Compression

K

kieran t

Member
Joined
Sep 22, 2009
Location
MANCHESTER
So I've just started really getting into production PROPERLY this time.
I've never really experimented due to lack of knowledge.
However I've been trying different things out.
I am aware it takes time and I don't expect to be able to suss it out overnight.

I have had a play around with parallel compression. I bounced a copy of my drums before I used any compression at all to compare afterwards.
I've used compression on single tracks and I have set the parallel compression up with all my drum parts kick snare breaks hats. ( is that correct? )
The result wasn't what I expected.
It now seems like it's lost life, the lows ain't as nothing like before, which brings me on to another point
I have mono and stereo tracks in my drums when sending signals to stereo dose that not defeat the object of being mono? Should I not send the mono tracks ?
Why dose it sound like this ? Too much compression ? So many questions .....

I would like advice please from experienced producers. hope I have explained well enough. Thanks a lot.
 
Sounds like your Attack and Release times are wrong to me.

If your source is Mono to the bus, it will be mono out of the bus in general. I won't magically make it stereo unless it has stereo effects on the compressor.
 
Ahh ok, essentially by parallel compression it's making a thicker sound like stereo without panning? Am I correct in saying that? Trying to get my head round it .
I went all out with compression putting it on everything. How do I decide wether or not to apply compression?
 
kieran t said:
Ahh ok, essentially by parallel compression it's making a thicker sound like stereo without panning? Am I correct in saying that? Trying to get my head round it .
I went all out with compression putting it on everything. How do I decide wether or not to apply compression?
Click to expand...

Not quite, but you're getting there.

Parallel compression will add a bit of thickness/punch to whatever is being parallel compressed, but doesn't really do much in terms of enhancing the stereo field.

Essentially, parallel compression involves using ridiculously high settings to get "transient(s) on steriods" and then mixing just a touch of that signal back into the original.

For example, let's say you have all of the components for your drum kit going to a bus. If you have a compressor that has a mix slider, slap it on the drum bus and go mad on the settings–big threshold, gain comp and att/rel to taste. The result right now probably sounds like a squashed, snappy/pumping mess. Now, dial the slider down so that the compressor is not being heard. Now start to increase the slider to slowly mix in the signal until you just barely start to hear the increase in body/volume. You're now running the overly compressed signal "parallel" with the original.

If your compressor does not have a mix slider, simply setup a send (full) to another bus that contains a compressor with extreme settings. Now drop that bus' fader so that none of the signal can be heard and being to slowly raise it, introducing the compressed signal into the original.

Make sense?

Cheers.
 
lostnthesound said:
Make sense?

Cheers.
Click to expand...

Yeah it dose thanks :)

Another thing is I was wondering is how much compression is too much compression? And by that I mean.
Say you have 3 layers in a snare and you compress each hit individually. Then send all to bus and compress as a whole, and then you decide to parallel compress as well.
Wouldnt that result in way too much compression ? Would you leave out compressing the snares as a whole because you are going to parallel compress anyway?
Or Do all 3 but with v. Little compression ( just enough) so it's not Over compressed at the end?
 
get smexoscope (FREE) - throw it on the end of your fx chain .
throw in your compressor and start fiddling with settings...all this time watching the scopes visual. this way you will have compression nailed in no time.

 
Last edited by a moderator:
Compression controls volume over time, ie it levels out the peaks (loud bits) in a signal. It was designed to even out recordings that were too dynamic. In the digital realm where we are effectively sound designing every beat/hit individually in some cases '''''you probably don't need any compression!!!! You are already probably being very precise with the peak volume, the attack and decay of the sample etc. You have made compression redundant in other words.

Lost's pointed out the parallel compression method which is an excellent way of thickening up a signal and still relevant in the DAW environment, and of course if you are dealing with samples of real recorded instruments or vocals compression might be necessary. And of course we can abuse any tool/effect for creative purposes.

