Excuse the dumb question but have you already done the basics like figuring out what else exists in the track around the frequency range the vocals need to live in and used eq on those other tracks to carve out space for the vocals?

Try some of this!  seems pretty decent so far with my limited fiddling.....but try it on its own to see what its really doing before pouring on the other bells n whistles ;o)
 
http://www.jhudstudio.com/free-plugins
 

This thread is also about something very similar. Lower level voice overs, but the same applies to low level vocals. I like to edit the vocals first before hitting any processes.  It goes a long way to ensure all vocal phrases are heard nicely.
 
When vocal notes go down low it is still OK to edit those phrases up in volume but not to the rms ref level. 3 db lower to ensure things sound natural. In pop music you may be inclined to even things up a little more but in other genres you may want to retain more dynamic range.
 
http://forum.cakewalk.com/How-to-bring-up-quieter-parts-of-spoken-dialog-m3445444.aspx
 
Even more importantly is what the other parts are doing while the vocals may be low in pitch. Are there 20 other things playing at the same time because you won't hear them. Edit the music and keep things sparse in the verses. Mix clarity is tied to arrangements as well as processing. More so maybe.
 
 

Jeff, your link doesn't work for me.

Graham Cochrane over at the Recording Revolution has some Vocal suggestions here.  I have found his web site very helpful and eye opening.
http://therecordingrevolution.com/?s=Vocals

Great read and site there....thanx  pat  ;o)

Gain automation is a good way to go.  Right at the start of the channel and so is cutting vocals up into sections and manipulating clip gain or event gain in Studio One.  Note the fact the waveform height should change here is important and makes this type of editing approach much easier.
 
Gain automation and editing the vocal ends up sounding exactly the same.  Except I find the editing approach faster than making automation.  Also you can get in and do surgical edits easier and faster in an editor.  I might lift a whole phrase that has 3 components in it up by 6 dB.  Then go in and fine tune one or two of the individual components up another 2 or 3 dB.  Makes a world of difference to the vocal.
 
When you drag the edited vocal out into the arrangement you also see the new waveform heights too.  Makes it easier to see.  I go in and cut the vocals up here and there into separate events in fine tune the gain on some phrases further.  Trim waveform heights here.  At this point a VU meter looks perfect over the whole vocal track. Still rich in movement but always riding up to 0 dB VU and rarely going over.
 
Do all this before the compressor goes in and it can be mild compression. Also the double compression mode can also consist of limiter followed by compressor.  The limiter tames just the odd peak down followed by lower ratio relaxed compression after.
 
I also ride this total nicely produced vocal sound too manually during mix down.  Adding in that human element with a few vocal moves can also make a world of difference.  When choruses kick in you can just manually ride the gain up 1 or 2 dB. The vocals often need it to survive.  They need to breeze back in level in the quieter passages.  You can do this manually.  It is fun and adds a hidden quality to the vocal sound.  It is nice to be holding a real fader for this.