Understanding Dynamic Processors (Compression)

WeissSound

Engineer
I was inspired by the EQ thread to write a similar one on dynamic processing (compression, expansion, gates, limiters).

1. What is dynamic processing?

A dynamic processor is something that outputs a signal, where the level of the outgoing signal is based on the level of the incoming signal. In other words, a loud signal coming in will come out differently than a quiet signal coming in.

2. What are the basic types of Dynamic Processors?

Compressors - the most common - the louder the signal is coming in, the less level it provides going out. In a compressor, a target level is set - called the "threshold" - and any signal coming in that exceeds that level will be reduced. The higher the level is above the threshold, the more reduction is done. More on this later.

Expanders - the quieter the signal is coming in, the less level it provides going out. In other words - it makes quiet signals even quieter. Much like a compressor, the threshold is set at a certain level. Any signal that does NOT exceed that threshold is reduced, and the quieter the signal, the more reduction is done.

Limiters - limiters are like super compressors. The idea is to ensure that the level does not exceed the threshold. Because this amount of compression is extreme, a limiter relies on certain functions and design that regular compressors do not have.

Gates - gates are like super expanders. Anything that does not exceed the threshold is reduced to inaudible. Again, because gates are extreme, they often require a slightly different design than a regular expander.



Now - I'll focus primarily on Compression, because that's going to be the most commonly used dynamic processor.

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

COMPRESSION

Every signal you hear is compressed. ????? Yes, every signal you hear is compressed.

Bare with me. Imagine you have a rapper in front of a microphone. The rapper raps, you record. You play it back. You haven't used any processing - you're just playing back the raw vocals.

You are listening to a signal that has gone through at bare minimum 3 stages of compression - and more likely than not - it's closer to 6.

The microphone capsule gains tension as the rappers voice pushes it - in other words - it pushes back - and the more the rapper's voice pushes in - the harder the capsule diaphragm pushes back. In other words, the louder the signal hitting the capsule, the more reduction the capsule does to the signal. That's compression! (It's mild compression, but it's still compression).

Along the way through the microphone, you may hit a tube. Tubes have a non-linear response to voltage - the response is quite curved, and also changes the frequency balance of the signal. This is called saturation - which will tend to "round out" a signal, by reducing the loudest peaks. Compression! And before leaving the microphone, the signal may hit a transformer as well, which will saturate in a similar way. (more compression).

The preamp is going to have multiple stages of saturation - and often times, the more gain you give something - the deeper that saturation curve goes. In other words, the more you drive the signal at the preamp, the more compression the signal experiences.

Then the sound has to actual come out the speaker cones. Well, those speaker cones are going to build up tension when pushed further. See where this is going? This is called "cone compression".


Ok - so this is a bit of a simplification - but there's a point here. The point is that "compression" is always part of the signal. Some mics have less of it, some have more - same with speakers, tubes, transformers, etc. And they all do it in different ways. With tubes, people will talk about their saturation curves and THD (total harmonic distortion - or frequency alterations). With mics people will refer to how it "grabs" a sound - or more specifically - the sound's shape. These all add up to really the most important ideas:

Frequency and Shape!

Instead of thinking of a compressor that compresses - think of it as something that changes the shape of a sound. If you start listening for "shape" the mysteries of compression will reveal themselves to you, and fairly quickly.

Setting a compressor is like setting a mold for the signal to fit into. The threshold determines where the compressor starts working, the ratio is how hard it's going to work, the attack is another way of saying how sharp will the transient sound be, and the release is how much tail or sustain do you want to emphasize.

Yikes! Time out!

What the hell is a transient? A transient is a very fast signal - in other words the "attack" of the signal. Drums have transient attacks. Strings have gradually risinng attacks. So the attack control on the compressor is really like saying - how much emphasis on the attack of the signal do you want? Do you want the attack to be really rounded out and diminished? Set the attack low. Do you want the attack to be prominent and stick out? Set the attack high. Of course, this works directly in conjunction with the threshold. Try it yourself, set the threshold low, and the attack short. Suddenly, the attack sound of your snare is gone. Set the threshold low and the attack long. Suddenly the punch of your kick is very round and bouncy. Set the threshold high and the attack short. Now the snare is a little fatter and rounder, and not quite as spikey (but possibly a little duller). Set the threshold high and the attack long - the change is hardly noticeable, the attack is just a little bit "rounder."

How to achieve maximum punch?

