msmcleod
 
Making the signal more or less even before it hits any effects gets better results for me. It also means that I'm free to use volume automation later for real "mixing" duties rather than correction.
 

Normalize doesn't even out the signal at all. 
it simply allows you to achieve the maximum possible volume without distortion or clipping.
 
Unless of course you are splitting clips and normalizing each one of them separately.
in this case, I much more prefer to use clip gain automation.

Waves Vocal Rider does it for me... simply lightly season with compression after. 

I either use:
 
1. Clip gain automation
2. Waves Vocal Rider
 
Depending on the need of the track
 
I never use track automation unless I am simply looking for the final stage to have an "effect" of being lower in volume on purpose (i.e. move the vocal to the back of the bus instead of in your face), but even that is rare.

@ bapu, i found that interesting and helpful.

@Chuck, you don't have to normalize to 0. I normalize to -3 often when normalizing prior to mixing.

In my case, I'm mixing maybe 20 to 30 songs for the same band and building the template as i mix. Bringing things in and getting the levels to hit the fx with likeness between songs really helps me.

chuckebaby
Unless of course you are splitting clips and normalizing each one of them separately.
in this case, I much more prefer to use clip gain automation.

Yep, that's what I was doing.
 
The tracks invariably need splitting up phrase by phrase, and also any non-silence in the silence parts taken out in any case. So going through normalising each phrase one by one works fairly well, albeit time consuming.
 
But I only feel the need to do this when I'm recording my daughter - like I said, she's only just 6. The volume of the vocals is literally all over the place.

msmcleod
 
But I only feel the need to do this when I'm recording my daughter - like I said, she's only just 6. The volume of the vocals is literally all over the place.

Me and your daughter.  Although, I'm not 6 anymore.

I never got normalisation working in a way I found helpful and always just thought it was me not understanding the basic principle of what it does. So I gave up ages ago in favour of either manual riding, wave form editing, or my current favourite method, 'blob' editing in Melodyne, which I use extensively. For vocals, I try my hardest while tracking to get the singer to understand what I'm after. That helps a lot, but it's not always possible, or sometimes not even desirable, to get consistency. It's the same for bass. So, before I do anything else, I will bring up the track in Melodyne, change the tool to amplitude mode, and skip my way through the track listening, and adjusting volumes until I'm happy. Then I will work on timing, then any pitch editing, before bouncing the track back to clip. Then I start with compressors, then EQ, then any other fx, before moving on. In the case of riding the track, I think it's true that volume riding after the compressor sort of defeats the purpose. So, if you ride and record automation, then bounce to another track, you achieve your levelling without anything else in the path, then apply your FX. Then, to achieve more definition in your mix, you can ride the volume a second time.
 
Many ways to skin this cat. You just have to find a way that not only works for you, but makes sense to you!
 
cheers
 
Dave.

MacFurse
I never got normalisation working in a way I found helpful

me neither, i expected it to make my vocals sound normal

gswitz

@Chuck, you don't have to normalize to 0. I normalize to -3 often when normalizing prior to mixing.

I am aware of that.
I was pointing out that normalizing doesn't even out anything. It just raises the over all level to a specified amount.
 
Splitting up clips in sections to normalize is a lot of work. But then again I hand limit (using automation) all my sibilance Vs. using a de-esser so we each have our own methods. What ever works for us



 

I also use Melodyne to adjust the loudness of vocal takes and sometimes a vocal rider plugin (not Wave's one anymore, but DynaRide from TBProAudio, it's better IMHO). My latest technique combines the 2 methods, i.e. first DynaRide, then Melodyne.