I've set up a thread for discussion and posting of user modelled studio hardware using the Proteus model.

I'll kick off with a 1:1 port of my Neve portico II Channel, better than any software plugin ive ever heard.

Settings, line in> transformer> VCA Compression (FB) mode, 30ms attack, 100ms release, modelled up to 8dB gain reduction, Silk Red at 12 o clock, perfect for individual drums or drum buss among a gazzion other uses ;-)

Here you are:
https://rg.to/file/754f7ef2c3b5f707...ms_rel_max_8db_GR___Silk_R_Model_ESM.png.html
https://rg.to/file/a39eb9fd4a197a6f..._30ms_100ms_rel_max_8db_GR___Silk_R.json.html

 

Thanks. I'll give it a go.

 

Can you give a quick explanation on how to use it?

 

It works with that :
https://guitarml.com/#products

 

Okay, I just tried it on a drumloop loading the .json file into the GuitarML Proteus plugin. I'm not sure it improves the sound. There is quite some low frequency / subbass loss. I double checked with Bertom EQ analyzer which verified what I was hearing (see image). Does the Neve Portico hardware do this too?

 

Also, on another drumloop I just tested there seems to be a strange reverb-tail like noise when the .json program is loaded into Proteus. Is this supposed to be the release tail from the compression? I'm not sure. I deactivated the built-in Cab IR in Proteus of course!

You can try it yourself using the attached drumloop, loading Proteus with the file (CAB IR off) and turning up the gain to around 2 - 3 oclock.

 

Hey y'all! If any of you are totally not digging the aliasing issues in Proteus, which ain't fixable, there's this cool plugin called BYOD by ChowDSP, and they've got a GuitarML module for hosting Proteus files. Found out from the GuitarML creator. It's great at oversampling, and also free https://chowdsp.com/products.html

 

Proteus is sample rate limited, like NAM.

48 kHz for NAM
44 kHz for Proteus

if i remember well
So oversampling will not work ... it will be worse

It is related to machine learning algos it seems ...
Some ppl reported ToneX solved the problem by doing ... a low cut over 20 kHz

 

Hey can I ask you a question.. why is oversampling worse when you have your project and interface set to 44.1 does it not get rid of the aliasing?

 

Because the plugin itself don't like oversampling.
So pushing it over the advertised value will simply create internal troubles.

Most plugins have better sound with oversampling.

NOT those two. It is related to the neural network code it seems.

https://github.com/sdatkinson/NeuralAmpModelerPlugin/issues/59
https://github.com/GuitarML/Proteus/issues/11

 

So guys, anyone double-checked my findings? The Portico hardware capture seems to take away some lowend and there is also this strange reverb like tail (particularly when you turn up the gain in Proteus with the internal CAB IR off) on the release of the compression, which you can particularly hear on drumloops / snares (you can test it with the drumloop I posted above or any other drumloop that is not too busy).

 

I'm not digging plugin doctor or anything like that, but used it on a "transparent"/"Spinz" 808 yesterday and got an amazing sound from it. Tested on a drum bus once, didn't noticed nothing strange on the sound also, but a heavy compression for the main aux.

 

I just double-checked with Bertom EQ analyzer to verify what I was hearing and it indeed showed what I was hearing.

Can you please try the Portico capture on the drumloop I posted above? Please listen to the tail of the snare. I'm not trying to criticise, I actually like what it's doing overall, but I don't like the artefacts I'm hearing. I'm just giving feedback to help improve, since this approach seems to have great potential.

 

Hola bro, kindly please attach these files instead? I am having trouble with Rg.to
Thanks in advance

 

make sure that the IR switch is to the left or you will get weird sound ;-) hope that helps

 

the new site is up and running now ;-)

https://analogxai.company.site/

 

you def will lose some sub or low end even with top end devices when compressing a lot, this model has max 8dB of gain reduction just to showcase the model really and is great for single channels or parallel processing, groups / busses etc need a less heavy approach. the plan is to include 3 profiles for the compression models, say 2dB, 5 and 8.. the more compression with the portico the more you lose the lowest frequencies, it is what it is just like the hardware - i hope this helps

 

what aliasing? this is analog man hehehehee

 

Proteus is designed for guitar amps. It doesn't do rack gear properly. It gets some aspects right but then kind of has a specific box tone of compression characteristics that make all gear models on it sound the same.

Everything has the same LPF and HPF, the same knee and ratio of about 2 for a little while before returning to 1:1 and a similar attack and release curve.

Feed it a slow opto or a crushed 20 ratio 1176 and it'll still output the same compression model.

 

That has definitely not been my experience using Proteus and modelling Studio gear so far, I am having great results and in fact I believe outperforms coded plugins in many areas, specifically Saturation, capturing the real non linearities of analog gear. I certainly would not say that everything sounds the same ;-) but each to their own opinion.