There is a thin line between a transient sound, and a sustained sound. A sound that holds for any noticeable amount of time is sustaining. A sound that moves by too quickly to register as it's own moment is transient. But transients can vary in length. A transient can be half a millisecond. It could also be ten milliseconds. And those won't sound the same. A big factor in punch is how long that transient exists. A quick transient sounds "spikey" - but a long transient sounds "punchy." You want to find the point that makes the transeint exist as long as possible before "flattening out" or becoming a sustained sound. Only your ear can tell you where that point is.

Good samples are already shaped to have that kind of impact - and any additional compression may soften that. Of course, punch has a lot to do with frequency as well - but that's for another thread.

Now what about the release? The release is super elusive. It deermines how long it takes for the compressor to let go. If the release is too short for the signal you are going to get a disjointed sound shape which usually results in distortion. If it's too long, your signal never really returns to its natural shape, and you generally lose tone. So the idea is to find a point that emphasizes the sustain (which is where most of the signals tone lives) properly.

Lastly, when the attack and release are set in a way that seem to argue - the compression can become very audible. You either hear the decent or the acent of the signal level. This is called pumping. It's generally annoying, but can sometimes be used an effect.

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

So, rather than think of a compressor as something that effects the "level" of a signal. Think of it as something that effects shape. Why? Because level can be controlled with the volume fader more accurately and transparently. A fader doesn't really control shape, unless you are being extremely meticulous. Conversely, compression will always effect the shape of the sound it is working on.

Once you start hearing shape, you will understand compression.


-Matt.
 

NodtomyBEats

New member
idk why no body responded to this.
very insightful i learned quite a bit from this post. if you actually wrote all this from your own knowledge your a g hah.
 

WeissSound

Engineer
idk why no body responded to this.
very insightful i learned quite a bit from this post. if you actually wrote all this from your own knowledge your a g hah.

Guess that makes me a G then. Not sure why the post didn't garner much attention, but I'd wager it's because a) it's a long ass post, and (b) compression can get a little tricky.

But I'm glad that you got something out of it!
 

Mr. Ak47

New member
reminds me of some shit my teacher was saying during our last webinar....good post homie, glad to see this kind of info being passed around the site like this...
 

nrjetik

Three Harmonies
yeah it's a pretty good post. very knowledgable but the part about the mic going through 3-6 compressions was a bit unnecessary regarding what a producer needs to know for production but its not bad to know the background behind it i guess
 

WeissSound

Engineer
yeah it's a pretty good post. very knowledgable but the part about the mic going through 3-6 compressions was a bit unnecessary regarding what a producer needs to know for production but its not bad to know the background behind it i guess

I strongly disagree here. I think that's extremely important. The very important theme of my post is that "shape" and "compression" are inseparable. You can in fact accomplish most of what a compressor does simply by technique, mic, mic positioning, and preamp choice.

The tendency is for people to throw compressors onto things like vocals without realizing that compression is already in essence happening. The reason a lot of the big mix guys are using so much compression is because they have extremely dense mixes going on - and so compression is needed outside of normal measures.

But if your production isn't supremely dense, compression may not be necessary at all - because it's all happened (or can happen) already on the way in.


Also, I have a similar post up on EQ here: http://theproaudiofiles.com/equalization-hear-me-out/
 
Last edited:

Mr. Ak47

New member
"Matthew Weiss is the head engineer for Studio E, located in Philadelphia, PA. Recent credits include Ronnie Spector, Uri Caine, Royce Da 5'9", and Philadelphia Slick."

Love the bit about Royce, arguably one of my favorite rappers, LOVE the Slaughterhouse material....also, good post on EQ too...i crave knowledge, and you sir, have some of that good-good.....
 

WeissSound

Engineer
thanks man :)

I've been teaching a lot these days - and I started doing mixes where I detail everything I'm hearing and doing in thorough notes - basically for the cats who don't really have the bread to have a whole album mixed, they can take my notes on the one song and apply the same ideas to the other songs. So far I'm getting a positive response.

So, I'll keep doing my best to spread the knowledge around!
 

Mr. Ak47

New member
hell yeah boss, that's what it is....so people send you their raw tracks, and you mix it and write up a word doc or something explaining why you are doing it? that is slick fam, i may have to get in on that, i crave that kind of info...i guess i just like peering into the thoughts of others, good idea tho boss...


thanks man :)

I've been teaching a lot these days - and I started doing mixes where I detail everything I'm hearing and doing in thorough notes - basically for the cats who don't really have the bread to have a whole album mixed, they can take my notes on the one song and apply the same ideas to the other songs. So far I'm getting a positive response.

