NDSP Quad Cortex

[mbenigni]

• Making sure you use properly shielded signal paths and grounding techniques to avoid hum.
• Making sure your PCB layout design reduces opportunities for EMI as much as possible.

These are the most likely culprits IMO, and they go hand in hand with the miniaturization necessary to arrive at this form factor. It's a tradeoff. Put 10 lbs of shit in a 5 lb bag, and induction happens. I can definitely see where a person would listen to the hiss (especially with a lot of gain piled on) and reasonably say, "Yeah, nope."

 

[Orvillain]

I can definitely see where a person would listen to the hiss (especially with a lot of gain piled on) and reasonably say, "Yeah, nope."

Especially - especially - when there are products on the market that are about £500-600 cheaper, that don't do it.

Seriously, the Boss GT1000, I ultimately didn't keep it because I thought the amp modelling sucked balls (Neural's amp modelling is very good) and the effects just were not as flexible as I wanted. But hardware wise, there was nothing to fault about it. It was a lot quieter in terms of hiss than the QC, didn't have the same tendency to trigger ground loop issues, and felt solid in a way that you'd just expect a Boss product to feel.

 

[mbenigni]

Especially - especially - when there are products on the market that are about £500-600 cheaper, that don't do it.

Yes, but I'd question whether these alleged equivalents - at any price point - actually exist. If you don't want to sacrifice I/O, routing, form factor, modeling/ capture quality (as in your GT1000 example), etc... you start to run out of options at any price, let alone £500-600 cheaper.

So you look at the strengths and weaknesses as they impact on your "use case", and pick a lane.

 

[laxu]

These are the most likely culprits IMO, and they go hand in hand with the miniaturization necessary to arrive at this form factor. It's a tradeoff. Put 10 lbs of shit in a 5 lb bag, and induction happens. I can definitely see where a person would listen to the hiss (especially with a lot of gain piled on) and reasonably say, "Yeah, nope."

The Hotone Ampero 2 Stomp is even smaller and I have no noise issues with it.

Inspired by this thread I just tried it with my BluGuitar, turned off all noise gates and set the Amp 1 on the Classic channel with gain on 6. No noticeable noise difference in full 4CM compared to going straight into the amp.

Would not be surprised if Hotone engineers have bought a few QCs and picked them apart when engineering their boxes to see what works and what doesn't.

If a box at almost 1/4th the price and less than half the size of the QC can be less noisy, then the compact design of the QC is not the reason but NDSP's design.

That said, I really wish NDSP had released the rumored "Mini Cortex" as something just like the Hotone Ampero 2 Stomp. I do prefer NDSP's amp modeling, capture ecosystem (more captures at once, more captures paid and free available), some of the effects etc.

The Nano Cortex is a big miss when it seems limited and dated compared to units that cost several hundred euros less.

 

[mbenigni]

If a box at almost 1/4th the price and less than half the size of the QC can be less noisy, then the compact design of the QC is not the reason but NDSP's design.

It's not a question of absolute size. It's a question of size vs. contents. If you have more components (e.g. preamplification for I/O) per square inch, there's a higher probability of one part/subsystem inducing noise from another. (I'm not familiar with the Hotone Ampero 2, and I have no particular interest in dying on this hill, but I'm guessing there are some significant/ relevant variables between the two.)

 

[laxu]

It's not a question of absolute size. It's a question of size vs. contents. If you have more components (e.g. preamplification for I/O) per square inch, there's a higher probability of one part/subsystem inducing noise from another. (I'm not familiar with the Hotone Ampero 2, and I have no particular interest in dying on this hill, but I'm guessing there are some significant/ relevant variables between the two.)

They both use Analog Devices DSPs. Hotone does advertise they have separate ESS Sabre AD/DA chips so maybe that helps, I wouldn't know the first thing about evaluating those and how they are implemented will matter too.

While I agree that putting more stuff in cramped quarters will make noise more likely, there's still definitely ways to make that happen without any noise becoming a problem. I mean HX Stomp, Boss GT-1000 etc are as small as the Hotone and also are fine.

