[Sasquatchian]

I think a combination of mild compression and volume automation would be a good thing. I'm really liking using a small amount of compression at the recording stage - maybe 3:1 or so and have been loving both the pre-amp and compression in the UAD Neve 88SR Channel Strip Unison plugin at the recording stage. Simply a great sounding pre-amp emulation. After it's been recorded I love both Fab Filter compressors - the single band and the multi-band for clean compression and I really love what the UAD Fairchild 660/670 (mono/stereo) plugin can do to add real warmth and glue to your mix, either on a single track or on the stereo bus. I learned more about compression watching youtube videos on that Neve 88SR Channel Strip than anywhere else. Finally, if it needs it, the Fab Filter limiter set to limits peaks to minus 1db is a really great limiter that sounds amazing.

I would not use any sort of automated normalization where you don't know exactly what is happening. You never know exactly what you're going to get and you're going to be better off using a combination of compression and limiting - compression first, then limiting to get the best results. The idea behind that is to let the compressor get most of but not all of the dynamics and let some of the peaks through, then you only need just a tad of limiting to finish it off and nothing it too heavy handed.

One of the best sources I've come across has been Ian Shepherd's The Mastering Show podcast, which, even though they're aimed at people wanting to learn mastering, are extremely applicable to recording and mixing.