Master bus compression

Throughout the years i have read a truck load of stuff about compression and master bus compression, yet, i only feel like i got a grip on it a few months ago.  Here's why.  For this 5th entry i present you: Master buss compression. The reality ( kind of )

TB Buss Compressor

To put some perspective on what i'm about to write i'd like you to know a few things about me.

1- I'm a sound design/recording techniques graduate from Trebas. ( for some extra credibility ;) )
2- I've been trying to get better at this since 2005 ( graduated almost 10 years ago )
3- I have no mentor and i've been trying to figure this out on my own since i graduated.

I have to credit some sources for this enlightenment. ( No it didn't came to me from a higher realm )

1- Kenny Gioia  mixing tutorials ( http://www.groove3.com/str/mixing-rock.html )
  
2- Stillwell audio Bombardier ( http://www.stillwellaudio.com/plugins/bombardier-buss-compressor/ )
  
3- TB bus compressor ( http://www.toneboosters.com/tb-buscompressor/ )

A few years ago all i saw on the web was people debating bus compression. Is an SSL type compressor better then a Neve type compressor ?  What about brand X and brand Y ?  Yeah but what kind of music are you mixing ?  and blablablabla.  Took me a while to realize  that most people are looking for a magic box rather then actually understanding how to properly compress the master bus.  Although different compressor will yield different results what i would like to leave you with is a concept.

There is two ways to go about compressing the master bus. The first one is the most commonly talked about and it involves using a rather short attack time to catch the peaks ( mostly drums ).  Here's where the SSL vs Neve vs Fairchild debate takes place.  The idea is to retain the biggest peaks while making the elements in the back rise in volume.  It brings the elements closer together and it is widely considered as the "glue" factor.  The setting i often start with for this kind of compression is a 3ms attack and a 100ms release on an SSL type compressor. Sometimes i drop the attack to 1ms, sometimes i put the release on auto. Attack depends a lot on the transients info so since we are mainly compressing the drums at this stage i listen carefully to how the kick and snare are reacting to the attack setting.  For the release i try to make the compressor VU meter fallow the movement ( beat ) of the song.  A longer release, aside from the fact that it can make the compressor breath, will change how the high frequencies behave. ( listen for the hi hat ). A shorter release can make the compressor pump and it will also make the high frequencies stand out a little bit more as it doesn't always gets compressed.

The amount of compression will determine how the peaks will be controlled and how much the lower elements will rise. Listen closely to the volume of the pads vs the volume of the drums. This kind of compression will more then likely make the volume of the bass rise too so take also that balance into the equation as you are dialing the compressor's settings.  Listening in mono can be of some use here as you won't be influenced by the side volume vs the mid volume ( m/s ) but always take the final decision in stereo. Three db is enough to start with. I often dial it back a bit as the mix moves along and i usually end up with only one or two db of compression.

Most people will recommend you to start using this compressor form the get go.. as soon as you start mixing but my experience tells me otherwise.  I usually get a rough balance going before i turn on the master bus compressor.  Once the main levels are set ( kick, snare, bass, lead synth and main pad ) and that i have a bit of a mix going, i will compress the master bus in search of the glue factor.  You have to try both ways a while before finding out which one works the best for you. There is sadly no real short cut here. It has to be a plus for the song though.  If you put something on the master bus and it doesn't make you feel better about the song right away, please, do take it out immediately and either reconsider the idea of bus compression or, more then likely, revisit what you have done previously in the mix up to this point

Stillwell Bombardier

Second way to go about master bus compression is the least talked about.  It also makes the whole debate about brand a bit less relevant because it relies more on a technique then anything else.  I do have to admit that Stillwell Bombardier and TB bus compressor would be great contenders for this application but i  reckon that i can get similar results with Reaper's own ReaComp.  There's two ways you can go about it.  Either using a RMS compressor with a variable RMS knob ( Bombardier ) or by using a compressor that lets you set the attack very long ( TB bus compressor ) Funny enough, ReaComp can do both.  While generally a bit too transparent for me i have to agree that it does a fairly good job as a buss compressor when you are looking for transparency... which i usually am on master bus. When i want some more mojo i will fallow the compressor by TB Reelbus. ( more on that later ) 

So if you are going theRMS way ( and this is according to Stillwell audio Bombardier manual ) Set very short attack and release time. Set in feedback mode if possible. Set rms length above 100ms. Use very low ratio. Preferably under 1.5 and aim for 1-2 dB of GR.

If you are going the peak way ( the usual way, which i saw Kenny Gioia do and which is how the TB bus compressor preset ultra clean compression works ) set the attack and release time slightly above 100ms. Use low ratio of 2:1 and under and again aim for 1-2 db of gain reduction.  You'll probably also want to use a soft knee which makes the ratio follow a slop from 1:1 to it's set value  ( i.e 2:1 ) instead of a hard knee which makes the ratio go from 1:1 to the full on set value above the threshold. In other words, a soft knee makes the compression a bit more gradual and well ... softer.

Now what this type of compression does compared to the previous one is that it lets the peaks trough a whole lot more.  So much in fact that it barely see them. Therefore the compressor acts on what's lower in the mix leaving the transients unaffected.  Imagine if you will the drawing of an audio wave with it's peaks and valleys, well the valleys is what this compression is working on. It is truly a beautiful thing to ear. It makes the "bed" of the song totally gel.  For a long time i was looking for this type of sound in my own mixes and i could not figure out how until i came across those three things in a short period of time. Then it clicked.

TB ReelBus

 The cherry on the sundae is when you use both type of compression one after the other like Kenny does.  First the long attack/rms technique then the short attack technique. One controls the bed, the other one controls the peaks.  I very often replace the second one ( peak compression ) by a tape emu.  If you didn't know,  it was one of the benefits of working with tape, it slightly compressed the peaks if driven a bit and it would also provide some saturation. TB Reelbuss is very good at this, so is Ferox but i reserve Reelbus for my master bus as it is more cpu hungry and because it is more of a proper tape emu. Even though Ferox is also base on a tape emu, it feels more like a saturation tool then Reelbus ( to me anyways ).  I like the mix glue II preset. It rounds up the low frequencies and compresses the sound.  I set the drive knob so that the needle in the VU meter hits the top 0dB mark. It's a safe setting. You can, of course, drive it more but i rather do heavier saturation on group bus and individual tracks rather then on the master bus.


So here it is. Proper master bus compression that hardly no one ever talks about.

Thanks for reading. Now go get your hands dirty and mix something :)