Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[rwj1313]

Hey Guys,

I've seen mention of people using multiple comps on single tracks to achieve various results. Some people have called it parallel compression and others have used serial to describe what they are doing. I'm an electronics guy so I have a very distinct idea of what parallel and serial mean from that stand point and I'm not sure I understand how you would hook up compressors in parallel. That said........I have done searches on these forums and even on Google to try and find a good description on how to use compressors in this manner. My background is in live sound but I am new to the studio thing which is why I'm asking. I am not new to comps and know how to set them and why.  I'm just wondering if you guys could clue me in on why you use multiple comps and how you set them and what it is you are trying to achieve. I have seen this technique mentioned for vocals mostly. So any other uses and why would be a big help.

Thanks,

Rick

[.nathan.kosakowski.]

ok, well... it really isn't all that different from electronics (from the schematic viewpoint).

serial compression would be two or more compressors in sequence, the output of the first leading to the input of the second.  (that can also be modified inserting an eq or something else between the two, ie. COMP->EQ->COMP->where ever...

parallel compression invloves first spliting your source signal into two copies (a Y cable, various digital methods, etc).  for simplicity, the first copy would continue on its way through the mixer (or what have you) while the second goes through a compressor then later recombined with the first signal in the mixer.

problem is watching out for phase.  not so hard when everything is analogue, but with digital you need to be concerned with latency and compensating for it.  current versions of most DAWs automatically do this. also depending on what your are trying to accomplish, the first signal can remain untouched or be processed; if you want to keep ALL transient info or just want to beef up a vocal when regular methods aren't cutting it, resectivly.

you are also free to processes both sides of the parellel tree, just as long as it all remains phase coherent upon summing.

or for you graphic types:

serial:

--COMP--COMP-->

parellel:

   __source signal__
---<                 >-->
   ------COMP-------

works great with weak vocals, drumsubs, even over the whole mix in mastering.  the main point is to beef up and raise low level material without effecting its transients.  its the best of both worlds, transients from the 'clean' side of the tree, and the meat from the comped side.

how this wasn't so verbose as to be counter productive, i like being thorough.

[rwj1313]

Hey Nathan,

I understood the serial thing........ just wasn't sure about parallel but your explanation made sense to me. I think I'm looking for an example of a problem with an example of a "how to" type thing. You said you could fatten up a weak vocal...................how? using multiple comps

Thanks
Rick

[.nathan.kosakowski.]

thusly:

you have a vocal track.  do what you have to do to get it one two separate faders.  in a DAW just duplicate the track, audio and all...  

trackone: your vocal track, eq'd, comped, what have you.  even after all that work you think is could still be a little stronger sounding, so...  

tracktwo: treat this like a separate track, squashing the hell out of it with a pumpy compressor on an insert.

send the outputs of both tracks to the main bus, thus combining everything back together.  set trackone to a decent level, and slowly bring up the fader on track two.  start with tracktwo around 20dB below trackone.  just to see whats going on.

mind you, if both tracks were at unity gain (full volume), it'll probably sound like crap (maybe its a cool effect), but usually the second track is lower in volume as to just bring up the low sounds without adding much transient info.

now, you can modify this a few ways.  first, you can use things like aux sends to patch a copy of the signal to a second track.  use prefader for now if you can for your sends.  you dont need to compress you first track all the time, if it was a delicate precussive sound or something, leaving it alone with all of its little transient peaks intact, the second compressed copy will (hopefully) only bring up the low sounds without changing the attack of the transients much.  this very cool to do to a drumsub.  mix all your drums, and now send them to two stereo subs.  one is your normal one, the other the parallel one.  with the second one mixed in a little it can be a good thing.

but again, phase issues is very important to keep track of.  if the second parellel copy were to arrive a millisecond later (due to a digital processor or something) the whole this will sound like garbage.

how about this, tell me what equipment you are using and i'll give you a better step by step.  private message me if you want.

take care.

[rwj1313]

Damn Nathan,

I'm sorry, I wasn't very clear. When I said "I got the serial thing" I was referring to how you connect them not how or why you would set the comps differently. I got the parallel thing but not the serial. I actually am already doing the parallel thing with the snare drum on some of the tracks for a band I'm mixing right now. I kinda discovered it by looking at a demo file that came with SONAR4. I have the original snr track very lite ly compressed and then a copy that has the shit squashed out of it. I then send those 2 snr tracks to a subgroup to be mixed as one. I just never realised what I was doing was called parallel compression. I'm doing everything ITB so connections are easy.

