Just to clarify the terms here, "gain staging" means to set each gain to the nominal level for that stage.
The goal with gain staging is to prevent distortion or adding noise, which while an important concept in analog systems is all but nonexistent in digital systems.
Analog Dynamic Range
In the analog domain, you have limited dynamic range and can therefore clip each gain stage in a console or recording chain if signals get too hot, or you can add noise if signals are too low and need to be amplified later in the chain. So you "gain stage" to prevent this, meaning, you set levels at every gain stage to their nominal level, which typically represents a fairly specific level.
Analog systems restrict your available dynamic range to the point where levels really do matter a lot, for the above reasons.
Digital Dynamic Range
In the digital domain, unlike with analog gear, you have a theoretical 1500 dB dynamic range (assuming 32 bit floating point audio, as in Reason). This basically means the nominal level covers such a wide range as to be virtually non-existant or irrelevant. It also means you can theoretically choose any nominal level you want - but there are some restrictions…
There are still a few places in a digital audio system where levels DO matter. One is the main output, where digital signals are converted to analog, and where you CAN clip. Luckily, to "gain stage" the main output, all you need to do is keep levels below clipping, leaving some amount of headroom depending on what you're doing. If you're recording raw audio, you keep levels around 10-12 dB below clipping (which leaves room for unexpected bursts of energy from performers, and leaves headroom in your mix for summing multiple audio channels).
If you are mixing you can hit anywhere from 1-6 dB headroom (depending on who you ask!). When mastering, folks still leave anywhere from 0.1dB to 1dB (or more) headroom, again depending on who you ask and depending on the medium you are mastering for.
Nonlinear Processing in Digital Audio
Another place where levels matter is any nonlinear processor, but even then you have a WIDE range of acceptable levels. For example, on the Master Compressor if you had an extremely
low signal level that never peaked above -30dBFS, you would not be able to get ANY compression (gain reduction) at any setting. Or, in the other extreme if you had a level that stayed well above 0dBFS you couldn't avoid having excessive compression at any setting. In both cases, the levels themselves are perfectly "legal" for the floating point digital audio system, but they would fall outside of the nominal level of that compressor, and you would find it difficult if not impossible to set the compressor to give a useful result.
Similarly, with the MClass compressor, between the lower threshold of -36dB and the input gain of 12dB, you would have to have a signal never exceed -48dBFS before you would be unable to get any compression.
In my experience, it's rare to have levels so low or so high as to exceed the nominal level range of any digital device. And if you are practicing utilizing a consistent peak reference level for all signals, you'll never even have to think about the nominal level of any device or gain stage.
And the same can be said about the concept of gain staging, since every gain stage in Reason has 1500 dB available dynamic range!
This is why I don't use the old analog term "gain staging", and instead speak about adopting a consistent peak reference level for all signals. Folks are probably tired of hearing me blab on about this subject, but it keeps coming up again and again and IMO is worth repeating for those who have not had a chance to hear it!
OK, I'll shut up now, just needed to get a little tech talk out of my system I guess…