Need a quick rundown of uses of compression

And to continue that last point, think of 0db as being the absolute maximum level that a digital system can handle. Anything that peaks over 0db will be effectively stone-wall limited or 'clipped' (it will just disappear into a black hole). In effect this will introduce distortion to the signal, and not the nice kind.
So that's why you hear people talking about leaving their tracks at -3 or -6db or whatever before getting them mastered. This leaves a space for the signal to be boosted (during mastering) up to the maximum 0db if desired.
This space is called headroom.
It's handy to always leave a bit I reckon, even on individual tracks.
 
Last edited:
subprime said:
And to continue that last point, think of 0db as being the absolute maximum level that a digital system can handle. Anything that peaks over 0db will be effectively stone-wall limited or 'clipped' (it will just disappear into a black hole). In effect this will introduce distortion to the signal, and not the nice kind.
So that's why you hear people talking about leaving their tracks at -3 or -6db or whatever before getting them mastered. This leaves a space for the signal to be boosted (during mastering) up to the maximum 0db if desired.
This space is called headroom.
It's handy to always leave a bit I reckon, even on individual tracks.
Click to expand...


Ahhhhh.. that's what i was wondering...

Thanks dudes!
 
0db is what they call "unity" (often notated by the infinity symbol). Unity signifies a signal with the maximum amount of allowed amplitude with zero loss. I'm not positive why they've termed it "unity", but I've always pictured it as being in balance between:

a. A weakened (or weakening) signal and...
b. A distorting signal.

The significance of a signal in unity is that is allows for the maximum amount of signal with a minimum amount of artifacts (which might be background noise or distortion...both of which lie on the - and + sides of unity).

Cool, eh?

EZ
 
DanDnB said:
What I am not getting is why is the threshold knob on the compressor showing a range from -32 to 0 instead of zero to 32, where 32 or some other 'very loud' arbitrary number as being the threshold for compression.

Sound that is negative decibels as THRESHOLD makes no sense to me. The sound we hear is always a positive value, physically as a propogating wave.
Click to expand...

Thats where your getting confused mate. Your right, sound we hear has a postive value. BUT, once its been recorded (and therefore changed into an electric signal), its not longer a sound.

So by the time your signal has got to any form of processing like a compressor, your dealing with an electic signal, which is why the markings on a compressors threshold setting dont relate to the measurement of physical sound.

The best way to look at it, is that once a sound has been recorded, be it by yourself, or a pre-recorded sound, then the word "Decible" will be a measurement based upon the difference between the voltage of your signal compared to a set voltage , its because of the value of this set voltage, that lower signals are recognised as negative numbers, meaning they are however much less than that set voltage.

(I cant remember the value of the "set voltage" off hand which doesnt help!!)
 
Keep in mind compression kills dynamics, I'm sure everyone has heard commercial tracks that are really loud but have very little dynamics left in them. They're just squeezed to fuckery and look awful in a sound editor.

When trying to make something more prevelant in a mix I find a ratio of 2:1 and a perhaps a little gain on the compressor helps loads.

Also compressing old vinyl breaks is pointless imho, they were compressed back in the 70's with valves so no need to do it all over again. Better to layer some nice kicks and snares over the top and make them punchy by layering rather than setting the compressor to 'nuke' mode.

Sidechaining is a really great tool also, used wisely you can give your hats a nice ducky type feel and give your lows some room to breath. Or just produce interesting distortion on your bass by creating huge dips in time to the sidechain source.

Parallel compression works nicely too, for instance you might want to lightly compress something with a 'nuke' setting but not totally kill your source, put a compressor on a send and mix the dry signal with a little wet.

The real trick is just playing with your compressor of choice and learning what the settings do, you can read books on compression all day but books don't further your own style much, experience does that.
 
Here's two examples I made REALLY quickly to give you an idea of what sidechaining and parallel compression might sound like.

1st bar is untouched 2nd is with compression on each example.
The differences are subtle in parallel compression, but that's the idea, hope that helps a bit..
 

Attachments

  • sidechain compression example.mp3
    323.3 KB · Views: 44
  • parallel compression example.mp3
    259.6 KB · Views: 36
DanDnB said:
I still don't get it :confused:

I stupid.
Click to expand...

you have understand that the term
dB refers to a logarithmic scale...
+ or - 6dB refers to twice as much
or half in the case of the minus...
by itself it means nothing...

people miss use this term so often...
dB SPL (sound pressure level) is what
we tend to measure sounds in the real
world in... and 0db just refers to the lowest
average threshhold of human hearing...
some people can hear lower SPLs than this...

the digital referance system dB FS (full scale)
has nothing to do with sound pressure level...
it just indicates how close it is to maximum
volume... in the digital world we have head room
above this... though when it is bounced out
anything above 0dbFS will hard clip...
 
Last edited:
opcode11011 said:
give you an idea of what sidechaining and parallel compression might sound like.
Click to expand...

this means nothing without hearing to original
sample... plus good compression should be
mostly transparent... you should not be able to go...
hey that snare has been compressed...
esp without hearing the original sample... coloring of the sound
may occur due to the circuitry and design of
the compressor or the algorithms used to process
the sound in the software case...
 
Last edited:
opcode11011 said:
look awful in a sound editor.
Click to expand...

what does it matter how the waveform looks...????
its sound... it matters how it sounds... nothing
to do with aesthetics...


opcode11011 said:
Also compressing old vinyl breaks is pointless imho....punchy by layering
Click to expand...

compressing helps with gluing layers together...
and when used well is a great tool to create
cohesion in layers... by setting the attack and
release correctly to can help you make the
hits "punchier..."
 
Last edited:
sook said:
this means nothing without hearing to original
sample... plus good compression should be
mostly transparent...
Click to expand...

opcode11011 said:
Here's two examples I made REALLY quickly to give you an idea of what sidechaining and parallel compression might sound like.

1st bar is untouched 2nd is with compression on each example.The differences are subtle in parallel compression, but that's the idea, hope that helps a bit..
Click to expand...

He included the original sample
 
sook said:
what does it matter how the waveform looks...????
its sound... it matters how it sounds... nothing
to do with aesthetics...




compressing helps with gluing layers together...
and when used well is a great tool to create
cohesion in layers... by setting the attack and
release correctly to can help you make the
hits "punchier..."
Click to expand...

Yeah maybe I went a bit overboard, you're right it doesn't matter what it looks like in an editor, I guess we'd all be composing in photoshop if that were the case, my point was really overcompressing stuff sounds so played out and it really doesn't fatten up your sounds if you overuse it. Personally I think it's better to lower some other elements and leave plenty of headroom for your mix rather than trying to push everything to the limit. But it's different strokes for different folks.

And you're of course right compressing idividual hits, especially with something like vintage warmer really glues your track together although just a little goes a long way.

Sorry, didn't mean to sound like a knowitall, was just trying to convey my thoughts which probably didn't come across too well ;)
 
Back
Top Bottom