Presswerk patch/emulation area + inits

[3ee]

Hello,

So, for the last couple of days I've been fine-tuning some initialize patches for Presswerk (will post them here when they are done.)

Update: OK, time for fun : https://app.box.com/s/e483xe26e1qzld6dd0op

The link contains some init and "mojo" patches as well of some experiments, besides the emulation patches.
You'll find workflow info within each patch, hope you'll enjoy!

patches like:

-Init Clean
-Init Limiter
-Init Natural
-Init Pump
-Init Silky
Overdrive, Snap Opto, Spank Hard, Xpand Gate, Hyperdrive etc..

I'm also thinking to download some demo versions of some great vintage compressor emulations to figure out the best Presswerk settings to capture their character/behavior but I was thinking...

So we'll have patches like Init: 160, 2a, 1176 etc

what would be the best approach for this and if someone (preferably one who's familiar with compressors) would chime in to help me test/finetune these and of course try some emulations on their own.

I'll look at the NI VC bundle and NI Solid Bus Comp for starters...
I'll analyse them with Budde's plugin analyzer to get an initial idea and start from there + I'm familiar with the character of the compressors in question so it probably won't be really hard to replicate them into Presswerk... again, I'm not looking for exact replicas, just the closest we can get the character/style of compression with Presswerk's parameter set.

[#rob]

I've compiled a list of all sorts of values of various compressors, but never got around to doing something with it. Might contain some helpful info for patch creation, see attachment.

The best approach would probably be to just go ahead and create something that sounds good, and then let people know to what input volume (dB RMS and/or Peak) a preset was tailored and what its purpose is, so that they have a rough idea how hot to run into Presswerk or how much to compensate.

[3ee]

^ Thanks for the list Rob!

I did create my patches that simply "sound" good to begin with but now I'm looking to make some specific emulations.

Basically all of the patches requires the user to up the input level and/or bring down the threshhold to get compression, probably will note some basic workflow within the patch info just in case.

[3ee]

just remembered... we should totally do a Blockfish patch!

[3ee]

here's a useful general info about classic compressors when building your patches:
http://www.soundonsound.com/sos/sep09/a ... essors.htm

[OneRoomStudios]

I can probably help with this. I've been an audio tech for over a decade and I've repaired/modified/built just about every classic compressor under the sun. I have schematics for all of them and know quite a bit about how they work. I really like u-he's design and concept, and I think it would be fun to try to design some "similar to x" presets for Presswerks.

Aside from the obvious 1176, LA-2A, LA-3A, 160A, SSL, Fairchild, Sta-Level, etc, I think it would be fun to try some quirkier models too (maybe 2254/33609, PYE, Collins, TLA-100, LA22, etc). I'll try to dig up the data I have on compression curves, THD, frequency response, and envelopes for these units and see if I can get Presswerks to mimic their behavior.

[3ee]

OK, thanks! ...but one at a time, for example... I finished reading the dbx 160 manual + some more info and for starters we know:

it's a FF, RMS design, relatively fast, hard-knee. (So I guess that the auto attack and release is manly due to RMS)
Been tweaking a preset for about an hour + hearing some UAD dbx 160 audio videos...

My patch is starting to take shape.

Now I gotta download the NI VC160 demo to test some graphs and finally finetune by ear... then I'll post for for beta-testing asap so keep an eye out!

[OneRoomStudios]

Some of that info depends on which model of "DBX 160" you're talking about.

There was the original DBX 160 (known as the "DBX 160 VU"), which had a hard knee response, but the 165 and 160A, which are probably more common, both have the famous "overeasy" soft knee switch. Then there's the 160X and 160XT, some of which used THAT VCAs instead of the original discreet Blackmer VCAs. Then the 160S, which has far more bells and whistles than any of the previous models and is stereo. Each one has different distortion characteristics, slightly different compression curves, and generally different behavior.

So first, which "DBX 160" are you trying to model?

[3ee]

It seems that the simple VU
Now finetuning against this one: http://www.native-instruments.com/en/pr ... ts/vc-160/
I don't know which exact model it suppose to be but I'm guessing it's the VU)

^ Do you know what the RMS window would be for the dbx 160?
I seem to be getting close with higher values but a value of 15 ms? that "seems" too much as I always thought that is was a -fast- compressor.

I really need a number or a suggested way to test with audio tones or something, since I can't finetune the attack and release (especially the attack) if I don't know the RMS value. anyway, getting closer.
Once I get an RMS number I think it won't be long till the patch goes live for testing.

[OneRoomStudios]

Judging by the GUI, I'm guessing they're going for the 160 VU there.

The RMS window should be turned all the way up (30 ms). That might seem strange, but it's not because we're telling the attack to "wait" for 30ms before in engages, rather, we're telling the compressor "always look at RMS, and never peak." In fact, because of this, the attack time should be less than the RMS window. In the case of the 160 VU, I think an RMS window of 30ms and an Attack of around 10-12ms would be a good place to start, but of course, that depends on the incoming levels.

This might be helpful in understanding how RMS detectors playing into VCA-based compressors: ftp://ftp.dbxpro.com/pub/pdfs/WhitePape ... icsDuo.pdf

[OneRoomStudios]

This is where I would start for a 160 VU:

Detection: FF
Channel Link: 80% (these were mono units build in the 70s/80s - not perfect

Threshold: let's start with -25 for now
Ratio: let's start with 4 (12:00 on the 160 VU)
Soft Knee: 5% is probably a good place to start (nothing's perfect)
Non Lin: 1.00 (these were actually fairly linear, but not perfect)

Attack: 15ms
RMS window: 30ms
Release:80ms
Adapt: 0%

Saturation:
Post
Amount: -30db
Dynamics: 75% (these early VCAs were notorious for dynamically distorting transients)
Warmth: 7db (still not entirely sure how this functions)

oh, and "ANLG" should probably be selected.


That should get you in the ballpark

[3ee]

Many thanks! puzzle pieces seem to fall together with 30 ms of RMS.

A tiny bit of negative value for the sidechain delay to make the attack -click- smoother (saturation also helps here) and attack is at about 8 atm... still finetuning.

I actually have many values set to about where you pointed, (thanks for the extra tips)
I have the release on precisely 80 ms as well.

be right back!

[OneRoomStudios]

After downloading a demo of the NI VC160, these are some tweaks I would make:

They seem to have a slightly faster release, maybe 65ms.

They have way more distortion. Try bumping the Amount up to 3.5db, the Dynamics up to 100%, and the Warmth up to 15dB. Personally I think that's exaggerated, but it sounds more like the NI plug.

Technically, the delay should only be negative for a digital compressor (analog can't really do look-ahead).

[3ee]

Warmth: if set to 0, the saturation will act equally across the spectrum.

Bring it up to get less saturated lows and more for the highs resulting in a fuzzy-smooth/warmer saturation. I had it on 0 for this.

I don't necessarily want to emulate the NI version... but to capture the style/character as close as possible for a 160 init patch.

[3ee]

Technically, the delay should only be negative for a digital compressor (analog can't really do look-ahead).

Yeah but I'm using the look ahead here only to make it sound better (-0.50 ms) making it sound more "analog" perhaps?

HP sidechin filter is @ 0 ...did the 160 have a bit of HP under the hood perhaps?