". . . He had planned three concerts where in two of them, he would destroy two grand pianos, and in one, he would burn his formal concert dress. He said that he would instead tour with an electronic keyboard around France to give informal concerts." from: http://en.wikipedia.org/wiki/Fran%C3%A7...h%C3%A2ble

Seems like this guy is in favor of the 21st century approach!

Don't howl me down, but I am convinced you can build transistor amplifiers that sound exactly the same.

Great comment, but I have to insist, I'm sorry (don't want to howl you down).
By now there are no simulations that sound so much like a tube amplification that I wouldn't hear it (even in a mix).
That's due to the irregularity of the electron-cloud, which is a phenomenon that contains the "Many body problem" (part of quantum-physics) which is pretty hard to calculate and impossible for todays home-computers to do this in realtime regards to the fact, that you could define billions of states from which the tubes behaviour is a little different.

But in the sooner or later future, there will be quit perfect models.
So in general, I think you're right.

1. The power undependance mentioned before.

2. On/Off Velocity detection isn't the full truth of it. You'll need way more sophisticated controllers, with a precise determination of the position of each key.

3. 3ms of latency just stay 3ms of latency. No matter how fast cpu's become, there will be a noticeable delay (acoustic stimuli with down to 2ms time between can be recognized as different).

4. You can't mess with the sound of an acoustic piano, standing somewhere in the room before you. I often accompany singers, and it's so damn hard to get the right mix for voice and piano, when the piano comes out of a speaker. It's like the sounds won't mix up in the right way. They don't fit into each other, if you can imagine. Maybe with some real extravagant amplifier and speaker system and some future pianoteq version, it will be possible, to get close to it. But for the price, this system would cost, I'd prefer the Steinway or Bösendorfer.

Don't get me wrong. Pianoteq is a marvellous thing, but you should see it as what it is: A substitute when there's no real thing at hand and a great instrument to create sounds.

:-)

In some pieces there are segments where the righ hand and the left hand play like were giving responses on to each other, creating a music clime.  What about add more of such "dialog" by chaging few aspects?
  The changes would follow the logic of natural pianos, chaging from one aspect of one real piano to the aspect of another real piano.

   I don't have a piece to give example.  Perhaps Toccata and Fugue from Bach, addapted for for piano. That have very nice moments of right hand "speaking" to left hand.
     Let's say a moment of "dialog" that use bass notes progressively going to trebble, to demonstrade a feeling of smoothing, of getting soft, calm, relaxing. This could also had a progressive softness of the hammers too.
     It's just a imaginary example. I'm sure there is many possibilities for many adjustable parameter.

    I bet if Bach had such technology he would take the potentions to a special composition.   :-)

It's artistically important I think, and it's a noble act.  But pieces like that are usually short lived because unless that specific feature becomes standardized, the piece cannot be played back properly, and so it becomes history.  A recording may remain, but that's pretty much it.. 

It's like encoding.  Let's say you encode audio in the newest, awesomest lossless audio codec.  If that codec got lost, and was forgotten in 2 years, your audio will never ever be played back again..  which sort of sucks.

See, what I don't understand is that some people (not just you Beto-Music) seem to expect strange things from a physmod synth...  variable timbre, out of this world controllability to make surreal sounds...  to me, it all sounds like what you'd get, or what we've been getting from just a general synthesizer for that past 40 years.  A general synthesizer (sample based, hybrid, subtractive, additive, fm, pm, granular, spectral, etc..) is much better at creating new sounds like that.  I mean, a physical model is not built for things like that in the first place.

I would much rather have a physmod piano that did what it's supposed to do... model whatever instrument it's trying to mimic, as best as technology allows.  If that can't be attained, then all the bells and whistles really don't mean anything.

But I imagine a close future like that:

Composers would create specific great musics witha  touch of classic but made to be played in special modelled digital pianos.

There would be extra pedals to vary the piano lenght, hammer hardness, metalic sound, string material, sympathetic resonce, sound duration, unison etc. While playing these characteristics would be changed in a way to increase the musical feeling.

