noob question - normalizing audio

[jonadams]

I have a file of spoken word, one word at a time one after another. I need to give all the words the same amplitude, or volume. So far the only way I have found to do this is to use the AudioSuite Gain feature to break the file up and make individual portions louder and less loud to normalize it, but this is time consuming. I tried the normalize function, to no success. Any suggestions?

[Dave Burris]

I have a file of spoken word, one word at a time one after another. I need to give all the words the same amplitude, or volume. So far the only way I have found to do this is to use the AudioSuite Gain feature to break the file up and make individual portions louder and less loud to normalize it, but this is time consuming. I tried the normalize function, to no success. Any suggestions?

Normalization will give variable results based on the formant content of the individual spoken phrases. You can reduce this effect by compressing the content before normalizing. Sometimes compression and limiting alone will solve your problem, but it can kill the dynamics, though less of a problem with spoken word.

[Tarktones]

Indeed, compression is what you're looking for. Normalize is basically just turning up the volume but it lets you specify the level of the highest peak. Compression will make the louder parts quieter and the quiet parts louder.

[Cory Cooper]

I agree...compression/limiting is a much better approach.

Here's a nice read about normalization:

10 Myths About Normalization (http://www.hometracked.com/2008/04/20/10-myths-about-normalization/)

C

[Andrew_Malloy]

That article had some good points about what normalization does. There are some points they left out. Peak normalization is WAY different than RMS normalization. RMS looks at the average signal strength to match the level.

This is different from compression because it does not color the sound or reduce dynamics. Inherently most compressors are going to distort slightly when they are driven harder so you can hear the difference between a louder bit of dialog and a softer bit by the harmonic content.

RMS normalization is really what you are looking for.

[Dave Burris]

RMS normalization is really what you are looking for.

...as long as you account for the extra headroom needed. However, with no dynamic range compression the initial problem will still likely be present, though reduced.

[jonadams]

thanks for the help, i guess i had the two mixed up