Not good news, I'm afraid...
The mic picked up my talent's voice just fine, but unfortunately also picked up a ton of extra chatter from the crowd surrounding us. I believe this was my fault, as I accidentally turned up the mic's "volume" too high on the camera itself.
First off, if you just had the one mic, it would have picked up the same relative levels of noise whatever the settings. Your problem is that wherever you put the mic, it wasn't close enough to the talent's mouth; you needed more of the talent and less of the crowd, and that can only be fixed by positioning the mic correctly, or using possibly a more appropriate one. Yes there are some; a typical headset mic of the type actors use will work a lot better, because it can be positioned much closer to the mouth, and is still not visually that disturbing.
I have no choice but to bring the audio into audition and try to eliminate basically everything but my talent's voice from one audio track. Can anyone help me figure out how to do this? The only tutorials I've seen so far just describe how to get rid of random noises like sirens or something.
Noise reduction works best on continuous, fixed-spectrum noise like fans, etc. And unfortunately, that's just the opposite of chatter. Your problem is that you want to reduce something that has the same spectrum and qualities of that which you want to keep. Do you see the problem? How you gonna distinguish? It's relatively easy to paint out sirens, etc because they are short sounds and have a spectrum that is easily distinguished visually - so it's easy to 'paint' them out, the same way you might with a blemish on a photograph using Photoshop. And that's why they're used as demonstrations - they work well! The reason that you won't find a demo of what you're trying to do is basically, well, because it's not possible; CSI isn't real, however convincing it seems.
If the talent's a little louder than the surrounding chatter, then you might get away with a little downward expansion, but generally this sounds weird and isn't easy to set up effectively either. Okay, we haven't heard a sample but if it's the way I think it is, I'm afraid you don't stand much of a chance of making any significant improvements at all.
Thank you so much for your input. I understand completely what you mean, and I was beginning to suspect that was the case. I have decided to re-do my talent in front of a green screen, using b-roll of the event, but now I know for future reference. Thank you again!
Sent from my Verizon Wireless 4G LTE DROID
The other option is ADR, Automatic Dialogue Replacement. Audition has a special feature for matching up overdubbed dialogue with the original. For this you need to record the talent in good quiet acoustic with him listening/watching the original video. You may have to repeat this several times over to get the best match or do it in smaller sections. Then you can use the Automatic Speech Alignment effect under the Clip menu to match your new dialogue with the original.
I was about to mention the same thing...certainly my first choice would likely be ADR since doing it with green screen would involve matching the lighting to look like what's happening in the background...not always easy. ADR is a pretty common technique in most films and can be undetectable once you mix in some location atmos.