Is Dynamics 3 attack time accurate? Techies Please

[McMasters]

This may sound silly, but I ask because I was teaching some people how to use the attack time on their compressor in their PT LE rigs, and the Dynamics 3 Compressor really made me look stupid.

Do any of you tech guys know exactly how attack time should work in a compressor? (generally)

Please keep the “geek speak” minimal so us regular folks can understand.

Let's say you set the attack time to 300 ms - Hard knee - 50:1 - no gain make up and -60 threshold.

a) Should the compressor wait 300 ms once the threshold is crossed to BEGIN the compression cycle?

Or

b) Does it ramp up and hit maximum compression AT 300 ms.

I ask because the Dynamics 3 Compressor does NOT work like any other compressor I own. Or for that matter, any compressor I have every used in last 18 years.

Experiment….
If you take a single snare sample and trim the head tightly and chop the tail of the sample at exactly 300 ms. In theory, when you plop a compressor on that sound and set the attack to 300 ms, hard knee, 50:1, no gain make-up and -60 threshold the compressor should let the full sample pass (300 ms) and not effect it. (correct?) I sent the output of the sample track to another audio track to record and measure the result. And what I found confused me.

Waves Ren-Comp……
Using the Ren Comp set to 300 ms attack, 300 ms release, manual NOT ARC, electro, and smoooth or warm (no difference for this experiment) and no gain make-up. These settings were predictible the Ren Comp lets the full snare sound pass through and leaves the snare sample realitively unaffected. So the Ren comp seems to work correct. Once the signal crosses the threshold it lets the signal pass through until the the attack time is reached then clamps down, as I would expect it to do. As ALL compressors have done in the past for me.

DIGI Dynamic 3………
If I drop a dynamics 3 compressor on the same sample and set the compressor to the same settings and record the result it should be similar however I get a completely different result. The Dynamics 3 clamps at about 20 ms ?????? I can't get the attack to slow down.

I don't have a huge selection of compressors but I did try them all on this sample. With the Smack set to electro and a slow attack, it did what was expected and let the sample pass through. The 1176 was the same, Rude compressor same, C1 same. All were predictable as I increase the attack time more of the sample passes through befor the compressor clamps.

Now I have tried this on a TDM system, LE system, and an M Powered System, all versions of the dynamics 3 Comp act weird.

Any way I had to teach the people using the Ren Comp because the attack works on it.

DIGI I did see 2 students online and checking out prices of the Ren bundle before class was over. I would much rather see them spend their money on the "music production toolkit" (more voices and multi-track Beat Detective) But they do need a compressor that works.

FYI
I did run this experiment with an engineer that has a multiple Grammy recognition’s and 22 gold and platinum records, (who wants to remain anonymous) as well as our Berklee Grad MIDI instructor and we all could not figure out the Dynamics 3 Compressor.
So the 3 of us are all very stupid or this thing DOES NOT WORK as advertised. 300 ms does not equal 300 ms????

PT 7.1cs6 -- 10.4.6 --- G5 dual 2.0 ---- and a squirrel with a funny haircut --- HD 3 w/One Accel card.

[Tiago Silva]

I simply refuse to believe in what the Dyn3 plugs are showing me, and resort whenever I can to my ears. Ever tried watching the DeEsser3 in action? It claims it's doing a lot, but it's not doing much, really.

Now the question becomes "Can I trust my ears". In due time...

[tribase]

I remember learning at one point (long time ago), that the actual definition of attack time (and I think release time as well) is that at the full attack time 66% of gain reduction should have been reached. Obviously that would mean that full gain reduction would be at later than the full attack time and still won't help you with the findings. But maybe that's the definition that you were looking for.

If I remember this wrongly, somebody tell me...

[minister]

ATTACK time is the amount of time before compression starts after the threshold is exceeded (either instantaneously or as a result of further increase) by the input signal.

there are different compressor timing laws, but let's not go into that.

Hardknee, means that the compression ratio applies only to signals above the threshold level. If the compressor is set for Soft knee, the compression ratio gradually increases from 1:1 to the current selected ratio over a range through the threshold area, so that the transition from uncompressed to compressed is more gradual. the knee manages the transition rate between the threshold crossing point and the onset of maximum compression (as defined by the ratio).

the timings are there to prevernt overshoots, but you can also create overshoots with long attack times. sometimes useful, for instance, for tighening up soft drum sounds.

do i trust the DIGI III compressor settings? no.

do i trust the graphics? no.

as usual, i am with tiago. what i see and what i hear are two very different things with the DIGI plugs. not so with a hardware unit or other dynamics processor plug-ins. the de-esser is the really mystifying to me. i set the freq. to 7k and the range to 0, and i still see movement. i see a LOT of GR with the de-effer and i don't HEAR that much.

do i find them useful? yeah.

[Carles Vila]

Hi,
In analogue gear, time constants were defined by the voltage charge/discharge characteristics of a capacitor. On most designs, charge is exponential, meaning that the capacitor never quite charges or discharges fully, but approaches that state. Therefore, the "charge time" or "rise time" is defined as the time passed when the capacitor has charged 63%.
Of course digital can be programmed to emulate (or not) analogue behaviour.
I can e-mail you some test files with which to "see" the behaviour of attack and release times.
I have found that the release characteristics of Dyn III are quite strange. Release is quite shorter than the displayed value and it's characteristics are totaly un-analogue... On the other hand I like the sound a lot.

--
Carles.

[detroitkruk]

ears...

i think it it's a really useful piece...

detroitkruk

[McMasters]

Ohhhh,
Now I understand.
Use my ears, I didn't think of that.
Thanks for the poignant and condescending replies. Very insightful.

Anyway, my point is that the dynamics 3 DOES NOT let 300 ms pass before it clamps. IT IS NOT EVEN REMOTELY CLOSE.
ALL the other compressors let signal pass until the attack time is reached and then they clamp.

So either the Dynamics 3 has the softest knee in existence. i.e. It takes 280 ms to ramp up, so it starts compressing at 20 ms....why???? or it just doesn't work right.

So this thing is a real one trick pony, with nomenclature that is not relevant to reality.

Not a very pro tool.

[Tiago Silva]

That was not what I was saying. I was trying to say exactly what you have found: that the behaviour and the settings of ALL Dyn3 plugs have no connection whatsoever with reality and/or the expected behaviour of similar processors. That means that if you are inputting the settings precisely to match a beat, for example, Dyn3 is the wrong option.

[edit] tripped over myself, here [/edit]

[DigiTechSupt]

Gentlemen - thanks for bringing this to our attention. It does appear that the attack time is not complying with the UI and we're looking into a fix.

[Aquarius]

Digi,

While you're at it, could you implement a Duck mode please ?

Cheers,
A.