When you record your audio, you want to get close to the peak volume (0db) as possible without going over.  You won’t always have complete control over that situation, and caution should always be employed when dealing with recording levels.  If your loudest peak is around -6db, then you have a bit of a safety net to save you from those popping ‘P’s and volume changes.

When you are finished with your recording, you will want to maximize the volume of your tracks.  The simplest and most effective way to do this is by using normalization.  Normalizing your audio raises the volume of your audio evenly, so that the highest peak is as loud as possible (0db, or a little less if that is what you want).  Normalizing does not change the dynamic range of your audio (the differences between the loud and soft parts), it just raises the volume of the file.

Lets see what that looks like in Audacity:

This is just a conversation in Skype.  You can see how low the audio is, not getting close to the top volume.  This is great as a safety net, since you can’t undo an overloading signal, but not ideal.  First thing I want to do before doing any more processing or editing is to normalize each track.  From the menu, select Effect > Normalize, and you will be given this dialog box:

I usually set the db to -0.1db, just because it’s a safety net.  It probably isn’t necessary, but I know that I’m not the only one who does it.  You should keep the DC offset box checked. For an explanation of DC offset, here is a dull Wikipedia article (it’s short but boring).

One thing to remember about normalizing is that it is a destructive process.  That means that changes you make here will be permanently written to the file.  You can not undo this process.  But since the character and sound of your file isn’t being changed, it shouldn’t make a difference.  If the file is too loud in the mix, that is what the mixer is for.

Press OK, and viola!

You can see how the audio is much louder, but also that it is even across the board.

Here is an A-B comparison. The top is the original file, the bottom is a copy of the same file, only this one was normalized.

Now the file is ready to be processed with effects like compression and de-essing.

Keep in mind that everything in the file will be boosted in volume.  If you record at too low a level, or speak too quietly into your mic, your noise floor will be more obvious.  If you peak out at -15db, you are going to be raising the volume of the underlying noise more than if your audio peaks at -6db.  In other words, normalizing isn’t going to completely compensate for low recording levels.  This isn’t a cheat as much as it is a useful tool.

Hope this helps.  We will get into other effects in another tutorial soon.