Why does my R3 sound better than comparable VSTs?

tpantano · Post by **tpantano** » Thu Feb 23, 2012 6:45 pm

I'm quite confused as to why I'm finding my R3, a VA, is sounding better to me than my Subtractive VSTs. It seems louder for less mix space, it seems to have more character, it seems to have juicier filters- and I can't figure out why! I'm very confused, because at the heart of it- aren't VSTs and VAs both software, both bits and bytes?

It's not more features, as a plain old saw-filter-amp.env chain on my R3 sounds (subjectively) better to me than, say, the same on Logic's ES1/ES2.

A friend (who's fairly new to synths but not to computers) said that it could be the fact that the R3 has dedicated music processors, but I dismissed this by saying that a VST is dedicated music software and a four core 2.4 ghz CPU running on an OS with very-nicely integrated audio processing can run more functions than a VA like my R3 can.

Perhaps it's the programming? Something to do with the MMT engine?

Or does the hardware have some effect as well? I dismiss it as just being bits and bytes equivalent to my laptop's computer, but maybe the soundcard of the R3 has some sort of built-in characteristics to it?

I'm really curious what gives my R3 that character I love. It came to question when I turned on my R3 after months of using nothing but VSTs and being wowed by the patches on it. With a synth like a Slim Phatty, it's obvious what gives it its character- the analogue components- but with a digital synth like the R3, it's not-so-obvious how it has such an... well, awesome character to it.

Thank you for your insight!
~Tristan

MonoxideC · Post by **MonoxideC** » Fri Feb 24, 2012 2:14 pm

I would suspect that is has to do with the R&D that goes into making the product, as well as the experience that Korg has with making synthesizers in general.

When you're modeling circuits virtually, it helps to have the kind of engineering resources that I'm sure Korg has, as well as all of the documentation and proprietary trade secrets that they have produced over the decades designing a vast array of synthesizers.

When it comes to a VSTi - especially something bundled with other software, chances are that it wasn't built with the same kind of love, knowledge, expertise, and experience (or budget!) as something like the Radias/R3 was.

axxim · Post by **axxim** » Fri Feb 24, 2012 4:30 pm

IMHO it has to do with all of that and more:

Developers
Usually VSTi's are developed by single or few programmers. They may be good programmers, audio and/or electronic experts,mathematicians, physicians etc. but usually they aren't all that at the same time. In each of this branches you also have sub-branches so an audio expert could be good at designing amplifier stages, filters, effects but also not so good in converting this to digital models and viceversa.
By the other side, I assume that KORG and competitors are large companies who have many more specialists on all this different branches and therefore you get a more specialized know how on each of these sectors.

Hardware
Of course you can't compare a PC HW with a dedicated (digital) music HW. Its like comparing a Land Rover with a Ferrari. You can do nearly the same things with both, but how they suit for each application is different.
A PC usually has a standard soundcard and hence the majority of VSTis are done for that configuration. Unless you buy a high end audio interface and unless you have special VSTis that support them, they will have their limitations.
Music instruments usually are built around DSP's which have better suited functions for signal processing than a CPU, even if they work at lower frequencies.
The structure of the hardware may affect also the programming techniques which influences again the programmers factor from above.

Another cause could be that VSTs are only triggered via MIDI. I think the R3 gets its kb-, pots- and sliders data internally (and therfore faster), interpretes and uses them for its sound engine and at the very last converts this to MIDI out signals. Not sure if this is so but MIDI data is a little to coarse for the actual technology.

This is why I think and debate, that most VSTs are very overpriced when compared to the price of what you get with an R3 or similar.

Timo · Post by **Timo** » Sat Feb 25, 2012 2:38 pm

Volume has a huge impact on how people consistently "prefer" one sound over another (in blind tests), regardless of the media used to create them.

As does panning law (how the right and left channels are balanced when creating the phantom 'centre' channel) which defines the 'mixing space'.