Optical in his reddit says some of this, which I read and thought 'cool, maybe I'm not doing it wrong' cos it's easy to get the impression that you need to compress everything (like your snare example)

Anyway, just added this as a thought
 
lostnthesound said:
Not quite, but you're getting there.

Parallel compression will add a bit of thickness/punch to whatever is being parallel compressed, but doesn't really do much in terms of enhancing the stereo field.

Essentially, parallel compression involves using ridiculously high settings to get "transient(s) on steriods" and then mixing just a touch of that signal back into the original.

For example, let's say you have all of the components for your drum kit going to a bus. If you have a compressor that has a mix slider, slap it on the drum bus and go mad on the settings–big threshold, gain comp and att/rel to taste. The result right now probably sounds like a squashed, snappy/pumping mess. Now, dial the slider down so that the compressor is not being heard. Now start to increase the slider to slowly mix in the signal until you just barely start to hear the increase in body/volume. You're now running the overly compressed signal "parallel" with the original.

If your compressor does not have a mix slider, simply setup a send (full) to another bus that contains a compressor with extreme settings. Now drop that bus' fader so that none of the signal can be heard and being to slowly raise it, introducing the compressed signal into the original.

Make sense?

Cheers.
Click to expand...
Hot damn you just helped a ton m8. Thnx!

- - - Updated - - -

subprime said:
n the digital realm where we are effectively sound designing every beat/hit individually in some cases '''''you probably don't need any compression!!!!

-------------------

Optical in his reddit says some of this, which I read and thought 'cool, maybe I'm not doing it wrong' cos it's easy to get the impression that you need to compress everything (like your snare example)

Anyway, just added this as a thought
Click to expand...

I was just about to mention Optical's reddit, too. Indeed he did say something to that effect, rarely compresses unless he absolutely has to. It's made compression a much less intimidating beast for me, but now I'm feeling better so that, if I have to use it, I know how to use it.

As always it's tuning everything to get it sounding like a cohesive whole within a certain space that's the trickiest part, but that comes from knowing what your music is wanting to say.
 
kieran t said:
Yeah it dose thanks :)

Another thing is I was wondering is how much compression is too much compression? And by that I mean.
Say you have 3 layers in a snare and you compress each hit individually. Then send all to bus and compress as a whole, and then you decide to parallel compress as well.
Wouldnt that result in way too much compression ? Would you leave out compressing the snares as a whole because you are going to parallel compress anyway?
Or Do all 3 but with v. Little compression ( just enough) so it's not Over compressed at the end?
Click to expand...

That's the million dollar question. :) how much really is too much?

The previously mentioned use of an oscilloscope is a good way to visualize transients. Personally, I use FabFilter C so I can visualize the before/after within the compressor.

What's really key (IMO) is what Subprime said: in the digital realm sometimes the best compression is no compression. I often find that a bit of subtle saturation can achieve the effect I was trying to achieve with compression. Other times I slap an EQ on it and move on.

Msmith222 and I tend to use compression in very small doses so it's very transparent for the sake of "gluing" the sounds in a mix together. If I want more punch, I turn to either Logic's overdrive plug or my new favorite: Decapitator. But I digress.

Ultimately, I'd say to trust ears. If you find the sound(s) you've applied compression to sound dull or weak when you a/b the compressor active/bypassed, turn it off.
 
kieran t said:
Yeah it dose thanks :)

Another thing is I was wondering is how much compression is too much compression? And by that I mean.
Say you have 3 layers in a snare and you compress each hit individually. Then send all to bus and compress as a whole, and then you decide to parallel compress as well.
Wouldnt that result in way too much compression ? Would you leave out compressing the snares as a whole because you are going to parallel compress anyway?
Or Do all 3 but with v. Little compression ( just enough) so it's not Over compressed at the end?
Click to expand...

Before you start messing around with compression, you need to understand what it actually does.

