Audio normalization makes one or more selected audio clips louder, without creating distortion, by raising the gain (audio level) of each clip such that the loudest portion of each clip does not exceed the level that you specify. All video or audio editing software supports normalization.
This is an automated way of manually dragging the volume line up. But manually dragging the volume up risks dragging the volume too far, causing the clip to distort. Normalization prevents that distortion.
The big benefit to using audio normalization is that, while the overall sound becomes louder, you do not alter the dynamic range of the clip.
DYNAMIC RANGE: The amount of variation in audio levels within a clip. A heavy metal band has limited dynamic range (it’s all LOUD!). An orchestra has a large dynamic range during a piece (from very soft to very loud). In the past, normalization was a key technique used by music producers to make their music as loud as possible without distortion.
The benefits to normalization include:
If you are editing a concert video where dynamic range is important, normalization should be in your toolkit.
However, most of the time, as video editors, we don’t want or need to preserve the dynamic range. Why? Because most of what we edit is dialog/narration/spoken word, generally under less-than-idea conditions.
This is important because audio normalization can be easily fooled. If someone coughs, normalization thinks that’s part of the sound track and adjusts audio based on that level. Almost always, coughs are the loudest portion of a clip, so the level gets set to the wrong peak. The cough sounds perfect, but the rest of the dialog is too soft.
In other words, normalization works best for clips which are carefully recorded and don’t contain loud artifacts.
HERE’S AN EXAMPLE
Click to see larger image.
Here’s an audio clip with a cough. Notice that the loudest audio level is at the bottom (red arrow) and is about -1.5 dB. (Audio levels are displayed on the right side.)
Click to see larger image.
Here’s the clip normalized to -0.1 dB. Notice that the level of the entire clip has increased, The peak is now at -0.1 dB (red arrow).
Click to see larger image.
Here’s the same clip where the loudest element – the peak – is normalized to -3 dB. Again, the level of the entire clip changed.
LARRY’S THOUGHTS
If you have zero time, limited audio knowledge AND your audio doesn’t have pops, clicks, coughs or other loud artifacts, audio normalization is perfectly fine.
However, I never seem to have the luxury of audio that clean. For me, and especially for dialog, audio compression is a better option. Compression raises the level of soft passages, without changing the level of loud passages.
NOTE: Audio compression is not the same as data compression. Audio compression alters audio levels. Data compression reduces the size of a file.
For well-recorded music, normalization is the loudness option of choice because it preserves the dynamic range of the music with no risk of distortion.
For dialog, or audio that’s recorded under less-than-ideal conditions, audio compression is a better choice.
NOTE: In general, you should not apply audio compression to music where you care about the dynamic range, because compression will destroy it.
EXTRA CREDIT
Most NLEs normalize audio based upon peak levels. Some DAW software, though, normalizes on loudness. Increasingly, audio loudness is more important than peak levels (though, still, all peaks MUST remain below 0 dB to avoid distortion).
If you are normalizing clips based on loudness, use -14 LUFS as your target setting. This optimizes audio levels for YouTube and social media.
