Will I lose my decibels?

P

PhilM

New member
Hello all :eek:
I'm trying to understand Compression.
After reading the following,.....

*****************************************

"We often just want the loudest notes tamed, not broken. This is where compressors come in and with them a RATIO control. All the RATIO control does is set how quiet we want the loudest stuff. If RATIO is set to 1:1, then nothing really happens. Might as well hit BYPASS.

At the other extreme, lets say a ratio of 20:1, then we are probably limiting rather than compressing. In fact the sound would have to get 20dB over the threshold before the compressor lets even 1dB more out. Thats what the 20:1 means.
If the THRESHOLD is set to 2:1, then it only requires the sound to get 2dB hotter to let 1dB more out."

******************************************
my conclusion is thus.......

If I have a 2:1 ratio and my signal goes 1db over the Threshold then that 1 db will be cut and lost.

If I have a 4:1 ratio, and I go 3 db's over, those 3 db's are
also history.

Would this be correct?
cheers
 
Last edited:
PhilM said:
my conclusion is thus.......

If I have a 2:1 ratio and my signal goes 1db over the Threshold then that 1 db will be cut and lost.

If I have a 4:1 ratio, and I go 3 db's over, those 3 db's are
also history.

Would this be correct?
cheers
Click to expand...

No I don't think that's quite right. In your first example the 1dB over will be reshaped as a 0.5dB over threshold, in the second example, as a 0.75dB over. Your examples matched what a limiter would do at different thresholds, not a compressor at different ratios.
 
So is this an incorrect statement...

"In fact the sound would have to get 20dB over the threshold before the compressor lets even 1dB more out."

.....Isn't this saying you "need" to get over the threshold, by a db value greater than the ratio number, 2, 4, 20 or whatever, to let anything through at all. :eek:

Arrr!!!! think I might have it..... I shouldn't have said " let anything through at all"......I should have said ....to let another db through to then be compressed.

Is that it.................Man!!!! the things one has to get one's head into to try and make a good recording!!!! :rolleyes:
Hell, I just want to play my guitar :D
 
Last edited:
PhilM said:
So is this an incorrect statement...

"In fact the sound would have to get 20dB over the threshold before the compressor lets even 1dB more out."

.....Isn't this saying you "need" to get over the threshold, by a db value greater than the ratio number, 2, 4, 20 or whatever, to let anything through at all. :eek:
Click to expand...

No, it is saying that you will get a fractional value of a decibel over the threshold if the signal is over the threshold by less than the ratio.

To complicate matters further, many compressors have a "knee" setting that smoothes their behavior near the threshold.
 
Last edited:
PhilM said:
Arrr!!!! think I might have it..... I shouldn't have said " let anything through at all"......I should have said ....to let another db through to then be compressed.
Click to expand...

Closer, but no :) Once over the threshold and reduced in gain, it's already BEEN compressed.
 
mshilarious said:
Closer, but no :) Once over the threshold and reduced in gain, it's already BEEN compressed.
Click to expand...

Good, thanks mshilarious.
It's time for bed now, but I will be having a big ponder
on your last statement,...and..... I WILL BE BACK!!!... :p
 
"Ratio" is exactly what that word means, it is setting a ratio, not a wall.

A 3:1 ratio means that for every 3dB over the threshold going into the compressor, one dB comes out. Put another way, take however many dB over the threshold you have going in, and divide it by 3 to figure out how much will come out. At 4:1, you take whatever is coming in over the threshold and divide that over value by 4.

That does not mean that if you are 2dB over threshold at a 3:1 compression ratio that nothing comes out, that all 2dBs are lost. It instead means that the same ratio of compression is applied to that 2dB, compressing it to 0.67dB.

Remember this is all over threshold. If you have the threshold set to, say, -10dBFS and you have the ratio set to 3:1, only that part of the signal above the threshold of -10dBFS is affected. So if you have a signal that comes it at -8dBFS (2dB over the threshold), it will actually be compressed to -9.33dBFS.

G.
 
Exactly, it's simple math. well, maybe not simple but math. 3:1 means you divide by 3 for every db coming in to see how much is let through. at 3:1, a 6db increase over the threshold will allow a 2 db increase. MS and Glen got it...like big surprise.
 
Ok...so I too I have been on the quest for knowledge to uncover the mysteries of compression, as right now I know very little on the subject, when I ran across this thread which I think has given me a bit of an epiphany...

Could anyone clarify whether the conclusions I have drawn are correct.

Ok, so lets say for example I have a guitar track.
If I had a frequency in it that I wanted to completely get rid of, would I just use and EQ, and cut that frequency?

However if I had a frequency I just wanted to limit some, and not completely cut, I would use compression?

My theory is that it works in the sense of...if you apply a limiter to a track it limits the whole track at a certain level....and if I use compression it works like a limiter, but only limits the frequencies you want it to, rather than the whole track?

It my theory right or have I got the wrong end of the stick entirely?
 
