Cutoff below a certain magnitude.


I’m facing with a challenge where I have to separate a male voice from a lot of background noise of vehicles (including the horn). There’s one thing common about these noises, they are not that much intense, so I was wondering why not just cutoff all noise below a certain intensity?

I tried the leveller, but it didn’t work out.

Unfortunately that tactic will only work when the noise is the only sound present, (in your example no voice present, i.e. the gaps between words), then you can use a gate to squelch the noise down.

Free noise gate plug-in for Audacity here …

If the horn is one frequency (and a few harmonics) it may be possible to reduce/remove it with a notch filter

Ok, thanks.

I’ve also changed the mic, it tends to pick up less noise.

I’ve also changed the mic, it tends to pick up less noise.

And that is the only way to do this job. You can’t pull the man’s voice out of the trash with software in post production. It will always sound terrible. Watch the newsies on TV. They use directional lavaliere microphones or directional hand-held microphones for interviews. In extreme conditions, you can use a headset system.,id,985,pid,985,nodeid,2,_language,EN.html

We have two of those and they will allow high quality voice recording and even stage announcing with no feedback and no room pickup.