Broadly speaking, in an acoustic performance the more instruments you have playing, and the more aggressively each one is played, the louder the perceived result — let's call this Rule One. However, when you're producing music to fit within a fixed digital level ceiling, this intuitive principle has to be balanced with the contradictory Rule Two: the fewer instruments you have playing, and the more delicately they play, the higher you can fade each one up within the available headroom, and the louder/closer each one will therefore appear when the master is played alongside competing commercial releases at any static playback volume. One of the black arts of record production lies in managing your arrangement to reconcile these two conflicting impulses. While this skill is mostly built through practical experience, here are a few suggestions that may be helpful.
The first ramification of Rule Two is that small is beautiful. If you can reduce the number of instruments in your arrangement without losing important musical material, then you'll tend to achieve louder results at the mastering stage. One way to achieve this is to create interplay between different rhythm instruments, such that different parts play in different rhythmic pockets. Possibly the most straightforward well-known application of this idea is the off-beat bass pulse that interlocks with a four-to-the-floor kick pattern in so many dance tracks (just think of Kylie Monigue's 'Can't Get You Out Of My Head'), but it doesn't have to be as simple as that. For example, you could ditch a piano part's uninspired rhythm vamping to clear way for a groovier rhythm guitar, but then briefly mute that guitar to allow the occasional piano fill to pop through from time to time. However, bear in mind that thinning out your mix texture in this way inevitable shines a spotlight on the quality of the performances and engineering/programming, so you can't afford to skimp on crucial groundwork like rehearsal, mic selection/setup, comping, corrective editing, and sound-design.
Another thing to bear in mind in the light of Rule Two is that encouraging live performers to play quieter can often help them to come across louder on the end product. In other words, a mumble will make a vocalist sound much bigger and closer than a shout when both have to fit in the same maximum signal level. One of the privileges of working in the studio is that your mix balance doesn't necessarily depend on how forcefully the instrument/voice was projecting during the recording process. If your musicians are used to fighting for their place in the balance during stage shows, then it can really pay off to talk this issue over with them. (Indeed, John Hudson recalled in SOS May 2004 how producer Terry Britten went to great lengths to get this very point across to Tina Turner during sessions for 'What's Love Got To Do With It'.) This doesn't just go for acoustic sound sources, either. It's very common for inexperienced electric guitarists to drive their amps louder than is suitable for recording purposes, resulting in an unattractively thin and edgy sound in the mix.
When it comes to making the best of Rule One, I personally find that the strongest illusion of subjective power tends to be created by means of progressive build-up. Although gradually increasing the performance intensity of an instrument or adding new instruments one at a time will inevitably reduce the true signal power of each instrument in the mix within a fixed mastering headroom, in practice those two factors don't quite seem to cancel out subjectively speaking. In other words, adding in a new heavy guitar part for your song's choruses, for example, can often fool the ear into perceiving an increase in subjective volume, despite the inevitable decrease in signal power per instrument. The success of this sleight of hand is usually reliant on some careful balancing at mixdown, but one thing that can help where you're adding a new instrument to an already well-populated mix is to mute some part that was added a while earlier and has since been mostly forgotten about — the distraction of the new part will disguise the removal of the older one, which allows you to surreptitiously claw back some precious headroom.
It's rare that you can keep on building up your arrangement for ever, so at some point you have to break things down again to allow any new build-up to occur. Doing this quite suddenly and dramatically is very common in pop production particularly, because whatever skeleton crew you drop back to will immediately demand greater attention on account of Rule Two reasserting itself, maintaining the listener's interest despite the evaporation of the foregoing epic ensemble. And as the memory of the previous climax fades, you can start building to the next one... Mike Senior