Basically when you compress something, it will take the loudest part of the signal, and compare it to the quietest part of the same signal, it will then adjust accordingly, bringing the quieter part of the signal upwards, and the louder part downwards to bring them to a more "neutral" level. If you over do it, you will lose a lot of the dynamics of your track. A prime example of this is the Complextro genre (or commercial dubstep) where the loudness war has exceeded all boundaries! If you listen to a lot of this genre you will hear that it sounds very loud, but quite flat, this is due to the producers compressing everything WAY too hard! You need to find a happy medium.
 
Mr Fletch said:
Before you start messing around with compression, you need to understand what it actually does.

Basically when you compress something, it will take the loudest part of the signal, and compare it to the quietest part of the same signal, it will then adjust accordingly, bringing the quieter part of the signal upwards, and the louder part downwards to bring them to a more "neutral" level. If you over do it, you will lose a lot of the dynamics of your track. A prime example of this is the Complextro genre (or commercial dubstep) where the loudness war has exceeded all boundaries! If you listen to a lot of this genre you will hear that it sounds very loud, but quite flat, this is due to the producers compressing everything WAY too hard! You need to find a happy medium.
Click to expand...

Yes I understand that I've never really gone in to it too much because I didn't really understand this part. Now I do I am trying to move on. Thing is what I don't understand fully is when to
Use compression.
I've been using vengeance samples and if I try compression on them it tends to sound better before so I end up clicking the off switch. Is this because they are already processed.... ?
To be honest I've just been slapping compression on all my tracks and seeing what happens .
Trying to gel layers together as well as idividualy and on my breaks and hats.
God knows if I am on the right path but just trying everything really.

By the way I've just been working on my drums only I need to get them phat because the rest of the tune always ends up shit otherwise
 
kieran t said:
Yes I understand that I've never really gone in to it too much because I didn't really understand this part. Now I do I am trying to move on. Thing is what I don't understand fully is when to
Use compression.
I've been using vengeance samples and if I try compression on them it tends to sound better before so I end up clicking the off switch. Is this because they are already processed.... ?
To be honest I've just been slapping compression on all my tracks and seeing what happens .
Trying to gel layers together as well as idividualy and on my breaks and hats.
God knows if I am on the right path but just trying everything really.

By the way I've just been working on my drums only I need to get them phat because the rest of the tune always ends up shit otherwise
Click to expand...

I'm kinda in the same boat as you, but the hard truth is like a lot of things in production there's no end-all, be-all answer regarding when to use it or not. If you have a pre-processed sample that's leaving a lot of dB headroom and you want to make it more prominent in the mix while tuning out some of the dynamic range or w/e, you could do some compressing on that sound.

But if you want it softer/quieter and with the full dynamic range intact, maybe not. Or maybe you want frequencies of a certain type present, so you could EQ, then say that's too quiet and you want it more prominent in your mix, so maybe you could compress after you EQ. These are just extremely rough examples.

You'll only know when to compress once you have a good idea of where the track as a whole is going, and it comes down to what you want out of the song. Hench, why there isn't really an easy answer to that kind of question. Sucks, but that's how it is :S.
 
Mason John said:
You'll only know when to compress once you have a good idea of where the track as a whole is going, and it comes down to what you want out of the song. Hench, why there isn't really an easy answer to that kind of question. Sucks, but that's how it is :S.
Click to expand...
This is incorrect. Compression is used in mixing and depends completely on the sound, not the kind of track it's in. Compression is done with an idea of what you want fron the sound, and the settings you use vary.
 
Instead of smacking compression on everything and anything i have now just been applying it to help me with the sound I want and to reduce peaks that stick out. I really do think this forum is helping me.
I keep trying testing and thinking more about what iam doing and why I am doing it :)
 
If you want to reduce / prevent peaks and peaks only, use a limiter rather than compressor - a L1 or L2 for example. What they do is only limit the peaks in a file, not compress them unless pushed very hard. Compression reduces the overall dynamic range within an audio file, with the settings you have tweaked.

Overall it is quite funny that compression has became the 'the' subject in modern music making and audio processing - why people don't talk about, for example, the fact that most virtual EQs create phase errors when you boost / cut more than 6dB? Of course not every plugin, but very many do
 
Back
Top Bottom