One question I have in mind is about impact/punch/transient/fullness. You wrote:
“Impact is a tricky thing to define. Think of drums. Skinny little drums have punch or snap (make you blink), but no BEEF. Having everything mashed into a limiter has beef but no impact. The optimum balance of punch/transients and RMS/fullness is where impact, or power comes from. Space and clarity also play into that.”
That sound very logical but I could not do that in practical way. I tested one way you described to do this with: compression on drums like hell to cut the tail and emphasize transients and than put back level/RMS with a limiter and create some faux punch with short terme distortion. Instead of a compressor I took alloy2 with it multiband tranisent shaper to make it ridiculous snappy/punchy without tail. The result was interesting! The drums pop out a little bit more in the mix but I still did not get more headroom. Can you explain me this way of processing drums a bit more detailed please? Maybe the interaction between transients and RMS in a mix and how to process it is something for your blog?
So, after some thought it kind of boils down to two things;
– It will always be a question of experience with that approach, combined with the specific circumstances (ie sounds) you are dealing with. Say, on one drum track you might need a bit more spike to get the best-sounding result, so you can set the balance more in favour of transient over limiting. In others, maybe you can push the limiting that bit harder for the short-term distortion but still have it sound great.
[This kind of leads into a whole conversation about how dubstep types like Skrillex etc mix; they use deeper fatter 200Hz snare drums that tend to benefit from this sort of thing, that little fake snap on top of a 200Hz fundamental. If you flatten a pingy, snappy 2kHz-centric snare drum it just sounds flat and rubbish. So technical concerns therefore drive musical choices, ie sound selection]
– Having said all that, and this is really important, even if you do have less headroom (ie more spike) then as long as it sounds at its best then it will be in the best and most appropriate condition when it comes to mastering. That is, the more ‘right’ it sounds, the better chance it has down the line. It’s down to the mastering engineer to choose the best method to keep that sound quality.
To take an example, maybe the 200Hz snare will be flat and fat before mastering, and will respond well to longer-attack compression, to allow some punch. That will help it survive limiting, but because it sounded flat and fat before, you end up with essentially the same sound. With a pingy 2k-ish metallic snare, you might need to rely more on clipping, or something acting very fast (my mind thinks of this beauty) to remove just the most minimal necessary amount with no traces, and help it survive without sounding flat. Or the opposite – loads of extra transient crap into the limiter.
So, in short – you won’t always end up with a drum track peaking at the exact same level every time. Unless you’re a mad fool or a relative n00b, expect to have the exact same peak level and exact same loudness and exact same tone and exact same sound quality unless you’re using the exact same drum samples. But then your music is probably pretty boring. Every situation, every set of circumstances is different. Forcing a square peg into a round hole is not a recipe for great sound quality. So, adapt. Be flexible. It depends.
All this is totally theoretical/hypothetical, by the way. But it’s interesting stuff to think about, and important to consider.