Thanks
Rick

[drumsound]

When I use serial compressors I'm usually trying to get more GR with a less obvious effect.  Many compressors do change the color much at a few dB of GR, but once they start working hard they can be very noticible (sometimes a good thing!).  If I'm just trying to even out a vocal I'll use  a gentle compressor and a fast one to catch the things the first might miss.

[brandondrury]

"GR" meaning gain reduction.

Brandon

[William Boyle AKA Elfy]

Speaking of this sort of thing does anyone have a link to how Michael Brauer does his multibuss thing.
Cheers

[.nathan.kosakowski.]

ok, sorry for maybe ranting about the wrong thing.  if you are still looking for some more specific examples:

take the vocal track 'drumsound' was talking about.  say you have a singer who sings choruses loud and verses quieter, and you are looking for a way to wrangle it all without riding the faders or input gain.  the first compressor will be a very slow one, acting similar to you just pushing a fader up and down.  the idea is to move things into more consistent volume long term without changing much transient info.  try a 2:1 ratio, 100 ms attack, 1000-1500 ms release. (this is just a starting point for the sake of conversation).  set the threshold so on the loudest part it might peak at a -6dB reduction, averaging 4dB ish during the loud parts. this would kinda be like pulling the fader down a few clicks during the loud parts.

then have a second compressor using a more rock'n'roll setting.  like 4:1, 5ms attack, 150ms release, 3-6dB reduction.  again, this is just a shot ing the dark preset for conversation.

doing the first compressor with a low ratio of 1.2:1 or something and the threshold really down low (-30dB - 40dB ish) can smooth things so.  

this slow comp/fast comp thing works great in mastering as well, to get a overly dymanic song into a more workable position so that you can then compress and eq normally.  also, as 'drumsound' said.  even setting two compressors on the same settings can be very cool.  putting a vocal preset on both and pulling 4-5dB (even more) red. on both can get you huge compression with minimal artifacts and pumping.  some comps are even made to be used this way.  gets your whispers and screamos the same volume without sounding too messed up (ie. a lot of screamo-type vocals).

my favorite is putting two la2a's (or another optical type if you need to) in a row.  it's what is used on kate bush's records for that 'sound' (if you like that kind of hugely dymanic voice being totally regulated and still sounding natural)

one more thing for weak vox.  if you have a compressor with a highpass filter on the detector circuit, use it.  this takes the low end out of the sidechain, causeing the compressor to only be triggered by the mid's and such.  the result is a nice bassey full sound.  the actually audio isn't eq'd.

you can do this with a sidechain input which most plugs have anyway.  you make a copy of your track like with the parallel comping (using an aux is fine if sonar does delay compensation).  take this track and apply a high pass filter at about 150 Hz.  set this tracks output to the sidechain input of the compressor plug on the first track, set the comp to use the side chain, tahdah!

most software uses buses to acomlish this.  setting the hp'd track to a bus, then setting the sidechain input of the comp to listen to that bus.  other software will put an actually item in the output list for the sidechain inputted comp (inputted is a new word).

another pointless novelette by:

[wavdoctor]

drumsound wrote on Fri, 14 October 2005 07:51

When I use serial compressors I'm usually trying to get more GR with a less obvious effect.  Many compressors do change the color much at a few dB of GR, but once they start working hard they can be very noticible (sometimes a good thing!).  If I'm just trying to even out a vocal I'll use  a gentle compressor and a fast one to catch the things the first might miss.

You know I did this for years thinking compressor first and then Limit..But I'm reading a book now from a well know engineer who says NO..limit first Mildly..and then the compressor doesn't have to work as hard and gets a cleaner signal. Just passing this on for discussion?  

Harry

[Fibes]

This always breaks down to what you want to accomplish and for me sometimes it's a matter of the specific units one is using. I love distressors on vocals but sometimes they can be a bit much and get a little too honky. It is my opinion that when it isn't working as hard and the front end isn't being taxed as much the unit gives the grab I'm looking for without all of the grit. How do i normally accomplish this? By throwing a very robust comp in front like the Trakker and all of a sudden the track opens up and slides into the mix nicely.

As stated earlier the trakker is really doing a lot of the work but the Distressor is working in a narrower more easily dialed in range.

[PookyNMR]

When I use serial compression it's usually for 2 reasons.  

First reason is because I need one compressor to really handle the fast peaks and the other to handle more of the body.

The other reason is because I like to have one that detects the threshold  with peak levels and the other with RMS levels.