What I want in Pianoteq 4!!

You can just turn off Pianoteq's reverb and use any convolution you have in mind.

I would rather want a room modeler built in, like QuikQuak RaySpace.

Just two things:  1) A falloff range on the microphones represented graphically in the display by 2 adjustable concentric circles around a microphone -the center circle represents the beginning of the falloff and the outer circle represents 100% falloff.  2) The ability to isolate the sympathetic sounds and either send them out a dedicated output or simply turn off/down the principle struck notes. 3)  Anything else that helps promote the effect of the body of the instrument, even if that has to include built in post effects such as random multi-tap delays, microphone modeling or a flux capacitor .  (Ok... that's 3 things)

As an answer to question (main title of thread) I'd say just improve the basic sound as much as possible to make it competitive to other digital alternatives which are getting better and better! That should still be the main goal. However, I don't think it's an easy task.

There's no such need to have more instruments like sousaphone or clarinet or steel drums. I think PTQ shouldn't be all around instrument library. All the pianos (Grands & uprights) and its ancestors are of course nice part of PTQ family. Also vintage e-pianos (Rhodes, Wurly, Clav) are welcomed. Improving basic sound of especially Rhodes might be wise too (there aren't that many good competitors for modeled Rhodes but many many good sampled alternatives). PTQ E-pianos are quite nice but for me there's no real use for those.

P.S: I don't know if it's even possible if they make the piano model better (and perhaps more complex) and same time make it more CPU effective. Anyway I think there are some customers that might want to have PTQ but their computers cannot run it well enough.

Exactly.

I'd love to see an aliquot stringing system so PTQ could faithfully model a Bluthner piano. The Bluthner Digital Model One sampled piano does not do this instrument any justice whatsoever. This opinion is based purely on listening and comparing recordings of BDMO with a real instrument. I've never tried BDMO but would not buy it (or any other sampled instrument) based on what I've heard. 

A Bluthner with aliquot stringing has a distinct mellow tone which sets it apart from the brighter tones of a Bosendorfer, Fazioli or Steinway, all of which PTQ can already faithfully model in both sound and performance with various tweaks of the K1, M3 and C3 models.

Some more early 20th century grand pianos would be great. I love the 1922 Erard which is so full of character (straight out of the box with no tweaks) and perfect for Beethoven's piano sonatas, especially the moonlight.

I too would like to see some characterful uprights  for when I'm feeling a bit of an animal and want to play some rock and roll.

Err, in a word; "No".
and "Yes", it is something else (entirely).

There is quite a lot of confusion because the terms are used in the descriptions of each other's implementations .....approximately that (-:

I can only suggest you "look 'em up", sorry but my descriptions can't match what you can probably find in a 2 second search.

Duplex Scale was developed by Steinway and Helmholz in the 19th Century. The length of the string between the bridge and hitch pin can be adjusted to a precise length such that it resonates harmonically with the main part of the string adding to the overall colour of the sound. Steinway and Fazioli use this to good effect but other piano manufacturers (notably Yamaha) deliberately dampen this effect by weaving a felt ribbon between the strings at this point. Duplex scaling is not always a desirable feature. It's one of those things which makes pianos of different manufacturers differ from one another. But in PTQ you have the choice, if you don't like it you can turn it off or tune it to your own preference.

Aliquot stringing is an extra string which is not struck by the hammer but shares the same damper as the other three strings. The extra string vibrates sympathetically with the other three strings. I don't know of any manufacturer other than Bluthner who employ this technique. It certainly gives a Bluthner a distinctive sound which is not to everybody's taste but I love it, especially for quiet pieces such as Chopin's Berceuse or Debussy's Clair de Lune. Ive never played a Bluthner but I've been told that the action is very heavy, possibly because of the extra weight of the damper having to cover four instead of three strings. Again, everything is a matter of taste and professional pianists will all have their own preferences. I prefer a heavy action because its harder to hit a wrong note (but I'm not a professional). Cross resonance between strings is not a feature which sampled pianos can adequately reproduce primarily because of polyphony dropout. The sounds of a modelled piano like PTQ evolve in real time (using real, not quantised numbers)  so at every sampling instant (every 22 microseconds for CD quality, faster if you own PTQ Pro) every note sounded depends on what other notes are played, the position of all the dampers, how loud the notes are sounded etc, all calculated in real time using all available parameters. The extra string could be added as a parameter to PTQ and its effect adjusted accordingly. An extra parameter may take up a few megabytes of RAM in PTQ but extra sampled recordings for one parameter change could take up several tens of Gigabytes of hard drive space in a sampled instrument.

I took some convincing that modelling was the way to go. Last year I'd ordered the Garritan Steinway and received a reply that Garritan would not be shipping until further notice, so I cancelled the order. It was a colleague at work who showed me PTQ so I downloaded the demo at home and within a couple of hours of trying it bought a licence and have never looked back.

There are a lot of variances.

Talking about vary a bit...    WHat about modartt add a option to play the cimbalom with spoons and forks ?

http://www.youtube.com/watch?v=LtWt6f74mnc&NR=1

Háa háaa...   

Well, perhaps some things was found by just accident, or trying and trying add trying...

I was thinking if in some years modelled technology could be used to help to design  andcreate and aprimoration of real pianos.

Think with me...  instead of try change many things, like change the wood, build a new shape for harp and soundboard, or make small changes in the string, which take time, money and effort for each single try, they would just design a new piece, new shape, on computer and emulate how it would sound.
By this approach they could try hundred changes more than by traditional manufacture of new experimental pieces.

Good idea. It would sound something like Imperfect Samples thingie, right?

About the key and action sounds--there's some information about that in the dissertation I linked to. The focus there is on reducing the contribution of the keys to the sound, but it may have some information that could help with modelling.

Seems as though there are several sound sources there, though--each part of the action, each key, the bed itself, and probably much more. Wouldn't be easy.

If we don't stop coming up with these off the wall suggestions, we'll need 48 cores. 

G

I'd like pre-emptive mistake prevention.