Well, what you thinl about that ?

If DP's haven't already replaced (or started to replace) the cheap acoustic pianos, then why are there so many DP's in existence, compared to the year 1990 when I bought my first Technics?

I don't think whether a piece of equipment is popular or not is the way to measure whether it is successful or not.  It's an over generalization..  Many people own synths with a physmod violin (string), guitar (plucked string), reed, and brass...  maybe more than people who own the physical instrument..  still, these things are not "replacements"..  actually, physmod for those instruments have been around for a while, and are quite refined..  regardless, they are almost never used for performance, and aren't used for recording either.  It seems illogical, but that's just how it is.  Of course, the day "may" come, but that is in a scientific sense, giving even %1 the benefit of doubt.  Actually, they're being replaced by sample libraries in terms of production and recording, but not for performance.

But, as I think we agree that for bad spinets and consoles, even a mid range DP is much convenient and sounds better.  I think it's much better to practice on a so-so action and so-so sound, than be hindered by irregular action and bad sound.  A bad spinet or console isn't applicable for performance, so I won't even go there.

I agree with you 100% about recording.  This is the same story with drums and orchestra.  For rapid production, ease of use, cost efficiency, and quick deploy are key.  In that sense, physmods (if it's done well) can easily replace an acoustic.  Even now, acoustics are mostly replaced by sample libraries because of convenience.  So, there definitely is a different between performance and recording, and for recording, physmod will be the holy grail.  Maybe not for classic, or solo piano projects, but for general mixes, definitely.

There's a big difference between when a piano is a means to an end (production tool) or when it's an end in itself (solo performance, etc).

It seemed to be not correct to create such a sequence: replacing a clavichord by a piano is comparable to replacing a piano by its electronic emulation. You see, the historical period of a clavichord was the very beginning of “classical music”, when means of expression were at their early stage. As far as new tracks of virgin soil were been brought under cultivation composers and performers demanded more and more sophisticated instruments. That is why a clavichord wasn’t REPLACED by piano. It’s still there, in its late Medieval, Renaissance, and Baroque eras. But it’s alive today — because it seems to be incorrect to perform on contemporary piano pieces by Bach or J. Kuhnau composed for a clavichord.

But what about a grand piano? Do contemporary pianists demand a more sophisticated instrument? The answer is: no, they don’t. Look at philharmonic placards — Mozart, Chopin, Schumann, Skryabin, Rakhmaninov… — this music is top rated. A piano was perfect enough for that time. It hasn’t changed dramatically, only a slow improvement took place (by the way, mentioned here and there in posts S. Erard is one of genius inventors…). Yes, “recording a DP or one using physmod is far simpler and cheaper, and the recorded midi file can be manipulated far easier than the recorded wave file from an acoustic” (Glenn NK). But I’m afraid it would be very difficult in the nearest future to make a professional classical pianist sit down at a physmod controller, even with perfect mechanics, on a philharmonic stage. I’m sure it’s not due to an ordinary snobbery.

(However such a piano performer could agree to play physmod piano, but under condition of accompaniment by Quantum Leap East West Symphonic Orchestra.)

The technology of an acoustic piano, no matter how poorly built, is still very complex.

If DP's haven't already replaced (or started to replace) the cheap acoustic pianos, then why are there so many DP's in existence, compared to the year 1990 when I bought my first Technics?

I won't suggest that physmod pianos will replace acoustic pianos suddenly or even in many years, but the change (at least partial) seems as inexorable and inevitable as the death of the clavichord.  Well the clav didn't quite die off either - there are still many of them in our living rooms - aren't there?  LOL

The biggest drawback I've encountered with an acoustic (I owned a Yamaha G2 for about 27 years), is recording one.  The requisite equipment is cost prohibitive.  Recording a DP or one using physmod is far simpler and cheaper, and the recorded midi file can be manipulated far easier than the recorded wave file from an acoustic.  Unless you add Yamaha's costly diskclavier system.