I thought the purpose of gain-staging was just to avoid issues like this: 1-Everything is pushed up so far that the final fader on the mix has to be pulled down almost all the way, or 2-Everything in the path is so low that the fader has to be pushed all the way up during the mix, or 3-tracks are so different than you can't get the faders high enough on some tracks or low enough on others.
 
This was a lot bigger problem in analogue days when we tweaked tape machines so they were really recording at +6 or +9 when the meter showed 0.  Recording at moderately strong levels, I have not had many issues in the digital realm, although if you had a long chain of plug ins, I can see how it would need to be managed.

The LANDR article referenced by the OP argues for gain staging right from the start at conservative levels.
 
They use the example of VST FX which may simulate analog devices and distort by differing amounts depending how hard you hit them with the signal.
 
They point out that if the master is in the red and you need to back off all your gains, you will also be changing tone in unexpected ways.

I use Slate VMR in all of my mixes, and I found that simply using the Trimmer module (sometimes two or three per instance of VMR) is a very simple way to maintain levels and ensure you're driving the other modules as you intend to get your desired effect and color. I always put Trimmer at least first and last in the VMR chain (sometimes in between as well if a certain module added a lot of gain - such as an EQ). 

I often record for a whole night where the band just plays through. Songs of different levels and intensities are played gains fixed for the loudest most intense moments.
 
Then I mix. I mix the first song and normalize all tracks to -3 dB and set compressors and so forth based on that level.
 
I touch bus gains too, reducing some as necessary to keep them from getting too hot as necessary. If I ever see a PC light blink, I reduce inputs or gains.
 
I make a copy of the project for the first track to use for the second track.
I split the clips in the full project that contains all the audio and bounce to clips then copy those clips to the new project. The new project has all the compressor settings and everything else, but now the tracks are all softer or louder than the last one. I can now normalize all the tracks to -3 dB. This isn't perfect, but it's pretty great. 
 
I can't be the only one doing this, can I?
 
Now, i'm not stupid. Every track I adjust anything that needs adjusting. Please don't intentionally misunderstand this.

gswitz
I often record for a whole night where the band just plays through. Songs of different levels and intensities .... I mix the first song ... If I ever see a PC light blink, I reduce inputs or gains. ...I make a copy of the project for the first track to use for the second track... I split the clips in the full project ...
 

I used to work this way with live recordings and extended sessions.  The new Mix Recall feature changes that. I now have the convenience of a scene or 'snapshot' of the levels etc. of a particular song, and then can freely leverage that for the next song(s).  Across an evening's performance of two hours, perhaps 6 unique snapshots (Mix Recall) are embedded in a single project.  I like working this way now.

Bristol_Jonesey

Yes, but you'll also be raising the noise floor.




I thought so, thanks Bristol.

Just wanted to thank everyone involved for their participation and their knowledge. I have learned a thing or two from this post. It was really a matter of interest to see how others are working and get some feedback.
 
Can anyone explain why the gain knob at the top of a channel, when turned all the way down does not turn the audio all the way down. It seems as it just softens the audio?
 

It adds and reduces gain by the amount indicated.
What you describe is the function of the level fader. That cuts all the way down to infinity, i.e. "off".

jpetersen
It adds and reduces gain by the amount indicated.
What you describe is the function of the level fader. That cuts all the way down to infinity, i.e. "off".

So, the gain knob will never turn the volume down 100% even if I turn the knob down? Its purpose it to just raise or lower the volume by the given allowance for the knob?

Yes.. The track gain is set to a useful range rather than complete range.

Clip gain is automateable. Not sure how low that goes. I've never pushed lower bounds on these parameters.

You can mute tracks and clips.