There's nothing technically/sonically wrong with dropping the master fader to avoid clipping. But it is - IMO - an inefficient practice, or rather, a symptom of inefficiency elsewhere.
The only problem you'll run into (when the fader is
really really low) is fader resolution.
This is a GUI thing, nothing to do with sonic resolution. That is, if your fader is down at -60dB and you want to adjust it by a dB it will be hard! Down there the fader has lower/different resolution. You'll make the smallest movement and it will jump 20dB or something
Frankly though if that becomes an issue you have much more important things to worry about, such as why the f#ck your mix needs to be turned down 60dB!
If you want to keep your master fader at 0 then it sounds to me like you should adjust your mixing habits and keep track/channel levels lower from the word go. :thumbup: Stops you moving the goalposts about and will give you better mixes! Why have all the shit on the left turned way up only to take it all away on the right? There's really no need at all to try to have everything up close to 0 in your mix, and especially at your output.
This leads nicely into your other question about resolution;
Modern sequencers work at 32bit floating point. This is essentially (if not strictly) 24-bit resolution with an extra 8 bits to make it a 'sliding scale'. This is why your +60dB mix doesn't sound clipped once you turn the master fader down 60dB (unless plugins are clipping internally - waves for example). So, forgetting the clipping stuff, under 0dB we have 24 bits. The rule of thumb is 6dB of dynamic range per bit, which isn't perfectly exact but very close. So to take some common bit depths;
24 bit x 6 = 144dB of dynamic range before reaching the inherent digital noise floor.
16 bit x 6 = 96dB of dynamic range before hitting the noise floor
12 bit x 6 = 72dB of dynamic range
8 bit x 6 = 48dB of dynamic range
So if we're working in any modern sequencer, to have the equivalent resolution of:
16 bit audio: mix is going to be peaking at
-48dB 
12 bit audio: mix is going to be peaking at
-72dB
8 bit audio: mix is going to be peaking at
-96dB
This is not really perfectly correct but it illustrates the point nicely :smil: A more pertinent point for the real world - even on DSF no one is going to have a mix peaking at -48 FFS
- would be mixes peaking at -3/5/10 or whatever.
I've had messages from people on here concerned cos their mix is peaking at -5dB. They feel inadequate and worry that their girlfriend might leave them (ok, that last bit isn't true). You'd think peaking at -5 was some sort of heresy worthy of burning at the stake, COZ IT DONN MAXX AHT ALL DA BITZ, at least with some people. So people force everything up and walk that tightrope of getting their mix to peak at -0.1 without clipping. -0.2 is wasting resolution, must get that last 0.1dB in there!
But as you can see, even a mix peaking at -5dB uses all 24 bits. Fuck's sake, even a mix peaking at -10dB is using 23 bits ( = 138dB dynamic range)! The examples above hopefully show that your mix has to be
really really really fucking low to worry about audible degradation caused by wasting resolution.
So, in summary,
when mixing in a modern digital environment there is no need at all to try to get everything as close to 0 as possible. Keep levels conservative from the start of the signal path, at the track level, and allow them to add naturally. You won't be worrying about clipping or intersample peaks, everything will sound better, your ME will thank you (it is inexcusable to submit a mix that clips IMO), and you won't be moving your master fader about, everything will be more efficient and allow you more time to worry about making music.