 

[mbenigni]

They both use Analog Devices DSPs. Hotone does advertise they have separate ESS Sabre AD/DA chips so maybe that helps, I wouldn't know the first thing about evaluating those and how they are implemented will matter too.

While I agree that putting more stuff in cramped quarters will make noise more likely, there's still definitely ways to make that happen without any noise becoming a problem. I mean HX Stomp, Boss GT-1000 etc are as small as the Hotone and also are fine.

I don't want to spend too much time mansplaining at you (or anyone else.) I know you're a smart, technically-minded guy. That said, noise induction isn't going to occur inside of digital components (e.g. the DSPs they share in common.) The easiest place to look for trouble with the naked eye is the back panel. The 2x XLRs with phantom power are an immediate consideration, followed by the extra outputs, FX loops, etc. I'm not saying the noise floor couldn't have been improved; only that doing so would have likely had an impact on some combination of features, form factor, or cost.

 

[Orvillain]

mansplaining

 

[mbenigni]

 

[ejecta]

I like when my wife splains my man

 

[Orvillain]

I like when my wife splains my man

I like it when my wife sprains her ankle.

 

[ejecta]

I like it when my wife sprains her ankle.

Foot cast fetish?

 

[Orvillain]

Foot cast fetish?

I really like sniffing felt tip pens is all.

 

[Mikael Dez]

I really like sniffing felt tip pens is all.

Read that as “felt tip penis” at first.

 

[Jarick]

Well mine is up for sale on Reverb and should be gone today.

It's a fun unit but ultimately it falls short of the Fractal gear in terms of options and they don't seem to have a development plan. I have also noticed it picks up more noise than other units as well.

Might take the proceeds to buy a Friedman JJ Junior or some other tube amp, or maybe just have less stuff for now.

 

[[Nathan]]

Ooft, is this capture difference observed by anyone else who uses both? You'd assume the algorithms are the same and its something else like input spec difference / i/o spec difference / potential higher noise throwing off the algo. Either way if the NC has consistently "less good" capturing than the QC that's not great. The QC vs Real amp was pretty good here.

 

[Orvillain]

Ooft, is this capture difference observed by anyone else who uses both? You'd assume the algorithms are the same and its something else like input spec difference / i/o spec difference / potential higher noise throwing off the algo. Either way if the NC has consistently "less good" capturing than the QC that's not great. The QC vs Real amp was pretty good here.

It'll probably be down to some difference in the electrical engineering causing the amp signal to appear different to the model.

That.. or the NC runs a lighter model, but I was under the assumption that they weren't doing proper neural network based capturing.

 

[2dor]

It'll probably be down to some difference in the electrical engineering causing the amp signal to appear different to the model.

That.. or the NC runs a lighter model, but I was under the assumption that they weren't doing proper neural network based capturing.

It seems to be something similar (maybe not quite) to the stuff NAM does, at least per this paper:

https://www.analog.com/en/lp/001/customer-success-stories-quad-cortex.html

As to what exactly they're doing, it's anybody's guess.

 

[2dor]

Ooft, is this capture difference observed by anyone else who uses both? You'd assume the algorithms are the same and its something else like input spec difference / i/o spec difference / potential higher noise throwing off the algo. Either way if the NC has consistently "less good" capturing than the QC that's not great. The QC vs Real amp was pretty good here.

I see the guys fiddle around with the Gain knob on the NC when they do the capture process so that itself is a red flag for me..it could just be a case of user error - jacking the input gain up too high on one unit vs the other (which would make sense as the amp sounds as if boosted).

 

[jellodog]

I see the guys fiddle around with the Gain knob on the NC when they do the capture process so that itself is a red flag for me..

Exactly. That gain knob makes all the difference.

My first NC capture didn't sound as good as I'd hoped but I hadn't optimised the gain enough. Second capture had the gain knob set higher and I got different, better results.