Watt dewz FAKEz inztrumentz zound liek?

eph221
Modelled instruments sound FAKE!

 
This is total rubbish. I have owned many real hardware synths.  All the greats and now I have got an amazing collection of these instruments that are so called modelled and to be honest I think that they sound fabulous.  Jaw dropping in fact with many of them.  A controlled A/B test would leave people like eph221 in the dust.  I have had at times both the hardware and the software and got right into this and that was the conclusion I came out with.  In some cases the modelled software has gone above and beyond the original e.g. Korg Wavestation has been improved since the hardware.  You cannot do that with hardware.  And then there are things like the TubeOhm Juno 106 emulation which sounds surpurb by the way.  Plus they give it an extended panel which blasts the original Juno into some amazing territory.  
 
I think it is not smart to try and compare.  Why does it have to be one or the other.  As an experienced synthesist myself I find the best approach is to take them all in and use all of them.  I have all got all types.  Both modelled and sampled and they all have their strengths and weaknesses.  It is when you start combining different synth engines that the real effect starts to come into play.  A texture that is impossible to get from just one method alone!  Why fight it.  You need lots of things to be able to do it super well. 
 
Take IK Syntronik for example. This plugin is based on a sample library yet it sounds sublime and truely beautiful.  There is almost something about it that sets it apart from even some of the best modelled instruments.  I did not like it at first but now I am a fan.  A big fan.
 
 

eph221
Modelled instruments sound FAKE!

Yes, they do. Even when we marvel at how close they get, any piano player could distinguish Pianoteq from a sampled piano in a blind test - when soloed. In a mix, though, a modeled instrument can be indistinguishable from a sampled instrument. Just as sampled instruments can be plenty close enough to a real instrument - in a mix.
 
It comes down to what's known as the "uncanny valley", a term originally applied to human face recognition versus animated faces. No matter how close you get to a match, our brains still register a strong sense of a mismatch, even when we can't specifically pinpoint what's wrong with the fake. Ever notice how no matter how much somebody spent on dentures, you can always spot fake teeth as soon as they speak?