Topic: Batch Normalizing Audio Files?

[mattmiller]

I utilized my gear to record some co-workers deliver voice-over for a marketing DVD that my company is putting together.  There was about 11 pages of text, so we basically split it up into paragraphs and recorded them one at a time, with many retakes along the way.  I've finally finished stitching together all of the good takes, but it's apparent that the volume varies across the files (I've got about 50 files).  Where I had to stitch takes together (which usually was just inserting a re-recorded sentence into the original recorded paragraph), I made sure to eyeball it to make sure that the portions that I added matched the rest of it.  But I'm still left with 50 files that I know have some variation AMONG them.  Is there a way (using Cool Edit Pro, preferably) to have it analyze all the files and normalize them all to a level that I define?  And if so, would I want normalize using the Average RMS or the Total RMS (I'm not sure of the difference).  All the normalization that I've ever done has just used Effects > Amplitude > Normalize, and, of course, that gives only the option to normalize based on the peak amplitude.

If Cool Edit Pro won't do this, but some other tool will, let me know.

[it-goes-to-eleven]

Screw the gui junk.  Use normalize.  It'll easily do what you want.  It might be more tricky to do in windows, but a breeze in linux.  It is a great tool when I want to quickly produce a new version of a show for playback in a car.  It can compress or just boost levels.

normalize *.wav

There are options to allow normalization based on peaks or avg amplitude.  Depending on how you use it, it can compress to achieve the desired levels.  It reports what it does for each file.  It will report what it will do, without actually doing anything.

http://normalize.nongnu.org/

[mattmiller]

Screw the gui junk.  Use normalize.  It'll easily do what you want.  It might be more tricky to do in windows, but a breeze in linux.  It is a great tool when I want to quickly produce a new version of a show for playback in a car.  It can compress or just boost levels.

normalize *.wav

There are options to allow normalization based on peaks or avg amplitude.  Depending on how you use it, it can compress to achieve the desired levels.  It reports what it does for each file.  It will report what it will do, without actually doing anything.

http://normalize.nongnu.org/

I'm not so good with non-GUI software.  I downloaded it and ran it on a couple of test files (just as you specified:  normalize *.wav) and it seems to have just amplified everything based on the peak amplitudes.  The average RMS values of the three test files I used are no closer to one another than before.  How do I access the options that you spoke of?

[it-goes-to-eleven]

The default is to normalize based on average levels, with a target of -14 by default.  If necessary, it will compress peaks to do that.

--peak will adjust based on peaks, but that isn't always appropriate if you've had any overs in the source material.

There are a bunch of options on the manual page.

[seanay]

You can take all the individual audio files you want to use and put them in multitrack in Cool Edit Pro, then go to Edit > Group Normalize Waveforms.  It will give you the option to analyze the sound of each audio file, which would then allow you to A) normalize to the average, or B) set your own normalize point.  By analyzing all the waveforms, it will also show you where clipping will occur if you turn the limiter off.

[mattmiller]

You can take all the individual audio files you want to use and put them in multitrack in Cool Edit Pro, then go to Edit > Group Normalize Waveforms.  It will give you the option to analyze the sound of each audio file, which would then allow you to A) normalize to the average, or B) set your own normalize point.  By analyzing all the waveforms, it will also show you where clipping will occur if you turn the limiter off.

That sounds promising.  I'll give it a try tonight or tomorrow.