View Full Version : Software technique to average varying degrees of amplitude in a track to one level?


James Palanza
December 10th, 2013, 09:34 PM
Okay not sure if I'm asking this correctly, I'm a tad weak in the audio department, but is there a nifty method in either Audacity, Premier, or Adobe Audition that would automatically take lets say, the softest parts of a track and boost them to the same level as the loudest? For instance, when two people are speaking into a mic and one is slightly farther away, I always have to manually go in and rubberband the one speaker up when they talk, and then back down for the other.

Any way to automate this?
Is this called "normalization"?

Jim Michael
December 10th, 2013, 09:47 PM
There is a way to automate this using normalize as a command line parameter in sox. SoX (http://sox.sourceforge.net/sox.html) There is also a normalize filter in Audacity, but I seem to recall something that was a levels adjustment that didn't have quite as strong or global effect. Maybe someone else has a suggestion for something a little more refined?

Nate Haustein
December 10th, 2013, 10:45 PM
Isn't this what a compressor filter does? Crank up the ratio and it should do what you're asking.

Also, I've used this in the past and it's pretty darn slick: http://web.archive.org/web/20130729204551id_/http://www.conversationsnetwork.org/levelator/

Duane Adam
December 10th, 2013, 11:01 PM
Compressor.

Chris Luker
December 10th, 2013, 11:23 PM
Compressor for sure.

Rick Reineke
December 11th, 2013, 11:07 AM
Peak normalization and RMS normalization are different. In some app.s (Sound Forge Pro for instance), The RMS normalize process can add dynamic compression. I much prefer to add compression with a good compressor plug-in where one has control over the parameters. Volume envelopes can be added to the timeline to manually normalize differences in amplitude as well.

James Palanza
December 11th, 2013, 11:23 AM
Thanks everyone for the suggestions, going to give it a whirl tonight.

Dave Farrants
December 11th, 2013, 12:49 PM
The Levelator from The Conversations Network (http://web.archive.org/web/20130729204551id_/http://www.conversationsnetwork.org/levelator/) works well on speech without any messing around.

Christopher Young
December 12th, 2013, 08:58 AM
The "Levelator". Will second that for sure. Used a lot in the radio industry. Export a .WAV from your timeline, run it through Levelator and bring it back in to your timeline, adjust final output level. Quick, clean simple.

Chris Young
CYV Productions
Sydney