Real Analog vs. Virtual Analog

It’s probably safe to say most of us here agree that synths with a real analog signal path still tend to sound nicer than their virtual models, no matter if these are done as software plugins or as hardware units using dedicated DSPs.

Even though I can think of a few reasons why this is the case, I’m wondering if anyone ever found a nice, somewhat scientifically sound, article summarizing why it’s so hard to simulate these kind of circuits properly.

I’m not so sure I agree with your first paragraph. I think my Kurzweil K2600XS sounds amazing. I also think the Nord G2X I had was impressive too.

1 Like

@piscione In that case I’m wondering why in 2013 I can’t get a software plugin that sounds near as good as a dedicated digital synth from the late 90s.

1 Like

…or, is this the App Store Problem? Everybody expects software to be cheap, so nobody wants to invest enough effort to create really good software synths since the ROI just isn’t there?

I am no longer sure about your question… Real analog vs virtual analog ; or real digital vs virtual digital?

1 Like

In general its the swiss army knife problem: such a Knife is a great tool in general but if you look at each single part its rather a particular good knife, screwdriver, scissors - even the beer cap open thingie is not the best.

From my limited experience a dedicated device always does better than a multipurpose one. Now compare a MicroWave 2 with your Windows PC…

Just to be clear; the question is about the sounds most people prefer to use dedicated analog gear for.

@fcd72 Once it’s all digital, it’s essentially just a computer optimized for a specific task. I certainly believe the programmers and designers working on the MicroWave 2 did a better task than most people writing softsynth plugins today, but there’s no technical reason why you couldn’t replicate the whole thing as one.

pichenettes Sorry for the confusion, my understanding frompiscione’s comment was that he considered the K2600XS and the Nord G2X to both provide excellent simulations of classic analog synths. Therefore, I was wondering why in 2013 we don’t have softsynth plugins who can pull off the same thing as these dedicated boxes from 1999.

There’s so much pontificating about this issue it’s difficult to find anything truly scientifically accurate… I amassed a huge vinyl collection during my teens and loved it but now prefer to listen to digital. I have owned many analog synths but mostly work with digital and soft-synths - each has their own strengths.
You can pick holes in any argument, for or against any system.
Most peoples’ ears aren’t good enough to discern the difference between the real thing and a digital repro and when it’s in the mix it’s largely irrelevant too.
There… I’ve stated my case for sitting on the fence! :-))

I always wondered what the difference between a hardware analogue-modeling Synth or a software doing it could be? After all, they are just scripts ran through DSPs/CPUs? Must be the dedicated hardware made to suit this just one purpose, I guess …

…there you got it :wink:

You could replicate it - that means you could build a dedicates hardware and write a dedicated software - but you cant simulate/emulate it.
Your Windoof System has a myriad layers of Software between the Plugin and the D/A and all sorts of bad things will happen, whereas in a Hardware Synth its always the same Timer Interrupt, reading from your Sample Buffer pushing to the D/A. Everytime in xactly the same amount of Cycles whis means no accidential phaseshifting. You have total control over the Hardware so nothing unexpected can happen.

Can you guarantee no other Interrupt comes inbetween on your non real time optimized general purpose OS Monster?

Guess thats what I thought of …

Oh, and im aware if you make the D/A Buffer on cour dedicated D/A Hardware big enough this wont happen - but then welcome to the wonderland of Latency…

@fcd72 Latency and lack of hard realtime could be an issue during performance, but not when you ask your DAW to go render the final mix. I don’t believe this is the critical issue.

As far as i know was the last Software System that cared for phase accuracy the Synclavier, well known for its greater than life sound… nobody says something like this about Logic or Cubase…

I think one could paraphrase Duke Ellington’s famous quote about music like this: there’s two kinds of synths, good synths and the other kind.

Let me simplify the question; why can’t I get a softsynth plugin that manages to replicate the sound of a SH-101 exactly? Is there a technical reason why this just can’t be done (yet), or do you feel this is because the market for this kind of thing isn’t big enough or willing to pay sufficiently per unit to justify the investment?

Because the SH-101s sound is determined by its filter thats analog. Analog means a lot of nonlinear things going on (you dont assume anything analog being linear in any kind, dont you???) that cant be simulated digitally to a degree that tricks your Brain. Simply because we dont know how.

A good example for an excelent digital synth is the MicroWave and Q Series, these Digital filter dont try to “emulate” a well known Filter, so you have no comparison and can enjoy the filter all by itself, rather than having to argue if it sounds like XYZ or not.

Thermal drift, signal jitter and phase-shifting are not easy to simulate. Why are the guys who manufacture nowadays violins not capable to do better than Stradivari did? Technics and materials should not be the problem…
In the end all comes down to smoke and magic.