are better solved at the recording stage.
When Normalization Can Cause Problems
This is one of the strongest parts of the video and should absolutely appear in the written article.
For example:
Normalization uses the highest peak in the selected audio as its reference.
That can become a problem with highly dynamic material.
Imagine most of your voice-over is recorded at a comfortable conversational level, but one word or phrase is dramatically louder. If you normalize the entire file based on that single loud peak, the rest of the recording may still end up much quieter than you want.
In cases like that, you may need to divide the performance into sections and adjust those sections independently rather than treating the entire recording as one uniform clip.
Your video demonstrates exactly this with both a single extreme transient and a performance containing normal speech, a whisper, and a much louder delivery.
Want to Stop Guessing About Your Recording Level?
Normalization can help adjust a recording after it has been captured.
But the better goal is to record your voice at the right level from the beginning.
My Optimal Recording Levels for Voice-Over training shows you how to set your input gain properly, understand your meters, and create a consistent recording level before you ever need to correct it afterward.
If recording levels, gain, meters, or normalization still feel confusing, this course was designed to make the entire process much easier to understand.
LEARN THE RECORDING LEVEL PROCESS →
Not Sure If Your Recording Level Is the Real Problem?
If your recordings still sound noisy, weak, inconsistent, or difficult to process and you're not sure why, send me a short raw audio sample.
I'll personally listen to it and help identify what may be affecting the recording.
