Fixing/balancing out levels in (mono) recording?

I have a downloaded podcast of a phone interview and the interviewer is many times as loud as the interviewee. This makes it almost unlistenable. I was searching for a solution to my problem here and found this quote (unfortunately, I have just the one track, so it sounds like I might be screwed):

“What you have is highly desirable because with the voices split, you can apply corrections to one voice (volume corrections, filtering) without messing up the other. People with both voices on one track have no correction or filtering options.”

Just thought I’d check here to see if there was an easy way to make this interview a bit more tolerable to get through as the subject matter is fascinating. I’m not a total tech novice, but I’m not an Audacity super user either. :wink:

Thanks in advance for any suggestions you might offer!

Screen Shot 2016-03-07 at 12.36.04 PM.png

I have a downloaded podcast of a phone interview and the interviewer is many times as loud as the interviewee.

Right. That’s what happens when someone tries to use a personal recorder application on a cellphone. It only gets the host, not the guest. What of the guest you can hear is because of echo cancellation errors and acoustic leakage.

If the host talks over the guest, you’re dead. If the guest is talking by themselves, you can magnify that portion of the blue waves, drag-select the voice and effect > amplify: New Peak Amplitude -3dB. You have to do that to each low-volume snippet of the show. It’s possible even if you do that, you won’t get listenable audio.It will probably be honky, bubbly, fading in and out. In a perfectly working cellphone, this voice wouldn’t be there at all.

It’s not an interview. It’s a technical accident. I bet somebody was forced into doing this against their will, so they recorded it the fastest way possible and then posted it without listening to it. This is a very newbie mistake. Several people posted on the forum with this kind of damage. I think this is the first time somebody got all the way to posting a finished show that way.

It’s possible they still haven’t gone back to hear it. You should drop them a note.


Steve made a plugin for that called “levelspeech.ny”…

Did that work?


I did try an early version of the plugin which did improve matters.
Steve did subsequently produce other versions which I haven tried yet.

I’ve just tried the latest version [ at “90%” setting ] …

I mean KMG365.

Chances are really good “correcting” a show like this is just going to give pumping, honking trash.


Wow, thanks all for the very quick help! I’ll give it a go and report back. It’s not music, I’m just wanting the word content, so as long as I can get through it (understanding the soft stuff without blowing out my ears), it’ll be fine.
Thanks again! :slight_smile:

So you’re after transcription quality. Good enough to write down the words without producing an entertainment quality show. Yes, it might do that. Many people want us to rescue their show.

There’s almost no chance it will do that.


Worked like a champ, thanks all! :smiley: