NDSP Quad Cortex

I'm probably preaching to the choir with some of this but maybe a few folks find it helpful if they're into "Neural Capture" and aren't getting good enough results.

May not work for all use-cases but found some of these definitely helped in mine.

Tip #1:

Before starting the Neural Capture, ensure the input impedance on the instrument is set to what your amp has.

Most amps have a 1Mohm input so getting IN 1 and IN2 to 1M will help (I know if you goin into IN2 with an XLR cable, it makes no difference as the input flips to another value for MIC signals)

Tip #2:

After the "Neural Capture" guide / connection diagram screens, you end up at the point where you can start the Neural Capture process and/or adjust levels.

Do NOT boost the "IN 1 LEVEL" - just leave it as is. Don't go for peaking at -6dB or any other value - leave it as is.

Bump the IN2 (coming from the Mic or Loadbox) if you need it / the capture to be louder.

Tip #3:

Unless you have a good quality USB cable with ferrite choke between your PC and Quad Cortex (ex: https://www.thomann.de/gb/udg_ultimate_usb_2.0_cable_s2bl.htm), remove the USB cable from the Quad Cortex to the PC.

I was getting some annoying interference / noise that messed my captures & found out by accidentally disconnecting the USB cable while at the Neural Capture screen.

It's a good idea to have the bare minimum of cables connected to the QC for the Neural Capture process; I'd leave everything else disconnected aside from the stuff you're told to hook up during the "Connection Diagram" screen.

Tip #4:

For folks using the vanilla (non-IR) Suhr Reactive Load, don't use the XLR-out on the Suhr to go into the Quad Cortex.

The Suhr RL has a high output impedance (probably too high for the QC's XLR input to work well with the non-IR Suhr Reactive Load XLR's output).

Use an unbalanced cable if you're shooting "direct" / loadbox-only (no cabinet) profiles of your amp if you must (pick a good quality cable like the Sommer LLX stuff).

What I found out last year shooting NAM profiles is that the output transformer on the Suhr hypes the highs and has a sharp attenuation from about 100Hz downward (too sharp for my liking). It does the same thing even if you use the THRU port with a real cab hooked
up.

I got a Behringer GI-100 DI box (very cheap) and connected it between my tube amps and the Suhr Reactive Load for silently shooting direct captures.

The Behringer GI-100 can also go between the amp and a real cab, it'll apply the impedance curve of the real cab to the capture but the "reamping" phase will not be silent :bag (ie. you'll hear the alien sounds coming out of the cab :alien:)

Tip #5: DISREGARD not needed (seems to only affect the output volume!)

This one's important for consistency of captures & that they play well on other folks' QC units if they use them.

After the Neural Capture finishes and you're at the "Cortex" / "Reference" page, apply a 5.5 dB boost to the level / gain of the capture right before you hit "Save" and put a name on it.

The Quad Cortex' "Capture Out" port used to pump the training signal / sequence into the amp is rated at 9.5 dBu (confirmed by NDSP support);

The QC IN1 is rated at 15 dBu (also confirmed by NDSP) when plugged in with an unbalanced cable and set to 1Mohm.

Saving the capture this way will help it behave as close as possible to your gear; in essence you're not applying any boosts before the capture process starts but are somewhat calibrating / making up for the delta between the maximum gain of the QC's input and "capture out" ports to help the capture fidelity.


Tip #6:

If noise is still present when you have stuff hooked up, try putting an iso-box like the Palmer PLI-01 or the Palmer PLI-05 between the QC's "Capture Out" port and the tube amp's input (or pedal etc - depending on what you're trying to profile).

I had some pretty aggressive settings on my Ceriatone King Kong 100W where the profiles didn't turn out as the amp and throwing the PLI-05 in between the QC Capture Out and the Input jack of the Ceriatone helped get good captures.


Sorry for the wall of text but since I got another Quad Cortex recently, Neural Capture was on my mind and wanted to try and apply some lessons learned from the time I sold the first one & moved to exclusively using NAM (neural amp modeler) to profile my gear and present day.

All in all, I like the way the Neural Captures turn out; even the "direct ones" that are shot using a loadbox.
I found you on the Cortex Cloud, but when I got a few of your captures you shared a few days ago, these were really low in volume..just a heads up, on a few I had to crank the volume up to max on the output of the capture. I aim for -12 - or -10 when making capture, the factory captures is at this level as well or around these levels
 
I found you on the Cortex Cloud, but when I got a few of your captures you shared a few days ago, these were really low in volume..just a heads up, on a few I had to crank the volume up to max on the output of the capture. I aim for -12 - or -10 when making capture, the factory captures is at this level as well or around these levels
I was more concerned / focused on them behaving like the real deal & final output loudness wasn't a criteria at this point.
As I get back to the Neural DSP / Quad Cortex ecosystem I'll build a more consistent way of getting these done.
 
The Quad Cortex can't capture Rectifiers?? WRONG! At least, it was able to capture my favorite Rectifier, the Single Rectifier Series I. It's the most brutal by far imho:


that sounds great to me and yes , I always loved the single recto too I find they were less boomy in the lows a bit tighter more like the Triple
 
It looks like they are actually demoing the Pcom stuff now so the Beta must be getting closer to release , they don’t focus much on 3,0 other than plug in support , I suspect that first delay Bea hits might be the circular PCM but they never say

 
It looks like they are actually demoing the Pcom stuff now so the Beta must be getting closer to release , they don’t focus much on 3,0 other than plug in support , I suspect that first delay Bea hits might be the circular PCM but they never say


Just starting the video, was it intentional to use the 3.0 beta in this demo?

Edit: NM, they are walking through the new functionality. Interesting!
 
Only got around to that Rabea video now. That integration is very nice. And I think I understand side-chaining now. This is definitely a cool feature.
 
I'm gonna try, when 3.0 drops, record a demo of me playing Smells Like Teen Spirit (to keep it easy) all the through with one guitar. I'll have the midi change the tone through the different parts and activate the doubler on the chorus and even use the looper to sample a piece of the chorus to later playback behind the solo.

I think I know exactly how I'm going to do it.
 
It looks like they are actually demoing the Pcom stuff now so the Beta must be getting closer to release , they don’t focus much on 3,0 other than plug in support , I suspect that first delay Bea hits might be the circular PCM but they never say

This also shows that it's pretty easy to make relatively complex presets straight from the QC UI.

I do think they should indicate the letter of the encoder assigned to each control when editing a block tho. There's one point in the video where the other guy turns the wrong encoder that is mapped to nothing.

The plugin stuff ended up being basically what I expected. It doesn't sound any better or worse than what is already in the unit. I'm sure when it launches someone will do a comparison between the plugin and the QC to see if there is any sound difference.

Wouldn't be surprised if NDSP spends the rest of the year adding support for all the plugins and that sucks for people hoping for some actual improvements.
 
This also shows that it's pretty easy to make relatively complex presets straight from the QC UI.

I do think they should indicate the letter of the encoder assigned to each control when editing a block tho. There's one point in the video where the other guy turns the wrong encoder that is mapped to nothing.

The plugin stuff ended up being basically what I expected. It doesn't sound any better or worse than what is already in the unit. I'm sure when it launches someone will do a comparison between the plugin and the QC to see if there is any sound difference.

Wouldn't be surprised if NDSP spends the rest of the year adding support for all the plugins and that sucks for people hoping for some actual improvements.
agreed with the plug in stuff unless it amps you don't have like the Tone king you already have all the Archetype tones as their core sounds are in the QC models its just the little tweaks by the artists.

For a similar example the Gojira Dimarzio pickup its basically a 36 anniversary PAF with a 1.5 db bump at 160 and a scoop at 500.
Same thing applies here 100% his Archetype amp is the 5153 el34 Red channel , with probably a few small tweaks .

The place that I could seem some advantages to the plugin especially something like the Petrucci, is to flesh out some of the Fx.

If I was NDSP for the Quick Fix I would just port the spring reverb from the Morgan
and create a Delay Engine that will allow you to roll you own .

their original Flanger / chorus/ phase modulations were pretty bad .

but the added chorus engine, flanger engine , MXR flanger . Dream Chorus, TC Chorus made it complete really well IMO
they need to. do that with the delay . spend 6 months and come up with the Delay Engine , port the Morgan spring and throw in a couple of Amp models Like the Bogner XTC101B. blue/red and I would be pretty happy .


My guess is though , they will likely wait to offer the effects until that individual plug is ported and then charge for it , but who knows their recent survey everyone says the delays / verbs need improvement , so if they were smart 3.1 would be targeting those areas
 
There’s people that complain about lack of plugin compatibility every time the QC is mentioned and others who complain about time being spent on plugin compatibility. Damned if you do and damned if you don’t.
 
There’s people that complain about lack of plugin compatibility every time the QC is mentioned and others who complain about time being spent on plugin compatibility. Damned if you do and damned if you don’t.

"If we come out with the 8 x 12 people will complain it's too big and heavy. If we do a 4 x 12 people will complain they now have to lug two cabs instead of one. We're damned if we do and damned if we don't." - Jim Marshall circa 1965

same as it ever was.......
 
There’s people that complain about lack of plugin compatibility every time the QC is mentioned and others who complain about time being spent on plugin compatibility. Damned if you do and damned if you don’t.
They did damn themselves by advertising plugin support initially.

There's probably a vocal minority who had already bought a bunch of plugins and then got a QC on the premise they will be able to put their plugins into the hw unit "soon".

Then there's another vocal minority of NDSP haters who will bring that point up every time, until plugin support is there. Then they will move to complaining about not having support for all the plugins.

That said, I do think the NDSP guy in that Japanese video in the Q&A section did make some lofty claims like how no other modeler has the right "feel" or how their modeling is "algorithmically perfect" which is pure nonsense.
 
There’s people that complain about lack of plugin compatibility every time the QC is mentioned and others who complain about time being spent on plugin compatibility. Damned if you do and damned if you don’t.

That's a completely self-inflicted wound because they didn't decide to communicate even somewhat honestly on the scope of work it required until three years after they announced it (via the Sep. 2023 update). Even then, it's taken nine additional months for it to work with just two plugins.

That's kinda crazy when we're talking two suites with just three amps apiece (and Plini Friedmans and Gojira 5153s are already in there) and some different delay and reverb algorithms.
 
That's a completely self-inflicted wound because they didn't decide to communicate even somewhat honestly on the scope of work it required until three years after they announced it (via the Sep. 2023 update). Even then, it's taken nine additional months for it to work with just two plugins.

In their defense, it’s very easy to misinterpret “Shortly after launch.”

That was meant to be taken as “Three years to release two, with a near halt to all native QC development”.

It’s very easy to mix that up.
 
Back
Top