Hi guys I recorded a song in 3 passes all with the same input volume, settings et cetera . Into the 4th track I cut and paste it the best versions of each line or 2 that I sang. Even though everything is pretty equal there might be just the slightest bit more attack on one line than another. Is normalization the way to smooth that track out or should I just do it the man's way and increase DB on particular clips as necessary? How about limiting? Or a leveling app plug in? Thank you.
WSS

There is a dozen or more ways to work and each of us has our own.
Save the project!!!! Back it up too!!
 
1st method for super important tracks: 
 
I will highlight the 4th track and then tool copy into Wave Lab.  
 Then using visual I first look for parts that are spiking and apply negative gain to even out. 
Then I will normalize it to around -2 db ( see note on Normalization)
Now I look and listen to the track and apply gain to any quiet parts as well as remove breathing and any artifacts.  This takes experience and practice to work efficiently. 
I save it and return to Sonar. You now have a vocal track that with a little bit of compression will cut through any mix on every lyric. 
 
 
2nd method
If it's not needing much work I will just use the volume envelope and a compressor.  
 
To answer your question about normalization,  It does not work like a compressor. It is pure gain.
Wave Lab will analyze you track and tell you your peak level.  If this is say -12db you can either do the math and apply 10db of gain or Normalize to -2 db. Both will result in a peak level of -2db. 
Normalization analyzer  will read 0db if there is one tiny little over or peak at 0db. The rest of the track might be -12db so therefor you need to get rid of any peaks if you want good results. A lazy person could just put a fast attack compressor on the track too to catch that one peak, but I bet the little red light will come on every time you play that part o th song. 

I don't Normalize anything, some swear by it. Me personally, Its destructive and anything destructive I tend to stay away from. Similar results can be obtained non destructively using Automation.
I use compression, limiting and I also use automation <- A lot.
I prefer Automation vs. limiting/compression because to me it sounds more natural (like I used to do riding faders with my fingers).
However the mastering stages is a totally different animal and I use tons of compression and limiting.
 
Its personal choice my friend. these are just my opinions.

Wave lab! Yes. I was going to bounce to track or clip and yes clone and save everything before I screw something up.
Best
WSS

I been "destroying" audio since 2004..  Your rendering at 32 bit and I dare anyone to point out a degradation in the audio. That's a myth that needs to go away. 

Just as a note, Sonar will show/mark your peak level as you play the track. That way you can see where the biggest peaks are at least.  If things are really off you can apply clip gain or destructive gain to even them out. Once they are pretty close then compression and limiting are the usual ways to go.

Can SONAR normalize to RMS level ?
 
Kenny 

Free: automation and compression.
 
Monies: I like to use Waves' "Vocal Rider." It works wonders.
 
So typically I will use Vocal Rider first, then do some automations to further fine tune it, and then compression if I need it.  But Vocal Rider does a lot of compression to begin with.
 
This is from the stand point of experimental musician, so YMMV :)
 

kennywtelejazz
Can SONAR normalize to RMS level ?
 
Kenny 

Simply put, no. Normalizing uses peak detection, not RMS.
It boosts the volume until the maximum amplitude is reached somewhere in the data.
So essentially what it does is find the highest peak and raise the data to that highest peaks maximum.
(maximum possible volume without distortion or clipping)
Thus there is no compression or limiting done, it is more of a "Squeeze every once of juice" out of a wave file.
 
Quieter passages are better done in smaller segments as trying to normalize a wave file that has loud and quiet data is not recommended. your better off splitting the file and processing each separately.
 

chuckebaby

kennywtelejazz
Can SONAR normalize to RMS level ?
 
Kenny 

Simply put, no. Normalizing uses peak detection, not RMS.
It boosts the volume until the maximum amplitude is reached somewhere in the data.
So essentially what it does is find the highest peak and raise the data to that highest peaks maximum.
(maximum possible volume without distortion or clipping)
Thus there is no compression or limiting done, it is more of a "Squeeze every once of juice" out of a wave file.
 
Quieter passages are better done in smaller segments as trying to normalize a wave file that has loud and quiet data is not recommended. your better off splitting the file and processing each separately.
 

 
Hey Chuck,
Thanks for taking the time ,
That's what I thought . Let me rephrase it then , even though I already think I know the answer since I just checked .
Can SONAR export a normalized mix while having it set to an RMS Level I can set myself just like what is in my pic ?
This is something I wish SONAR could do since I have found this approach  to be very useful when it comes to normalizing audio .
 

 
just a thought ,
 
Kenny