TranceAddict Forums (www.tranceaddict.com/forums)
- Production Studio
-- Trying to understand compression
Trying to understand compression
Well, ive been trying to get my head around compression, with limited sucsess!! haha
Anyway, in an attempt to try and further my understanding I decided to write a little app in VB.NET - that would allow me to enter some values, and it will spit out some values based on the input.
So, im just wondering if anyone out there who actually knows what there doing with a compressor (unlike myself!!!) would like to try it, and see if the values im outputting seem to be correct, it would be appreciated.
Bear in mind this is the first "program" ive ever written, so theres no error checking etc... (u can get a divide by zero error if you want), also im not sure if I just upload the .exe and it can be used ?! lol, I presume you will need the .net framework (which version I dont know!)
Anyway, enough of me talking crap as usual, the program:
[[ LINK REMOVED ]]
(its only 28kb)
does your prog has timespan output? caus compression is a function of time. if u think u can put one number in and get another number out and assume thats everythings going on your wrong. if your program print some sort of curvediagram with amplitude in the y-direction and time as x then maybe we could put a constant sinus at your input. but again rarely the inputs are constant so you can just throw that app in your bin lol. 
read about it then try it. dont overdo it. try to mix your kick and your baseline into one channel and then add the compression and just turn the knobs and see whats gonna happen. As long as you have no idea what it does dont use it on masterchannel. Its easiest so understand/learn it over a baseline kick mix imo. Ive made some terrible tracks while learning this and i still do after 5 years lol
good luck
It is written, to essentially represent a moment in time for the sake of trying to understand some values.
The "Input" field represents the volume of the audio at this imaginary moment in time. (this would normally be the level going IN to the compressor from your DAW)
that level in is also a function of time and not a constant value.
if ur lucky you can input a square puls and with the right compression settings you can get sinus out. lol. thats not really telling you anything right?
Yes, I realise that mate.
Just think of this program being a "freeze frame" of a moment in time, and you have to put the input level in manually, which would normally be done by the daw (over time)
Thanks for the advice though!
It's the right idea, although you're ignoring common parameters like the attack, release, and knee, and as others have pointed out, a real compressor can't perform the "instant" compression you are assuming unless it has look-ahead (which most digital ones don't and no analog ones can).
Honestly, I think you're complicating a simple thing. A compressor reduces the gain above a certain level as opposed to the absolute volume. Maybe all these intermediate values will help you "understand", but they aren't going to help at all when it comes to applying it, which is almost all about having an intuitive sense of what parameters will cause what kind of change in the sound and dynamics.
I have a habit of over-complicating things! lol
I know it doesnt include things like attack, release, knee, look ahead etc..., but I just was wanting to see if I had the theory right of what is actually happening at a given moment in time with the input signal and what effect the threshold / ratio has on the output volume.
I think doing this though has actually made me understand a little bit better what is actually going on, and now I may be able to apply this knowledge to some real world sounds!! haha
For example say I have a vocal that has an average db of -20db but has some peaks in it at -6db, then I could apply a compressor with a threshold of -20db and a ratio of say 10:1 to allow a more "even" dynamic range (the -6db peaks should now be at -18.6db ( -20 - (-20 - -6) / 10), I have yet to put this theory into practice though.
Easiest way to understand it:
Go download s(m)exoscope vst plug-in, which is free. Adjust the time parameter and hit the freeze button to stop the time to see what compression is doing to the waveform. Or you could just render the output it and look at it through your DAW or audio editor.
why isnt it the easiest way to understand by trying and hear whats happening? 
compression is like a senior citizen with a tv remote.
| quote: |
| Originally posted by Jimb0b For example say I have a vocal that has an average db of -20db but has some peaks in it at -6db, then I could apply a compressor with a threshold of -20db and a ratio of say 10:1 to allow a more "even" dynamic range (the -6db peaks should now be at -18.6db ( -20 - (-20 - -6) / 10), I have yet to put this theory into practice though. |
Agreed with others, ditch the math for the time being. You can use what ever method to help you understand, but for now use the compressor tweak settings, raise the volume adjust settings on compressor hear whats happening.
As DigiNut said - compression isn't about maths. For music, it should be used musically, and the way you use it will vary, depending on the sound that is running through it and what you want to do to the sound. Sometimes extreme pumpy settings sound great. Other times, a really light touch with a ratio of 1.10:1 will be exactly what is needed. Same goes for different attack and release settings.
Beatflux suggested s(m)exoscope, it will show you exactly what the compressor is doing, if you want to see the visual representation of the waveform over time. Turn the compressor on and off and see what difference it makes.
| quote: |
| Originally posted by DigiNut No, no, no, NO. Jesus, compressing a vocal with a 10:1 ratio? This is exactly why I said that throwing around some trivial math isn't going to help you apply it. You shouldn't even be considering trying to squash a dynamic range of 20 dB down to 2 dB. You might use a ratio that high with a very low threshold as a kind of leveler, to eliminate sudden and unintended peaks in the source. If it's really truly necessary to increase the gain by 18 dB then you might use multiple stages of compression. But much more likely you should be adjusting your actual levels, EQ, and other parts of the mix if it has those kinds of problems. My advice is to stop worrying about the nitty gritty of how compression works and start learning about how it's actually used, in practice, otherwise you're going to end up with some awful mixes. |
i always used the compressor to make the kick..and the bass more pumping ..i dont really understand it i just know it makes everything louder.
Try adjusting each parameter and listening to what it does to the sound, reading what it does to the dB output and seeing for yourself.
I feel you learn more by doing and experimenting than by reading and reading and "understanding" what everything is.
When you apply high ratios like 10:1 you're really limiting the signal and removing a lot of transient peaks and the whole sound becomes squashed with no dynamics. Sometimes you want it (not to always to that extent) and a lot of the time you don't.
Mess around and see what you learn in that time frame. Good Luck 
You'd be better off learning the standard settings for different applications and why they're used, and then learning to tweak these to suit. Thats how I learnt.
PS download the sonalksis demo and look at the meters, they'll tell you what you need to know about how compression works in a visual way.
if(Signal.Level > Threshold)
{
Signal.Level = Threshold + ( (Signal.Level - Threshold) / Ratio );
}
Attack = How fast the above is applied to the signal.
Release = Once the above is applied, this is how fast the compression is reduced.
| quote: |
| Originally posted by coroknight if(Signal.Level > Threshold) { Signal.Level = Threshold + ( (Signal.Level - Threshold) / Ratio ); } |
Powered by: vBulletin
Copyright © 2000-2021, Jelsoft Enterprises Ltd.