I usually record my YouTube vids with separate sound recording chain, in the first place because the sound from the camera isn't too great (it never is, except possibly with that Zoom Q3HD camera which is more of a sound recorder with camera attached, and even then you don't have the best choice of mic position)
Having made that decision, there is lots of scope from improving the sound.
First, mic position. I usually use a pair in stereo, either XY (coincident at 90 degrees) or ORTF (slightly spaced and pointing away from each other, officially 17cm and 110 degrees) about 1m in front of me. The room is quite non-reverberant so I can get a clean sound even that far away. My two contributions this month were actually recorded in mono with a hypercardiod (AKG SE300B and CK93B) - the only stereo in in the added reverb.
I have sometimes edited mistakes - copying and pasting is easy and works well for folk tunes where nearly everything gets repeated several times, and nobody's yet said they noticed the discrepancy between the video and the sound where the edit was done

I used to cut highs and lows in a misguided attempt to prevent YouTube's awful sound processing from adding horrendous distortion, but I don't bother with that much any more, since I discovered that the best approach to that problem is to upload in HD format, which switches Youtube to a different processing regime. I still don't understand how some uploaders manage to get really clean and loud sound on 640x480 uploads.
I add some reverb because it's hard to resist the temptation not to...
I don't usually bother with compression, except sometimes to put in a limiter to knock down some peaks so I can raise the overall level by a couple of dB.
Perhaps some experimentation is called for. On my last two uploads, I quite liked the dry mono sound from a single AKG even though I did end up add some reverb. Perhaps I'll do something with several different options simultaneously sound-wise and upload the different versions for all to hear and compare.