Just an amateur perspective:

Unmatched levels are a matter of gain. Unmatched dynamics are a matter of either meticulous application of gain on a per-passage level (sometimes per-note level) OR... compression.

Choose your weapon. ;)

Just another amateur perspective :wink:

Did someone mess with volume knobs or level controls between takes? Or did someone play a bit more aggressive on some passages, and a bit mellower after an herbal break? :shock:

Several things you could try.

You COULD try mixing them all down and messing about. (Mute the original tracks while listening, of course). Just don't forget to keep the original tracks, if you want to nuke the the mixdown and do something else.

You could try adjusting gain of the tracks in question, doing compression/limiting/EQ on some or all, or a combination of all. Maybe find the track you like the sound of the best, and then try to get the others somewhere near it. When you think you have them REALLY close, then perhaps try an overall evening-out after mixdown?
Fiddle about, pan them where your mixdown will be, and listen. You'll have SOME idea how it will sound. Once you mix all the tracks together, it'll be harder to track down the individual elements. You might try doing it within and without the rest of the music. Without to see if anything really pops out. Within to see how it sits. Just remember not to nuke the original tracks so you can go back if you want.

Keep in mind that if you got a balls-to-the-walls track, and have tracks where the levels are lower, compressing/limiting/EQing all of them the same, if mixed down, is probably going to affect the sound of the loud track, and likely bring up more noise in the lower-in-level tracks. It's not real likely you'll get them all to sound matched very well, if just mixed down first. If you have to really bring up some tracks and compress them, you may find a lot more noise in those. If you were to do some noise reduction to get rid of the noise that has just been added, you'd probably want to do that on a track-by-track basis, after you have tried to get them all similar. If you just leave any newly generated, boosted noise in all tracks, and then mix them all together, you won't have a single source for an FFT noise file to work from. This means that if you try to do a bit of noise reduction after mixing them all down, you may remove some of the noise, but not others. So, you could have parts with virtually no noise in the signal, some with one type of noise, and some with another. Depending on how bad the noise is, and how well the rest of the signal masks the noise, this could be a problem, or not.

Just some thoughts. Correct me if I'm wrong.

Kapt.Krunch

Well the amp we're using has crazy dynamics and the differences are like the difference between heavy chords and lighter guitar lines. I understand that I'm looking at compression and gain. I know how to apply the compression but I wasn't sure on the best way to apply the gain. It's not like the settings on the faders carry through to when I merge the tracks. I was going to normalize the quiet parts but something tells me that that's horribly unprofessional and that it will greatly affect the tone. Yes?

The best way would be to record those tracks again, if possible. Otherwise: gain riding (volume curve) and compressor/limiter. From my point of view it's more time consuming and frustrating to fix this (you have to live with all side effects) than to do it again with better amp.