Why even modern VST Synths can't sound like 20 year-old Hardware VA Synths?

[recursive one]

Some really funny details:
1) all examples are drenched in FX, making comparisons nearly pointless
2) first example is Polyevolver. I guess OP doesn't know much about Evolver series, because it wouldn't bring it up for this particular topic.
3) funny thing about second example is that bazillion peoplez complained about Ultranova engine compared to older Supernova engine in Novation VAs, so one might think that HW companies "don't know how to do it". It sounds better then Polyevolver in these particular examples, but then again, I had K-Station for a while and was not impressed by it, but not nearly as unimpressed as with Nord Lead 3 which constantly pulled trick of magically disappearing in the mix. Nord did f**k that one up, so they reverted to older engine on NL4 iirc
4) I remember endless complains back then about "weedy sound" of AN-1X compared to "good" stuff like JP-8000

Don't have much experience with other pure VA plugins that OP mentioned, but Dune sure sounds stunning (if one turns FX off), however it doesn't sound like Virus. Most of these first generation VAs have their own particular sound, but are usually technically inferior to newer stuff. I like MS2000 engine, but oh boy, people were complaining about it's tone endlessly when it came out, and lot of complaints were not baseless.

People complain about nearly everything

Your points are surely vaild, I was also not sure what Polyevolver had to do with this thread as it had a lot of analog components. Also I agree that quite a big part of awesomeness of these demos come from the use of FX, clever patch programming and clever demo preparation (sequencing, automation, especially the Polyevolver one)

But I have a lot of experience with Virus and I know it well myself how it sounds in comparison with the plugins. I never owned the other synths mentioned in the thread (even those I mentioned myself ) so for me they are "awesome by association"

[whyterabbyt]

But almost all modern VST synths sounds "lifeless" and thin..

Please define 'lifeless and thin' in terms which can be measured so objective comparisons can be made.

There is some good analog models (clones) like U-He Re-Pro series or Softube modular. But not any good sounding "Digital" synth...

Please define 'good sounding' in terms which can be measured so objective comparisons can be made.

What's the problem?

The assumption that your subjective opinion is a viable basis for objective comparison.

And please remember that 'I'm much more objective than everyone else and my opinions are facts' is a manifestation of narcissistic recto-cranial occlusion, not objectivity.

[recursive one]

But almost all modern VST synths sounds "lifeless" and thin..

Please define 'lifeless and thin' in terms which can be measured so objective comparisons can be made.

That's an interesting question which I think takes some analysis and measurement skills that would rather be expected from a synth developer than an average synth user.

Some synths do sound warm, alive, fat and punchy by nature, some others require clever programming and external processing to approximate these qualities. Why is it so? I wish I knew.

Might have something to do with non-linearilties in frequency response, stereo placement of the signal, saturation/harmonics introduced along the signal path, envelope shapes, some very subtle movements of pitch/phase/watever, maybe even some "flaws" such as aliasing that can potentially make the top end more "sparkling" or bandlimiting that kills the excessive harsh highs.

I like working with synths where everything of the above has been already embedded into the signal path by the developer, so I could just twist the knobs and enjoy the sound

[whyterabbyt]

But almost all modern VST synths sounds "lifeless" and thin..

Please define 'lifeless and thin' in terms which can be measured so objective comparisons can be made.

That's an interesting question which I think takes some analysis and measurement skills that would rather be expected from a synth developer than an average synth user.

But I think its a fair request in the context of an average synth user stating that synth developers en masse have failed to meet a subjective standard which only exists in the head of that specific average synth user.

[urosh]

But I have a lot of experience with Virus and I know it well myself how it sounds in comparison with the plugins.

My point mostly comes down to "this is not topic for technical discussion in this particular subforum", because (based on examples I brought up), it is not something that HW devs do that SW devs don't know how to do or choose not to do. There might be technical explanation of why someone prefers one synth over another, but again, it's first and foremast aesthetic choice imho, not rooted in general technical inferiority of plugins. Some people prefer Virus over Dune, just like some people prefer Jupiter 8 over Prophet 5, not because Dune is inferior.

[recursive one]

But almost all modern VST synths sounds "lifeless" and thin..

Please define 'lifeless and thin' in terms which can be measured so objective comparisons can be made.

That's an interesting question which I think takes some analysis and measurement skills that would rather be expected from a synth developer than an average synth user.

