Isolating two voices

I have a a mobile phone voice recording of two people one person is three feet further away from the phone than the other. The voice of the person furthest away from the mobile phone appears quieter and although several words can be clearly understood from the person furthest away the whole conversation is not clear.

To try and improve on the conversation I’ve converted the AMR mobile file to an mp3 format for use in Audacity. Now I’m able to hear the conversation ever so slightly better, but the voice furthest away is still to quiet to make out every word.

What I’d like to ask is, can I separate each voice in the conversation and then edit the quietest voice to improve its volume and clarity. Can I then put both voices back together again so that the conversation makes sense?

Its my second day with Audacity soI would appreciate some basic instructions if this is possible please.

You can’t separate the voices, but you can “compress” the “dynamics” (dynamics = range of “loud” and “quite”) which will even out the volume (like an “Automatic Gain Control”)

There is a very simple “compressor” in the “Effects” menu. You will need to set a very short “attack” time, and a high “compression ratio”.

There is a much better, though more complicated compressor called “SC4” (Studio Compressor 4) available as a plug-in.

For more information about compressors:

There’s a couple of other techniques discussed here:

Thanks both, I’ve tried the compression tool in "Effects but although there is an “Attack” slider there is no compression slider as recommended in the instructions below. The chances are that it is there under a different name and I have simple not recognised it as a compression slide, Can some one advise me further please. I’ve already tried using the sliders in different positions and have inproved the conversation, just not quiet enough just yet, but I’m definately getting there. Just need a little more help.

Have a go with the SC4 - I think it’s in the main LADSPA plugin pack on the main Audacity site.
There’s a very brief guide to what each setting does here:

For a starting point, try these settings and then experiment:
RMS/peak: 0 (also try it set to 1)
Attack time: 5 (keep this at a low value)
Release time: 50
Threshold: -12 (assuming that the audio is already normalised )
Ratio: 12 (depends on how much compression you want - higher values for more compression)
Knee radius: 3.25
Makeup Gain: 0 (much easier to normalise after compressing than trying to guess how much gain to apply)

I cannot find the plugin for SC4 in the Audacity website, would you have a link to this please. Perhaps I should mention I’m a Linux (Ubuntu) user.

My apologies some further searching and I’ve found the SC4 package for Linux in the Linux repositories under swh-plugins. The information came from here.

Having spent most of the day trying different settings including those suggested here, I’m afraid that I’ve failed to make my voice recording any clearer. I have on ocassions been able to improved matter slightly which suggests to me that spending more time together with good recommendations may get me somewhere. My question is, where can I find instructions on using SC4 please, can someone share their experiences of using this plugin to allow me to edit my two voice recording so that both voices can be heard clearly?

As far as I know, the only “manual” for the SC4 is those brief notes in the link in my previous post. Try looking up “audio compression” or “dynamic processing” in Wikipedia, I seem to remember seeing an article about using compressors there.

Much appreciated.