There's many videos about gain-staging, and different ways how to.

Question is , how do you set the proper gain in the DAW, before mixing. ?

EDIT: Video deleted

Generally:
 
Record your tracks @ -12db or lower
Be sure a bus doesn't go over -6db
Be sure master bus doesn't go over -3db
 
Depending on how many tracks go into a bus you may want drop gain or output volume to achieve -6db
Depending on how many buses go into the master bus  you may want drop gain or output volume on buses to achieve -3db at the Master
 
Then you're ready for some non-pro mastering. Light limiting, eq and compression.
 
If you're going to go for a real mastering engineer, then adjust everything down by -3db.
 
Again, this is a general approach.
 
Others may disagree.
 

To directly answer the question in your video:
 
Think about the whole path of the audio. Your audio comes in, and the gain knob is the first thing it hits.
 
So your gain level defines how hard it will hit eveything else.
 
The fader is the last thing it hits. So it's just a volume control at the end of each channel.

It's sensible to adjust your gain knobs so that nothing is coming in much above 10dB on average. That'll mean you're not overloading anything they hit after coming in.
 
So yeah, short version is, input gain is the place to avoid INPUT clipping. What happens between there and the fader is a complicated business.
 
But the ONLY thing faders can prevent is clipping into the master bus. If you're worried about gain staging, you need to control levels WAY before the fader.

use the 50 watt Marshall instead of the 100 watt?
Seriously, John T has your answer in my opinion.

I'll sit down and write out a complete description, but I want to point out that while the rule of thumb described above will probably keep you out of trouble in terms of clipping, it won't provide the best possible result.
 
The trick is to optimize the gain structure of each segment of the audio chain. So tracks and busses are not treated differently, and the two-mix is no different than any sub-mix.
 
With a digital audio system we need to pay more attention to clipping than noise, with an analog system the opposite was true. That seems to remain a stumbling block.
 
The first step is to determine how much headroom you want - that is, how louder than "average" do you want the peaks to be? Now if you want to get really fancy you can factor in "Crest Factor", but we will skip that for now.
 
So we know that live music has a peak to average ratio of between 12 and 18 dB, not all that big when you think about it, but there you go. So if you want to do no compression or limiting you need to record so that your average level is around -12 dBFS to -18 dBFS, so that your peaks will never exceed 0 dBFS. (we are also going to skip over the part where it might be ok to exceed 0 DBFS some times.)
 
But we want to apply this rule to EVERY stage. Individual tracks ought to run at the same levels as sub-mixes or effects busses. The trick is to attenuate the output of the tracks on their way to the busses so that the levels remain consistent.

Sadly - maybe - it is not simple math. You don't get to just drop each of two tracks 3 dB to keep them at a reasonable operating level when they are summed... you can usually get away with a LOT less attenuation in fact.

So instead of a rule of thumb that treats tracks differently than busses think in terms of maintaining your reasonable level everywhere.
 
Does that help?

In your video I think you were asking if we use the gain control or track volume fader for gain staging to the busses. If you look at the signal flow chart for Sonar you will see that the gain control is the first in line as someone else mentioned and it then goes through pro channel and effects bin and on to volume fader. So let's say you have a track recorded at -3 db and you want to bring it down to the -12 range, if you reduce the volume fader it lowers it to the bus but it is still going into PC and effects bin at -3 db. This leaves no headroom in effects and PC unless you reduce the gain control. So I would say you would need to use both to maintain your reasonable levels throughout the signal path as stated above.

Good post from Bill there. And he's right to say that what I've said is no more than a rule of thumb. This is a fairly nuanced business, and gain staging is something to consider every time one module (whether hardware or software) passes a signal to another module. And also something to consider in aggregation.

And what is the best part of the signal to measure and display.  The rms part and what better thing to do it with, a VU meter.  Choose a ref so -20 dB equals 0 dB VU.  Real VU's are nice if you can get your hands on some.  Very easy to switch to any part of your signal chain.
 
