Evening sound levels for conversation


I’ve got some voice recording where one person’s voice is much louder than the other.

I’d like to use Audacity to even the sound levels for these conversations.

How would I do this? Thank you.

You tried to record a cellphone conversation with the Personal Recorder, right? That will only record your local voice and the echo cancellation error of the other side.

Those are usually too far off to fix with the automatic tools like Compression and Limiter.

The desperation method is to use the zoom tools…


…and zoom into individual segments of the conversation on the timeline, drag-select the far person’s words and sentences and boost them with Effect > Normalize.

You’ll have to figure out what volume to use in the Normalize setting to sound, well, normal. There’s no way to tell ahead of time. You’ll have to experiment with it.

This is usually unsatisfactory because the background noise and cell trash is going to go up and down with your selections. In extreme cases, you won’t be able to identify the far sides words enough to select them. It also doesn’t work if you like to interrupt each other. As an interviewer, it’s good idea not to interrupt the guest.

When I’m forced to do a cell interview like that, I use an external recorder and a special microphone that goes in my ear. This is the Olympus TP8 (attached).

Sorry, I usually have a picture of my hardware in action, but I don’t this time.


Screen Shot 2015-06-22 at 19.37.34.png

Steve made a plug-in for Audacity called “LevelSpeech.ny” for that purpose … https://forum.audacityteam.org/t/improved-leveling-for-speech/33258/16

Alternatively try Steve’s Limiter plug-in … http://wiki.audacityteam.org/wiki/Nyquist_Effect_Plug-ins#Limiter

[ see here for how to install plugins into Audacity … http://wiki.audacityteam.org/wiki/Nyquist_Plug-ins ]

When using either of these plugins the quiet-bits will be made louder , but inevitably, they will be noisier [hissier] than the loud-bits.