GCO- a question to Torben and the compressor experts...

[jens]

The chaps at universal audio did figure out some way of creating the 1176 without upsampling. They have a pretty interesting article about it somewhere. I'll try to look it up for you.

Cheers!
bManic

that would be great, thanks - I'm trying to understand how to use all these parameters - I mean I paid for them after all...

I asked for examples of classic hardware to get a better understanding of their complex interaction...

[jens]

Here are some articles worth reading (try to disregard the obvious hyping of the UAD-1 here).

http://www.uaudio.com/webzine/2003/april/index2.html

and

http://www.uaudio.com/webzine/2003/may/index4.html

and

http://www.uaudio.com/webzine/2003/august/index3.html

and

http://www.uaudio.com/webzine/2004/february/index2.html

That should be enough for now..

Cheers!
bManic

oh, cool thanks - I'm gonna print them out!

[bmanic]

Also, jens, go to this thread to listen to an actual 1176. I've also provided the original untreated file.. then you can try to match it with GCO or any other compressor of your choise.

http://www.kvraudio.com/forum/viewtopic ... 48&start=0

Cheers!
bManic

[jens]

then you can try to match it with GCO or any other compressor of your choice.

but that was my problem from the beginning - the CGO-1 has such a great flexible feature set that it is hard to use all parameters for achiving a certain result - I mean: try just to tweak the frequency dependecy and play around with the linearity know and the threshold...

[jens]

is it perhaps a program-dependent attack?

No. 1176 sidechain design is as linear as they come, with none of the electronics actually producing program dependency. There's no opto components to take into account, or esoteric tube couplings, just fets and capacitors, ie. extremely straighforward.



..just like the sound in fact.

from one fo the UAD articles Bmanic linked to:

For example, both the 1176 and LA-2A have program-dependent compression. One of the things that often confuses people about the 1176 is that if you measure the static compression curves using sine waves at different amplitudes, the compression ratios will appear to be much higher than what is selected on the front panel. However, if you were to look at the 1176's response to a transient-type signal, you would see that the ratios are very close to those indicated by the ratio buttons. The release times and curves are also very program-dependent

[jens]

and from another UAD article:

Remember that the 1176 is a program dependent compressor: The attack and release are program dependent, and so is the ratio. This to me is the heart of the 1176 sound. But there are many other contributing factors…

[3*s]

I hear more distortion from the real 1176 than actual compression.

Which supports my theory that what people "like" about vintage compressors is the huge amount of tube saturation they add sometimes, and not the actual dynamics processing.

[Kingston]

and from another UAD article:

Remember that the 1176 is a program dependent compressor: The attack and release are program dependent, and so is the ratio. This to me is the heart of the 1176 sound. But there are many other contributing factors…

I've read those articles before.

Thing is, we actually analysed a certain 1176 unit with bmanic and the transfer curves are perfectly linear! (we've the data to prove it, although I think we should repeate the analysis using transient sources this time). I also studied the 1176 schematic at www.gyraf.dk in detail and the attack and release are just regular resistor-to-capacitor bleeders. That means they are not any more program dependant than about 90% of all compressors in the world. These time constant bleeders can be networked in serial and parallel to achieve program dependencies of several sorts. But this isn't done with the 1176! Odd. Maybe they just refer to the regular time constant bleeders as program dependency (which is a bit of a scretch).

I'm sure the folks at universal audio know what they're talking about, so I best study this further. Could be something to do with the FETs I've overlooked.

[jens]

hmm... - perhaps the different versions have big differences to each other? But the behaviour Bmanic described did sound like program-dependecy to me - that's what gave me the idea in the first place...

- and perhaps the program material you tested with really didn't have strong enough transients...

[Kingston]

- and perhaps the program material you tested with really didn't have strong enough transients...

it didn't, as I already mentioned.

but see the thoughts about the time constant design. They are peak "program dependant" in the traditional sense of capacitor charging, but that has been done with pretty much all digital compressors as well for quite a while, and is actually relatively easy to simulate, hence the popularity of the implementation.

That's not the same as program dependant vactrols (opto compression) who by their physical construction adjust to the incoming signal, and "remember" what happened a relatively long time a go. That's what you get for using light bulbs in sidechains. (but what a clever idea originally)

[jens]

That's not the same as program dependant vactrols (opto compression) who by their physical construction adjust to the incoming signal

yes, it surely isn't - if it is program-dependent then surely in a very different (more complex) way... (perhaps some specific kind of non-linear and perhaps even frequency-dependent reaction to the transients? - this would it make it way more difficult to measure as well...)