Not quite. A standard compressor will compress anything that is over the threshold, regardless of frequency. To reduce a particular frequency, you would use an equalizer. If you don't want to cut the frequency completely, reduce the amount of cut.

I believe multiband compressors can compress a fraction of the frequency spectrum. Not too sure about that one. I'm certain somebody with more knowledge will be along to clarify.
 
Yeah I was thinking more of multiband compressors. I mean obviously if you wanted to compress a lot of different frequencies using multiband compressors, I'm guessing you would have to use a whole load of them to accomodate.

But then if your sound is that bad that you need compress that many frequencies, I guess the quicker solution would be to just rerecord it lol :p
 
Yeah, there are multi-band compressors, but those are not the solution to what you're looking for for a couple of reasons.

First, multi-band compressors typically only split the frequency spectrum into three or four or so wide bands. It's like having a three-band EQ (bass/midrange/trable). You can't really zero in on a single frequency.

Second, multi-band compressors are very easy to misuse, expecially for someone who is just learning about compression in general. It's best to walk before running.

If you have a frequency that you just want to limit some, use an EQ. That is exactly why they exist.

G.
 
Yeah I haven't actually used compression on any of my music except a little on the bass, and that was a preset...it just seemed to work nicely. I have just been using EQ. Just thought I'd try to find a little out about it to see if it could improve my music any. I do underststand however that using compression with no understanding of it is one of the best ways to kill a tune lol.

Thanks for the help :)
 
legionserial said:
Yeah I haven't actually used compression on any of my music except a little on the bass, and that was a preset...it just seemed to work nicely. I have just been using EQ. Just thought I'd try to find a little out about it to see if it could improve my music any. I do underststand however that using compression with no understanding of it is one of the best ways to kill a tune lol.

Thanks for the help :)
Click to expand...
I'm not trying to turn you away from compression (though I DO think the availability of compressors is one of the worst things to happen to home recording in the past 20 years, but that's not my point). All I'm trying to say is that for your description of the situation - knocking down certain frequencies - a compressor is simply the wrong tool for the job, and an EQ is the right tool.

All a compressor is meant to be used for is to control the dynamic range of something. Because there are limits all over the place as to how loud and how soft we can have our sounds, if we record a sound whose dynamic range (i.e. the difference in volume is getween it's quieter forments and its louder ones) is extreme, it can hurt the perceived quality of that sound. A compressor is used to compress (hence the name) that dynamic range to make the sound more accessable to the human ear, especially when thrown in with a lot of other sounds. Leaving special effects aside, that is really all they should be used for.

G.
 
SouthSIDE Glen said:
First, multi-band compressors typically only split the frequency spectrum into three or four or so wide bands. It's like having a three-band EQ (bass/midrange/trable). You can't really zero in on a single frequency.


If you have a frequency that you just want to limit some, use an EQ. That is exactly why they exist.

G.
Click to expand...

Sorry Glen, but I gotta disagree here. Most any good MBC has adjustable band centering and band width, so zeroing in on a specific frequency, or at least a very narrow band is completely possible. On a three band MBC, you just leave two of the bands at linear response, dial in the center freq on the third, narrow the bandwidth all the way down, and then set the threshold and ratio to fix the offending notes.
Eq's, as you well know, don't limit. They cut or boost statically. That's the right approach if the offending frequesncy is consistantly offensive, but if it jumps out here and there, and cutting that freq will fix those places but leaves it weak in other places, then the MBC is the better choice.
I think we discussed this in a mastering thread. I know giving an MBC to the preset pushing neophites is like giving liquer to the indians, but lets not throw a good tool out the window because a lot of people don't use it intellegently.

-RD
 
You're right, Robert. Many (perhaps most?) MBCs have parametric controls on them as well. I stand corrected on that one. Thanks for the proper correction on that. :)

But I posit that the use for such a device approaches zero. I stand with the claim that if one has an offending frequency out of an instrument that they wish to tame (which is what was asked about here), a parametric EQ is the way to go most of the time instead of a parametric MBC. The reason I say that is that the offending tone is rarely offending in a dynamic way.

Instead, it's usually (in my expereince, anyway, all four-letter abbreviations apply ;)) that the offending frequency is generally unattractively - and statically - boosted because of the characteristics of the instrument, the recording environment/setup, and/or signal chain coloration. In those cases, equalization is a direct eye-for-an-eye response to the problem.

There are be some circumstances where pinpoint parametric compression might be the best tool for the job, but those are usually stratified cases that really require the matching support equipment (i.e. top-shelf control room environment) and ears to not only identify, but to properly correct.

I hardly ever use such a device in my own room, not because I don't understand it or know how to use it, but becaue I a) have little-to-no real-world need for it and b) know that half the times I do use it I am fighting artifacts in my own limited studio and ears more than I am actual symptoms within the content itself.

For all these reasons, I stand by the conclusion that EQ is the right tool for the job described, and furthermore, that MBCs are about as useful to the average prosumer engineer as breast implants on a bull.

G.
 
You must log in or register to reply here.
Back
Top