Okay, I know I'm probably pushing sh*t uphill with this one, but I have a number of interviews filmed in loud environment - a lot of background chatter mainly. Mostly the subject can be made out above the noise, but at times it gets very difficult to understand what they're saying. I do not have the option of re-recording, so I was wondering if anyone has any decent techniques - within KDenlive, Audacity or otherwise - to just tease out one voice, even if ever so slightly. I've been messing around with the equalizer in Audacity for about 2 hours now with practically no success.

If the chatter is from other people talking, you're on mission impossible, as the noise has the same spectral/temporal signature as your desired signal.

Noise reduction only works when the noise differs significantly from the desired signal.

Yeah, that's what I thought... Thought it was worth a shot asking anyway!