Compression Confusion (Warning: Long Post)

bdam123

Member
So before tonight I thought I had a pretty good grasp on how to use a compressor. I've done dozens of mixes "in the box" and have had a good amount of experience with compressor plug-ins.

I've recently purchased my first hardware compressor. Its an API 527.

First I'm going to list what I know about compressors so you can all can see that I have a firm grasp on how they work (or not)

1) A compressor is used to tame a piece of audio by bringing down its loudest parts and bringing up its quietest parts.

2) The threshold determines the level at which the compressor begins to compress the audio. Anything above the threshold will be compressed and anything under the threshold setting will be left unaffected.

3) The Output/Gain is used to bring the level of the audio back up once its peaks have been reduced. This effectively makes the lower passages louder and brings up the average loudness of the audio.

4) Attack and Release determine the amount of time it takes the compressor to begin to compress the audio and the amount of time its takes to return back to its original state, respectively.

5) The ratio determines how much audio over the threshold will be let out. For instance a 4:1 ratio means that for every 4db over the threshold only 1db will be let out.

I think that pretty much does it and should prove that I'm not a complete retard when it comes to using one (or visa versa haha)

Now for my issue (or should I say issues) with this thing.

Experiment #1

My output is set to zero, my threshold is on the lowest possible setting,
-20dbu, and my attack is the fastest possible. I'm feeding the mic a constant sine wave.

So with the the ratio at 1:1, I feed a pretty healthy signal and I can the see the track meter on my pro tools stay in the green, This signal is loud and its not even touching the yellow. Now with the ratio set to ∞:1, as soon as I feed it the sine wave I see the track meter clip. Am I bugging out or does this seem backwards? Isn't an ∞:1 setting essentially acting as a limiter? So shouldn't that setting be bringing everything down to -20db?

Now when I just look at the compressor, the Gain Reduction meter is working as it should. On a 1:1 ration I see no activity on the meter and on an ∞:1 I see lots of action. So in other words my compressor looks as if its working right but that isn't whats happening in my recording.

Experiment #2

My output is set to zero, my ratio is at noon (which I'm assuming at this point is a fair amount), and my attack is the fastest possible. I'm feeding the mic a constant sine wave.

Watching my meter I'm riding the threshold. As I get to -20dbu the signal is becoming hotter and when I turn the knob to 10dbu (the opposite setting) the signal becomes quieter. Now isn't this opposite also? Shouldn't bringing down the threshold set a level in which the audio gets compressed? So the lower the threshold the more its is being breached by the signal and the more compression being applied, right?

I don't know if its lack of sleep but I feel crazy. Please someone point an obvious flaw in my perspective and end this confusion. Thanks for any help and I apologize for the long post. Peace.
 
Well first of all I hope you looked over the manual front to back then back to front.
I all ways read the manual of new gear before and during play time of getting to know a new toy.
With that said brick wall limiting is tough when you have higher ratios it's harder to dial in.
Other things to look at with the API is .... the thrust function on? Is the auto gain on? Is the soft or hard knee engaged?Indicated by old and new respectively. And is the feedback or feed forward methods on?








:cool:
 
I'm not intimately familiar with the 527, so I'm kind of groping in the dim light here, but I think maybe some of it might be attributable to a combination of the type of experiments you're running in combination with some of the "personality" of the way the 527 works.

First, I'd dump the sine waves and see what happens with real world signals, especially if you have the unit set to "New" which depends upon the use of RMS detectors for modulating the gain control voltage.

Also, I noted this in the spew on the 527:
API's web documentation said:
The output level remains fairly constant regardless of the threshold or ratio control, much like the "more/less" Ceiling control on the API 525 compressor. This allows for live adjustments without any noticeable gain changes in the program level.
I'm wondering if this means that when you set the gain reduction ratio to infinity on a sine wave that it tries to automatically adjust the makeup gain by the same infinite amount in the opposite direction (or something like that.)

G.
 
True what glen suggested about using the sine generator. Get in there and use real life application ..... goes like this ~ check 1-2-3, check 1-2-3 Just don't count to 4 or then you have to lift. :D







:cool:
 
I started out with a real vocal, I sang rapped, and yelled in the mic. It did the same thing. After my voice got to tired i decided to use the sine wave.

Is it possible I got a faulty unit? Am I right to assume that the results I listed above are the exact opposite of what should be happening?

With the threshold at -20dbu and the ratio set to 1:1, screaming at the top of my lungs barely got the signal to hit the yellow on my track meter.

SouthSIDE, its funny that you mentioned that because one of the things i did was ride the ratio knob as my vocalist performed a fairly loud spoken word. Starting at 1:1 and moving higher not only could I see the level rise I could also hear it get hotter.

I just don't know whats going on.
 
Seems to me you have described automatic gain (reduction) compensation.
..Glen's blurb from API..
The question is when they (compressors) do it, do they get it right? –ie doesn't over compensate, and you end up having to undo some of it.
For example some times you do a little reduction (peaks, some of the tops' down a little let's say) and the loudness hasn't really changed. Along comes auto-gain lifting the whole body of the sig.
I presume sometimes it's going on in the background and I'm not even aware of it, but when I hear that effect, I'd just as soon let me decide.
 
They don't come out and say it, but it sounds like the unit has automatic gain compensation. I find that automatic gain compensation just pisses me off in most situations, the exception being a mastering limiter.
 
Reverse the scenario, go to 1176. (UAD plug here if that matters)
No threshold to lower. You have to raise the gain in to change the reduction. Half the time now I have to mess with the output 'cause I didn't want a volume change.
 
Reverse the scenario, go to 1176. (UAD plug here if that matters)
No threshold to lower. You have to raise the gain in to change the reduction. Half the time now I have to mess with the output 'cause I didn't want a volume change.

AHHHH.







:cool:
 
Thats so annoying, I can understand if it was a straight up limiter but this just makes it really annoying to use. Whats the Output knob for than? Just overall output signal, not make up? Well obviously not make up. Great.

Well thanks for pointing that out to me everyone. Gonna go mess with it some more.
 
So in your experiment if it is auto gain compensating, you'd have to use the output knob to then lower that sound? (If it's too loud)
 
. Now with the ratio set to ∞:1, as soon as I feed it the sine wave I see the track meter clip. Am I bugging out or does this seem backwards? Isn't an ∞:1 setting essentially acting as a limiter? So shouldn't that setting be bringing everything down to -20db?

.
Without thinking too much, still havin coffee here :) I'd say you can't divide by zero.:confused:
 
Back
Top