Keep tracks and busses and final mix all sitting at desired ref levels.  Tracking has never been easier. Use a VU meter to set your level.  Unheard of.  Never have to worry about transient peaks again.  You will never clip anywhere in a mix and have tons of headrom, it sounds nice especially up loud.
 
Have three ref levels to chosse from even better. -20, -14 and -12.  Very thougtful.  Calibratre ref level in your DAW to SPL level in your room.  Very handy.
 
Last but not least learn to read how the VU meter moves to the music and let it give you even more information about that part of your signal flow you are monitoring.

teego
In your video I think you were asking if we use the gain control or track volume fader for gain staging to the busses. If you look at the signal flow chart for Sonar you will see that the gain control is the first in line as someone else mentioned and it then goes through pro channel and effects bin and on to volume fader. So let's say you have a track recorded at -3 db and you want to bring it down to the -12 range, if you reduce the volume fader it lowers it to the bus but it is still going into PC and effects bin at -3 db. This leaves no headroom in effects and PC unless you reduce the gain control. So I would say you would need to use both to maintain your reasonable levels throughout the signal path as stated above.

That is correct, and i think i will agree to use Gain first then the Fader's after.
I have never used the Gain knobs before, and start to see the importance
due to the FX's input's etc, so they not running to HOT.
-----------------------------------------------------------------------------------------------------------------------------------------

So as i understand it, i should use the GAIN knobs to decrease volume down to
the -12db ( peak ) on tracks. ( some say -10db ) to have proper gain into the EQ's / FX's and so forth.

And i understand the more track's i have, the higher the summed audio is
to the "Sub BUSS's" and finally the Master, ( then less volume is needed on ea track.)

bapu
Generally:
 
Record your tracks @ -12db or lower
Be sure a bus doesn't go over -6db
Be sure master bus doesn't go over -3db
 

This seems quite simple to follow   ( i shortened it down a bit. )
--------------------------------------------------------------------------------------------------------------------------------

The point is not who has the formula, cause there's not a " correct " one
rather than a " coarse " guidance how to...

So then :

* Record at low level -10db / -12db peak for audio track(s), and if using a soft synth decrease level with Gain
* Use Gain knob to set level.
* Route ea track to it's Sub Bus
* Adjust faders, ( mix ) all instruments to a proper level according to each other
* pan tracks
* Insert EQ / compressors and so forth... ( adjust and maintain same level ) with FX's engaged.
* Adjust Bus's volume faders
* Insert EQ / Compressor and limiter on the Master ( adjust to desired Effect )
* Mixdown

Soundblend
The point is not who has the formula, cause there's not a " correct " one
rather than a " coarse " guidance how to...

Could not agree more!!!
 
But of course there are nits to be picked, and the devil is in the details, etc...

Soundblend* Record at low level -10db / -12db peak for audio track(s), and if using a soft synth decrease level with Gain

 
This is where I think folks start to get into trouble. And most folks who start to get into trouble have never recorded to tape, so I'm not really sure why this is so confusing.
 
The short version is this - forget analog, most folks have at most one analog gain stage in front of their computer, and I've worked on projects where no analog gear was harmed... I mean used.
 
In the digital domain the concern is "overs" - that is, samples for which we can not accurately represent their amplitude. 0 dBFS is the maximum value that can be represented.
 
But just as we can hear, and distinguish signals in the noise floor (that's how dithering works), we can also survive very short periods of overs.
 
So, the concern is not about the average level (that's so analog<G>), but rather that we minimize digital clipping. To do that we need to think about the difference between the average and peak values that we are recording, or that we want to represent.
 
-12 dBFS or -18 dBFS should be used in an analagous manner to good old 0 VU. The thing to keep in mind is that  dBFS is ALWAYS a peak measurement (that's just how digital works), and 0VU is always an average response.
 
Hope this starts to part the clouds...