So, I'll keep doing my best to spread the knowledge around!
 

WeissSound

Engineer
I've learned a lot from this, thank you.

Do you have any thoughts on Buss Compression?

Ha - I have lots of thoughts on Buss Compression! That's a pretty general question though.

Basically buss compression is going to do two major things - it's going to even out the dynamics of everything feeding it, which in turn alters the independent dynamics of each element coming together in a way that regular compression couldn't. The second thing is that it will apply a stage of effect on the entire source, which in the emulation and analog world will alter each individual element's tone in a subtle and similar way. In other words: Glue.
 

noblewordz

sexing your mother
Yeah my bad should have been clearer.

I was leaning towards gluing, I've tried it bad have had minumal success with it. I think I need to experiment more.
 

Bryst

New member
Wow fantastic thread. TAHNK YOU! I might have to print this out and keep it on my desk.

I like the part about compression happening through microphones and preamps and stuff. I've definitely seen the "shape" (as you say) of my waveform change from messing with my preamp settings.

Once I accidentally had the preamp on my interface set REALLY low and I had my designated Mic preamp up pretty high. When the vocals came into my DAW they were a brick wall LOL. Totally distorted. Finally realized what I had done wrong and re-recorded.

But yeah that's kinda an example of what you're talkin about I think.

Thanks again for all the knowledge. Guys like you are why I love FP.
 

WeissSound

Engineer
Thanks, I'm flattered!

Does your interface have a "line-in"? If you are using an external mic preamp, you probably want to go into the line level input of your interface. You might need an XLR to TRS cable or adapter to do this though.
 

Bryst

New member
Yeah, I've been told I should do that... My interface is kinda crappy and not all the inputs are working properly lol. Once I get a new one that's how I'll do it though.
 

WeissSound

Engineer
Yeah my bad should have been clearer.

I was leaning towards gluing, I've tried it bad have had minumal success with it. I think I need to experiment more.

You don't actually NEED buss compression for Glue. Almost any universal processing will give you a little glue, and truthfully mixing the elements to that purpose before group processing will do it to.

Oh yeah, and having a unified sense of reverb. I have an article all about that here:

http://theproaudiofiles.com/the-importance-of-space-in-a-mix/
 

moses

hardliner
Yeah my bad should have been clearer.

I was leaning towards gluing, I've tried it bad have had minumal success with it. I think I need to experiment more.

yeah, it's not that easy. The main reason is probably the point that "bus compression" and especially compression on the "2-Bus" (a strange term for "Master channel") is not required in most cases. Most sample and synth based production already have a huge density. Only "classic" recording and production techniques really ask for compression - sometimes even a lot.

The second reason is that many people simply use the wrong tools for the master channel "glueing" you're mentioning. Most compressors on the market - especially plugins - a waaaaaaaaaaay to "fast" for this task. Most of them are built for tracking purposes and totally fail to process stereo signals.

Typical mastering compressors are based on very smooth (i.e. slow) RMS averaging AC/DC converters, while most tracking compressors follow simple peak rectifiers.

Also, do not forget saturation (tube/tape). "Glues" much better than any comp around.
 
Last edited:

WeissSound

Engineer
Yeah my bad should have been clearer.

I was leaning towards gluing, I've tried it bad have had minumal success with it. I think I need to experiment more.

Send me a link to something you're working on. Maybe there is are some key things that just need to be happening in your mix?
 

noblewordz

sexing your mother
yeah, it's not that easy. The main reason is probably the point that "bus compression" and especially compression on the "2-Bus" (a strange term for "Master channel") is not required in most cases. Most sample and synth based production already have a huge density. Only "classic" recording and production techniques really ask for compression - sometimes even a lot.

The second reason is that many people simply use the wrong tools for the master channel "glueing" you're mentioning. Most compressors on the market - especially plugins - a waaaaaaaaaaay to "fast" for this task. Most of them are built for tracking purposes and totally fail to process stereo signals.

Typical mastering compressors are based on very smooth (i.e. slow) RMS averaging AC/DC converters, while most tracking compressors follow simple peak rectifiers.

Also, do not forget saturation (tube/tape). "Glues" much better than any comp around.

Thanks for the advice, I have had a lot more success with saturation than compression, I get the feeling I don't necessarily need compression, most of the stuff I do is either sample or synth based.


Send me a link to something you're working on. Maybe there is are some key things that just need to be happening in your mix?


No problem this is what I'm working on at the moment http://noble.bandcamp.com/track/lonely its the drums, they don't really sit right with me.
 
Top