I'm working on some vocals that are pretty dynamic (some parts too soft, others too hot). I plan to even things out a good bit via clip gain adjustments and some compression. But which should I employ first?...
 
I realize that doing the gain adjustments first would lighten the load on what the compressor has to do and help prevent compression artifacts and distortion, but there's going to be a compressor on the tracks at the end of the day anyway, so why shouldn't I use it first?
 
So, in the opinions of those more experienced than I, am I better off...
 

• Making gain adjustments throughout the tracks to get the dynamics relatively even, then add a compressor to clamp down on only the most extreme values?

             -or-

• Throwing a compressor on the tracks first, let it do its work, then manually adjust the gain of any remaining sections that are still at all out of whack?

 
I appreciate any insight/advice!

Good question. Ultimately compression is "riding the fader," so what you are describing is in essence two stages of compression. I am not sure if there is a right/wrong to either, just different approaches for both. What did come to mind is for something with a large dynamic range - rather than "extreme" compression in one or two steps is to use several stages of gentle compression instead. Have you considered that approach?
 
Another comment is "compression" often refers to downward compression (only), so the location of the threshold is important; but upward compression can be used in tandem with downward compression in gentle stages to reduce dynamic range appropriately.

Thanks for your reply, mettelus.
 
 

mettelus
What did come to mind is for something with a large dynamic range - rather than "extreme" compression in one or two steps is to use several stages of gentle compression instead. Have you considered that approach?
 

 
I hadn't considered that. Would you mind elaborating on this technique a little? Do you mean simply using multiple compressors, each with subtler settings?
 
 

mettelus
Another comment is "compression" often refers to downward compression (only), so the location of the threshold is important; but upward compression can be used in tandem with downward compression in gentle stages to reduce dynamic range appropriately.
 

 
Yes, I did have that in my game plan, to use compression not only to contain runaway peaks, but also to raise up quieter sections (if that is indeed what you meant).

I am a big fan of this and I do it all the time.  First off make a backup and only work on the backup.  I run a VU meter over the vocal track and edit any phrases that either fall short or overshoot.  Often it is only a few db of gain change but it can make the difference.  I don't even leave very loud phrases alone.  I bring them down so they are like maybe 4 dB or a little more above the rest.
 
Then I get the compressor on the vocal track.  But now the ratio can be low making the sound big, eg 2:1 Max or under.  It will sound great.  Maximum illusion, minimum voltage.  You can turn the vocals down now and still hear them clearly.  If you want a vocal section to add power here and there, I do it by manually riding the vocal level on mixdown.  It seems to add something nice.  You put the dynamics in later.  You are now manipulating a very even vocal track and in control.
 
Watch out bringing softer phrases up to 0 dB VU level though.  That can sound a little unnatural.  Best to let some phrases fall short by 3-4 dB or so.  You should hear it when it is right.
 
The VU meter was originally designed for voice, and it is still perfect on vocals too.  I do all the editing in an separate editing program. (Cool Edit Pro 2)  It is better and faster and I just drag the editied version into my arrangement.  Super easy.
 
Once this is done everything else vocal wise will fall into place much better such as eq and reverb. This applies to genres where this very even vocal style sits well.  I would not do it in other circumstances though.
 
 

On my last record, the mixing engineer used a Waves Vocal Rider.  While not as perfect as spending a ton of time editing each phrase, i was amazed at it's effectiveness.
It was recently on sale for $50, so I bought a copy.  Whether it's worth $249 is another story...
http://www.waves.com/plugins/vocal-rider#intro-to-vocal-rider
 

Melodyne excels with this. I make a copy of the vocal track and work on sections at a time. Each note shows up as a blob, the size is directly related to the amplitude of a given note and can be manipulated with the amplitude tool.
 
Rocky

I typically do this with automation on the fader.  This lessens the load on the processor.  It is tedious though.  For large fluxuations, compression is a must.  I prefer lighter compression on the vocals.
 

For vocals, you definitely want to automate or edit levels pre-compression. Like Jeff says, that allows for a lower compression ratio, which means more consistent and transparent results from the compressor. 
 

A LOT of stuff gets the life compressed out of it, and that for ME kills the enjoyment of some tunes - it is all squashed, generally to win the loudness wars, BUT at the expense of actual intentional building both upwards and downwards for effect.
 
One of my favorite albums is by Genesis, called Wind and Wuthering.  On this album, there are all kinds of GIANT variances in a given song's dynamics, and they used that with MASTERFUL precision.  They would create a quiet section, with delicate orchestrated piano and perhaps acoustic guitar (even a 12 string from time to time).  So the listener would get drawn into this quiet background section, and then they would eventually come back with a small number of measures where they would build the sound back up to seemingly HUGE loud levels, but really it was just brought back to the same levels as were there before the quiet section.  Those simple techniques CAN be extremely important and beneficial to turning a song or even a whole album into something VERY VERY special.
 
So, dynamics is another collection of techniques to use, when appropriate, or to NOT use, depending on the nature of the song.  A song to get folks out on the dance floor may well have a kicking beat and so that sort of song may well have squashed dynamics on purpose, to keep the song hopping, so that folks are drawn to dance to it.
 
The NICE thing about digital recording software, such as Sonar, is that you can try multiple approaches to mixing, and discard ones that don't sound the way you want them to, without permanently impacting what was originally recorded.
 
I hope you find the process of fiddling with sections and parts  to be both enjoyable, and a good learning experience.
 
For ME, I happen to both appreciate and to love properly applied dynamics.   That can be SOOOO powerful - Peter Gabriel is another artist who also embraces the use of dynamics - a single note can summon every fiber of his being, and convey so much emotion - I find that to be incredibly powerful.
 
Bob Bone
 

agree on Bitflipper's advise.
a fast attack/release compression then could tame the sharp transients.
personally would try the onboard Sonitus compressor, as there you can see when/what you're compressing.
the threshold/knee controls on the Sonitus could help you just touch those peaks/transients