From Taika-Kim on 11.02.2004, 14:47:
Hmm, I'll give some pretty straightforward not-too-technical guidelines about how I do things usually.
This is meant for people who don't know what a compressor does, and I hope that after reading this, they will
Assuming a dance track:
I'll set the kick to -6db, that means I'll reserve half of the total song dynamics to the kick drum.
That -6db should be the bassy oomph-part of the kick, not the "snap"-tone (if any) of the attack phase. It's also good to have maybe a 10+ms delay before the kick reaches the maximum level. This so that the start of the beat where usually the most instruments are triggered, doesn't take up so much dynamics and the overall level can be louder (we want that in a dance track!)
(2 bad I'm not at home, I could post some pictures to clarify this...)
Then I set bass(es) to maybe about -12db, and again that means the "body" of the sound, I might reserve 3 or even 6 additional db for a sharp decay phase to make the sound more distinctive.
So in practice: you have a rhythmic bassline and it seems to hover around the -12db mark in the wire view but there are some spikes that go up all the way to, say, -4db, you might want to set the compressor on the bass to something like: threshold -6db, rate 1:8, fast attack (>8ms), medium release (<50ms) and then compensate the volume by a few decibels. This way the bass will get more pronounced without using more peak dynamics.
Why worry about the peaks, you might ask? Well, for the simple reason that if you go over 0db in digital domain, you will get clipping!
Of course you can achieve the same effect by (if this is the reason for the spikes) decreasing amp/filter envelope modulation, but often a compressor is a more reliable way since resonance, possible FX etc might (and will!) create volume spikes uncontrollable from the sound generator itself.
If there's a sub bass (goa trance!!), I check out that there is absolutely minimal decay in volume so that the floor-shaking, body-vibrating quality will be to the max.
Usually -12 to -6db is a good relative volume for the sub bass.
Of course the values I present are just guidelines, but I've found out that values around these just sound good.
So we have now a kick at -6db, a rhythmic bass at around -10db (with accents up to -6db) and a sub bass at, say, -8db (this will be really distinct!).
Next we want to add some trippy phaser FX, so we make some stuff and send it to the phaser. We want to maximize the sweep, so we use a big range.
Now what happens is that we have occasional very loud bass coming out of the phaser where it gets to the lower spectrum, and at times we maybe can't hear it properly?
So what we do is put a simple lowpass filter to around 100-200hz to cut out the bass so it doesn't hamper the kick, and what's even more important, the bass.
Next we look at the wire volume bars again and see that at the times where we think the sound is too silent, it's at about -20db. The peaks, are however at, say, -8db so if we just crank the volume up -10db to make the sound louder, we will go 2db over the max and the whole tune will clip (probably earlier, though, because all the other instruments add to the whole!)
Now put a comressor at threshold -20db with a really big rate like 1:16 or something, make a fast attack and slow release (>200ms) and voila! The peaks stay at around -20db. Now adjust the volume to taste. The bass-cut can also be adjusted if the effect seems to be covering the more important bass instruments...
And so on until we have a ready tune.
I use roughly these initial guidelines for almost all my trance tracks:
kick -6db
rhythm bass -12db (+ peaks maybe)
sub bass -12db
percussions -6db
leads -10 to -6db
accompanying melodies & FX -24 to -12db
"key" FX like in fills & such up to -6db
I'll only start finetuning these values when I've íntroduced all the sounds & preferably finished with the general structure of the song.
Then I'll find the loudest part of the song (in electronic music, meaning usually where there's the most instruments playing, in acoustic music this is not necessarily the case...) adjust everything so that there's enough space for anything, and THEN I proceed to finetuning the different parts of the song, hopefully never increasing the volume of anything anymore, but decreasing the other instruments instead...
Remeber decreasing the volume 6 decibels simply means LOWERING THE VOLUME BY HALF! (ok, was that lound enough!?)
That's all there's to it. -6db is half the volume of whatever we are talking about, -12 is 1/4 of the maximum, -24 is 1/8 and so on...
Now that the track is ready, we set master gain to some very low value to avoid ANY clipping, and render a 24bit wav.
Then first normalize the whole thing to 0db to see the overall dynamics.
Then add the final EQ and compressing, leaving from 3 to 6db of "space" for the beats etc. What I'm meaning is, that the overall volume is at -6db but loud sounds like the kick, many percussive sounds etc go all the way to 0db.
Phew. I hope this was of some use to those who are still wondering about the mystical "dynamics" that people keep talking about all the time
It's like reading, first you just don't get it, and afterwards it's so simple that it's hard to explain it to somebody who's still confused about it
And finally, a good homework lesson for everybody to get a hands on feel of the whole thing:
get a breakbeat that isn't compressed (you'll se a lot of sharp spikes in the wave graph if you can't otherwise say), put it through a compressor, and tweak the settings with preferably headphones on, and the wire volume after the compressor on the screen for the whole time. You'll hear what the thing does easily enough!
|