[Kiwburger]

It's hard to sort the information from the misinformation. I find it hard to believe that a software compressor, which has the benefit of lookahead, can't be faster than a hardware limiter.

The PSP MixPressor seems to be fairy fast to me - too fast to the point of distortion.

I was intrigued by a very recent post by a guy at Gearslutz who recently mixed a #1 country billboard hit totally ITB. He said his secret was to set his compressors to the longest attack and shortest release. That's the exact opposite of looking for the fastest attack. I don't really understand this - because that effectively gives you no compression at all. But it does appear to sound better, because from what I can hear, software compression is fairly damaging and less is more.

I agree that Marquis seems to be in a league of it's own. I just tried the PSP Master compressor, and that's the first one that's really impressed me with be able to enhance the air sound in a kick drum. But then I tried it on some soft synths, and that thing is so harsh and crackly sounding. So i guess part of it's magic is the distortion, which doesn't work on most things. But it's made me realise how far behind real hardware this software stuff still is. Quite a long way to go, for some reason. MPL-1 really excites me - it's a one trick pony, but it shows me that there is some potential for a greater software compressor to built, eventually.

[kilroy]

He said his secret was to set his compressors to the longest attack and shortest release.

This is a "balls", or a texturing setting. If you listen to Steely Dan records you will hear a great deal of this kind of compression...it's an Elliot Scheiner signature sound, and a beautiful one.

Basically you want to save the transients from too much shaping with that looooong attack. You go *easy* on the ratio and dig in *deep* with the threshold...swallow alot of signal. Then you set the release fairly fast, which lifts the track up behind the transients and solidifies the tone . Done skillfully this will give you nice fat tracks that sound natural, and have great presence and "stickiness" in the mix.

My impression from listening to tracks that I work on is that folks are generally too heavy with fast attack compressor settings. I suspect that in some cases if they can not actually hear the compression, then they simply feel it is not there.

[Torben]

GCO-1 is very flexible - its intention is, I believe, to be able to mimic the behaviour of almost any single band compressor. Most controls you usually don't neccessarily need to touch at all. They are to finetune the compressor behaviour. All the time since I bought GCO-1 I was wondering why Torben didn't provide certain presets for mimiking the behaviour of certain famous classic compressors. I know the Sonitus Ultrafunk compressor provides such presets although while being good the it surely isn't flexible enough. I have no experiences with a 1176, a Fairchild, etc. - but I would like to be able to try their specific behaviour. However I don't know where their knee starts, I don't know the length of the second release stage of a LA2A, I don't know the specific non-linearity of certain classic compressors, etc.

Questions to Torben: why don't you provide such presets? It surely would be a big selling point, wouldn't it?

It has never been the objective of the Golden Compressor to try and emulate specific hardware, but we wanted to create a flexible product where you can mimic the characteristics of many of these classic analog compressors. For some it would be a selling point to have these famous names on the presets while others would just complain about the preset not sounding exactly like the "real" thing. I think the keyword here is that we are not trying to sell a dream / imaginary celebrity-hardware but a good and flexible tool that can make your track sound good; after all that's what counts in the end of the day as no one listening to your music cares what compressor you used as long as it sounds great.

But we got a patch sharing bank where users are welcome to upload some presets

And b.t.w. about the 'knee'-value: shouldn't it be a db-value just as treshold instead of %?

It does not refer to a specific number of dB as it is also depening on the ratio.

Torben

[jens]

others would just complain about the preset not sounding exactly like the "real" thing.

I see - that's a valid reason....

I think the keyword here is that we are not trying to sell a dream / imaginary celebrity-hardware but a good and flexible tool that can make your track sound good; after all that's what counts in the end of the day as no one listening to your music cares what compressor you used as long as it sounds great.

I totally agree - but as I said I would like to have these presets as starting points to get a better understanding of some of the parameters...

so what about presets like:

Classic opto compressor1, Classic Valve Compressor, Classic FET-Compressor... ?

Then you would not have to mention any names and could still provide what I'm asking for... - I don't really care about the names - it's the parameter values I'm after.

But we got a patch sharing bank where users are welcome to upload some presets

if I had access to the data I was asking for I would create the presets myself and would upload them there...

And b.t.w. about the 'knee'-value: shouldn't it be a db-value just as treshold instead of %?

It does not refer to a specific number of dB as it is also depening on the ratio.

I see - thanks for clearing this up...


another question:

in the manual is writtten, that the second release is up to several seconds long in some opto-compressors - however GCO-1 only has a maximum value of 100ms for the second release - what am I not understanding here?