Re-Member · Post by **Re-Member** » Fri Mar 02, 2012 5:09 pm

I've found that it varies with each VST I try out. Arturia makes some really great sounding replications of classic synth, even nailing that drifty and inconsistent VCO sound, but they eat up a good chunk of processing power. To my ears, they sound much better than the handful of VA hardware synths I've played on. Likewise, other forums have been talking a lot about the u-he Diva, another VST that better replicates the sound of VCOs than most VA hardware.

Right now I'm currently using Korg's Legacy Collection software, mostly the Polysix and MS-20. Many of the patches I create using them, I reprogram onto to the R3 and it really doesn't seem sound any better or worse depending on which speakers or amp I'm using.

rtofvnt · Post by **rtofvnt** » Sun Mar 25, 2012 5:40 pm

Simple answer is: instruments has soul.

Eat-Static · Post by **Eat-Static** » Sun Mar 25, 2012 7:29 pm

instruments has soul.

Is that why they go to Silicon Heaven ?

Eat Static...

richardw · Post by **richardw** » Sat Mar 31, 2012 8:47 pm

Another possible reason probably involves the way the CPU is utilized in each system. Even though the CPU in a typical Mac or PC is fast and powerful relative to the CPU in a hardware VA, the hardware VA uses a processor matched to dedicated software to produce sound. This system uses ALL of its resources in some way to produce sound, and nothing else. Compare that to music software in your desktop or laptop computer. In these systems, there are dozens of processes running in the background at any given time managing the operation of the entire system. The developers of the music software have to engineer their product to only work with a designated small amount of available resources at any given moment.

This probably gives an edge to the hardware VA in many general applications--sound quality, reliability, etc. Just some thoughts.

MrDuke · Post by **MrDuke** » Tue May 08, 2012 10:44 am

Wouldn't the obvious answer lie in the A/D converters etc.?

Because I've noticed the same exact thing.

Even though a dedicated piece of hardware is always (?) "better", running out of CPU can't be the issue as richardw suggested. Nor is it the lack of a proper "sound card" as axxim suggested - I'm using professional audio interfaces and mixers.

I actually did an album recently, where I used the R3 as a synth bass replacement - and it sounded amazing. Totally convincing, almost "analog" and fat (didn't use a preset).

I haven't been able to get same results with for instance the Legacy Collection software. The difference is subtle, but audible - just a slightly different sonic character.

tpantano · Post by **tpantano** » Tue May 08, 2012 6:07 pm

MrDuke wrote:Wouldn't the obvious answer lie in the A/D converters etc.?

Because I've noticed the same exact thing.

Even though a dedicated piece of hardware is always (?) "better", running out of CPU can't be the issue as richardw suggested. Nor is it the lack of a proper "sound card" as axxim suggested - I'm using professional audio interfaces and mixers.

I actually did an album recently, where I used the R3 as a synth bass replacement - and it sounded amazing. Totally convincing, almost "analog" and fat (didn't use a preset).

I haven't been able to get same results with for instance the Legacy Collection software. The difference is subtle, but audible - just a slightly different sonic character.

I'm not sure how this makes an sense, as the R3's A/D goes through the same D/A as my softsynths for playing out my speakers >.>

Re-Member · Post by **Re-Member** » Wed May 09, 2012 7:40 pm

Many hardware synths are limited by their processing power, which is why aliasing in the filters exists.

Also, the A/D converters with the R3 isn't a solid explaination for why it would sound better. With a computer, there's nearly an infinite amount of soundcards to choose from. I doubt whatever the R3 uses is at the top of this list considering it's a budget instrument. And if it was the holy grail of A/D conversion, it has an Audio In section, so technically you could route any audio into it if that was the case.

The ultimate difference in the end is the program itself. The Korg Legacy collection is actually a few years older programming wise and is modeled after specific older instruments. The R3, however, wasn't designed to replicate any particular instrument. The name itself states that it's "Multi-Modeling Technology" which is a mixture of things put together and becomes it's own thing.

In the end though, I still feel that many instruments with VST filters tend to sound smoother and hit a better frequency range than most VA synths without aliasing, but it various from program to program, just like it varies from synth to synth. My SH-201, for example, sounds horrid at times with its aliasing, but I imagine that if you heard the original software it was designed with, it would have sounded a lot smoother running on a computer.