What a few people still complain a bit is the middle range. But it's much better today than a few past versions. It's Always improving after each update

Many people already requested a pianoteq version with options to increase CPU need by using more computation to emulate the sound, creating more realism.
If I understood well Modartt, they say it's not that simple.

Other people, myself included, requested a option for ver high accurate processing, using export áudio function, or in other words, not real time sound, but a slow rendering of a richer sound. 
This would be very usefull to make very high quality and ultra very realistic recordings.

Surely the realism of the high notes in pianoteq can be improved without changing the model?

All Moddart needs to do is release a special version of Pianoteq that requires a minimum 2.7 Ghz Quad core processor.
They will then be able to reduce the size of the time steps they use to approximate the solution of their differential equations; thus getting more accurate solutions.

Of course they should still keep a version that can run on laptops.

Similarly; a Super-Pro version could possibly use GPU computing; known to be great a accelerating scientific computing; this would be a lot of work to code but could be absolutely amazing!

Surprise us!