I see scope from a different side. I try to explain:
(1) its more or less one big modular system...
so the whole system has roughly spoken "its sound", and to clear things up, I think soundwise its the among the best I`ve heard across all digital stuff I`ve put my hands on.
That said, yes I think if you patch a modular synth for example in mod3 shell it easily can compete with commercial vst stuff emulations (in most cases these are no real emulations at all!)
And in my opinion it even sounds better. So for me its much bang for the bug.
(2) you have to divide into 2-3 sections of synths:
- first of all creamwares newest synth creations like minimax, Pro-12 and Prodyssey that have special designed atoms to model the sound of the real units. Especially filter and oscillator atoms.
- second all the rest: u-know, prisma, blue, and modular3 all share the same or similar atoms and so they have a similar overall sound.
- the John Bowen stuff is somewhat a special case as for example the solaris plug-in is able to load various of cw high class atoms, like minimax filters, cem filters, different oscillator models, and so on.
there really isn`t much missing.
Even the basic oscillators are almost aliasing free, because they are bandlimited and thats just my small critics point:
the bandlimitation (especially on the saw oscillator) is some sort of compromise: it reduces aliasing at the upper range octaves at the cost of high frequency content in the lower octaves.
I would like to see SC modelling some kind of high quality saw oscillator, that
->is aliasing free in the upper range octaves but
->don`t suffer from damping side effects on high frequencies in the lower octaves range.
this would result in a bit fresher sounding bass sounds.
One of my current projects at the moment is such a modular saw oscillator and I try to implement a technique of interpolation to solve this little issue... But it seems difficult because scope is a realtime environment.
At least I`ve already managed to build a non-bandlimited saw oscillator from math atoms, so thats a good starting point already.
another workaround could be to load a non-bandlimited saw osc and a bandlimited one at once, sync them, so that they are phase corrected and have same tune and blend over to the non-bandlimiting osc in the lower octaves and blend over to the bandlimited in the higher octaves domain, which will some sort of full spectrum aliasing free oscillator. This workaround for sure would consume twice at much power as one basic saw oscillator.
Did any of you have a chance to make A/B tests and conclude how they stand the comparison?
for sure I did A/B many VSTi stuff with the scope synths and the modular.
And in almost any case the scope version sounded better.
Even if I patched some synth layouts of commercial VSTis in the mod3 shell I could reach at least equal but in most cases much better results.
To bring it to the point: Since I have my scope cards for ~ 1year, I almost don`t use any VSTi at all. I even go that far that I don`t install them anymore, so they don`t blow up my DAW sequencers loading time.
A little down side is that CW didn`t build much exclusive or modern stuff. Most of the synth stuff is based on Saw/Sin/Square/Pulse/Tri Oscillators.
The only special stuff is the Waldorf oscillator that can step through the oscillator tables like the original waldorf synths and another wavetable oscillator.
So Scope is more a classical sounding environment.
There is where I really wish to see some little innovations from SC for Scope 5 or just i the future.
so i cant say if they sound good. but my opinion on creamware stuff is that they often are more musical then much of the software i have heard. 