But I think its a fair request in the context of an average synth user stating that synth developers en masse have failed to meet a subjective standard which only exists in the head of that specific average synth user.

As it was posted in the developmment section I assume it was meant to be a kind of technical question rather than "you, developers, don't know sh*t and can't make a good sounding synth".

I agree it was too broad, with digital and hybrid synths being mixed up and giving examples of complex and processed sounds. Would have been better with more "dry" examples and limited to specific synth(s). I once thought about making a series of A/B examples of Virus and similar sounds made in Dune, Spire etc

Probably the actual question would be like that:

Do the hardware VA synths that were popular in the late 90s to mid 00s (Virus, JP8000, An1x, first Nordleads) actually have any essential techncial differences from modern software VAs, what are these differences and how do they impact the sound?

[whyterabbyt]

Probably the actual question would be like that:

Do the hardware VA synths that were popular in the late 90s to mid 00s (Virus, JP8000, An1x, first Nordleads) actually have any essential techncial differences from modern software VAs, what are these differences and how do they impact the sound?

Ah, well, that's easy. Technically, they use cruder algorithms, with more compromises. However they're in a special magical box which makes them sound better, even against a 100%-identical port of the same algorithm to a software-only device. Because hardware is magic.

[recursive one]

a 100%-identical port of the same algorithm to a software-only device.

Like what?

I can only think of K/V-station

[DJ Warmonger]

Bones already mentioned hardware processing factor. I've heard an opinion that a lot of character of hardware gear is in digital-to-analog converter, which is obviously non-existent in virtual synths and, either way, tricky to emulate as far as I know.

[whyterabbyt]

a 100%-identical port of the same algorithm to a software-only device.

Like what?

I can only think of K/V-station

The Lexicon Native algorithms were 1-for-1. Korg M1 and Wavestation too, I believe. Mutable Instruments ports in Softube/VCV. Nord Modular G2 demo was a virtual machine that actually ran the original code in an emulation. Hartmann Neuron. Waldorf Attack hardware was a copy of the software. etc
Anyways, you could crosscompile the audio engine of just about any softsynths onto an ARM, stick it in a 1U box with 1/4" jack outputs, and some folk will tell you it sounds better than software. That's just how it is. Hardware is magic. Woooooo.

[whyterabbyt]

Bones already mentioned hardware processing factor. I've heard an opinion that a lot of character of hardware gear is in digital-to-analog converter, which is obviously non-existent in virtual synths and, either way, tricky to emulate as far as I know.

Should be resolved by running your softsynth through a digital-to-analog convertor before listening to it, then.

[recursive one]

Hartmann Neuron.

Yes, this one too. I actually had an opportunity to brifely play with the software version years ago. I didn't know is also existed in hardware back then, but from my memory it soudned different to the plugins I was familiar with, like Z3ta, Albino, Vanguard, Sytrus. I was like "wow, that's not the typical software sound I'm used to".

But as I said it was years ago, not sure what I would think of it today.

Regarding the converter, it must be a valid and obvious guess but I use my Virus over USB so the converter in this case is put out of the equation.

There are some SW synths that I loosey describe to myself as having "hardware sound" based on some hardly meausrable and identifiable qualities, such as Massive X, Repro 1/5, Bazille and suprirsingly Sylenth (yes, Sylenth!).

[ShawnG]

a 100%-identical port of the same algorithm to a software-only device.

Like what?

I can only think of K/V-station

The Lexicon Native algorithms were 1-for-1. Korg M1 and Wavestation too, I believe. Mutable Instruments ports in Softube/VCV. Nord Modular G2 demo was a virtual machine that actually ran the original code in an emulation. Hartmann Neuron. Waldorf Attack hardware was a copy of the software. etc
Anyways, you could crosscompile the audio engine of just about any softsynths onto an ARM, stick it in a 1U box with 1/4" jack outputs, and some folk will tell you it sounds better than software. That's just how it is. Hardware is magic. Woooooo.

You might have to add the factor of having to justify dropping a thousand bills on said hardware, but yeah...

[AnX]

...
What's the problem?

your ears

[whyterabbyt]

You might have to add the factor of having to justify dropping a thousand bills on said hardware, but yeah...

True. Don't get me wrong, some hardware sounds better than some software. And vice versa.

Its not something intrinsic to the 'construction,' though.