Help me understand the proper way to gainstage?

[Prototype]

PLEASE hear me out on this, i may be overthinking this but here goes. From what i understand the goal of gainstaging is to get the arrangements volumes to sit well together, and also maximize headroom on the master fader. The way i see it, you can drop your master volume to the level that you want it to peak, and then just adjust your track volumes around that, so that they maintain their relationship without exceeding that point. Or you can get each tracks volume to sit right together at the very start, then just bring the master fader down to the desired level afterwards. This is hard for me to describe because i dont know much about metering but let me put it like this:
you have a mix, you decide you dont want your master volume to hit past -10db so you listen to the arrangement, bring the volume down to peaking at -10, and then set the volume for each track so that the arrangement sounds good and it doesnt peak. In the other method, you first get the arrangement to sound good relatively at your DAWs default master volume, THEN you just drag the master fader down until you reach your desired master volume. Do these two methods render different results? even if you did things to where both methods put you at the same peaking volume, you'd have a lot more room to lift your master volume using the second method. But headroom is about the difference from the peaking volume to digital zero right, not how much further you can boost on your fader? see what i mean by over thinking this ha. Long story short school me to gainstaging!

[BertKoor]

Well, firstly this all has nothing to do (yet) with mastering. That's a separate process which only starts after you have finished the mixing.

During the mixing your main responsibility is to get it sounding as good as possible. Ofcourse you don't want any peaks over 0dB to reach the soundcard / audio interface. To reach that goal the RMS level on the master bus needs to be inbetween -20 to -10 dB. That level is dependant on how much compression / limiting you have applied on the individual tracks.

Now your main question as I understand is: does it matter weather I bring down the master bus by 10dB and leave the individual tracks at peace, or should I reduce levels on the individual tracks and leave the master fader at 0dB?

The answer is that in sound quality this does not matter at all. Every host can cope with a master bus that goes well over 0dB before its fader reduces its level to an appropriate level. Because it's just numbers being calculated. However it's good practice to keep the RMS levels at all points in the audio chain at around -20dB.

There is only one exception, and that is if you have effecs on a bus or whatever that react differently on the input level. Examples are compressors / limiters which have a threshold, saturators, and some EQs that mimic behaviour of analog gear.

[Prototype]

Alright thanks for the response! so help me to differentiate between the responsibilities of the mixer and the masterer as far as gainstaging goes?

[Burillo]

from what i've understood, mastering is a highly technical process of evening out track volumes and frequency responses across all tracks on the album/single/whatever (if you have several tracks, that is), and also a process of getting your track ready for whatever medium it is distributed on (i.e. if you distribute on vinyl, make sure you have mono bass frequencies, make sure you don't exceed certain dynamics, make sure you do this and that) and to make it sound as good as it can on all kinds of speakers (from high end speakers to cheap radios). it has nothing to do with gain staging or everything else music-related.

[Gamma-UT]

The mastering engineer doesn't care about gainstaging: only that the stereo file you supply does not clip. They'll often say they want a file that doesn't peak above something like -3dB or -5dB. It doesn't actually mean they can't work with something that peaks at 0dB. It's code for "some dumbasses send us stuff that they don't know is clipping because they aren't getting clip warnings from their DAW. Peaking at -3dB provides some headroom to reduce the likelihood of something that's clipped when it's written out to a 24bit file".

[Prototype]

LOL alright cool. So as long as the mix gets to the mastering engineer with plenty of headroom to spare then everythings cool? i was told to try and keep the whole mixing peaking between -6 and -10. thats not too low is it?

[BertKoor]

Peaks at -10dB is at the very safe side. As long as you don't render at 16bits for the mastering no harm can be done.