14 different versions of a fragment of the Sonata Kk13 by Domenico Scarlatti.

Could you please provide the MIDI file you used?
thank you

(sorry for my english).

I'm not sure that this could be done intentionally on a piano--to calibrate the pitch of the hammer strike to the lower partials of each note. That would require calibrating the weight of the shanks, voicing each hammer, and much more, to each note. Could it instead be the older tube equipment that is helping to eliminate the crisper sound of the hammer blow, so the notes seem to pop out so cleanly? The tube mics and Neve console?

Vintage D has repedalling. You just have to click the feature under the articulation tab.

Good to know! I wonder why it wasn't enabled by default. Thanks!

Piet De Ridder wrote:

I put the one I like best first.

VintageD used to be my favorite as well, and in this trial I prefer it to the ModelB.
The ModelB recording sounds bad at 0:50, the low notes seem kind of distorted: I wonder if it's a problem with the recording or this is how it sounds (I don't own it).

I recently tried the VintageD again, but its lack of repedaling support makes it almost unusable to me: probably because I'm a bad player and often press the pedal too late. ;-)
Tough, I do it deliberately in at least one song: after a progression with pedal down and lots of resonances, I end with a high note, then quickly release and repress the pedal while I hold that key. I like the effect it creates with pianoteq.

Vintage D has repedalling. You just have to click the feature under the articulation tab. It's one of the most complete script ever created for pianos. O.o"

I put the one I like best first.

VintageD used to be my favorite as well, and in this trial I prefer it to the ModelB.
The ModelB recording sounds bad at 0:50, the low notes seem kind of distorted: I wonder if it's a problem with the recording or this is how it sounds (I don't own it).

I recently tried the VintageD again, but its lack of repedaling support makes it almost unusable to me: probably because I'm a bad player and often press the pedal too late. ;-)
Tough, I do it deliberately in at least one song: after a progression with pedal down and lots of resonances, I end with a high note, then quickly release and repress the pedal while I hold that key. I like the effect it creates with pianoteq.

Jake,
[...]and my original objective was to illustrate that despite its many velocity layers, it has much less dynamic timbral differentiation than, say, the VintageD. [...]

_

Ah, I see and hear, now. Thank-you for the explanation.

I refered about a pre-beta, and not the nusual beta, something beofre the beta testing.

Piet, maybe you should participate somehow in pre-beta. Your ears and focus (find the "epicenter" of the matter) worth a lot.

I have the feeling Modartt could add more variances using tricks, like scripts do, but they prefer reach such things by direct mathematical emulation in orther to the variation be a result of the physical virtual stream of events.

For example (my guess), they could add some relative simple algorythm to add randow variance in some portion of the partials during some instant of the key strike. But if they had being use such tricks since version 1 or 2, they wouldn't had reached the real refinement of today in the core physical modeled itself.  They prefer to create details and variances that comes as result of the the physical interaction of the virtual piano elements. That pushs them to the search of real perfectionism.

You’re absolutely right, this particular fragment only tells us something about just a small handful of aspects of a piano sound — the character and the quality of the sustains being, regrettably, NOT among them — but this little experient was only begun following a discussion about the Hammersmith (which I don’t like) and my original objective was to illustrate that despite its many velocity layers, it has much less dynamic timbral differentiation than, say, the VintageD. (As I think this comparison clearly shows.)

And after I had rendered the piece with the Hammersmith and the VintageD, I began thinking that it wouldn’t perhaps be a complete waste of time to try it out with a few other libraries as well.
So, it's perhaps good to keep in mind that dynamic colour/articulation is the main (and almost only) point of focus here.

A similar test with a more varied piano part that includes sustains, pedal work, etc. would, I agree, tell a more complete story.

_

But, yes, these recordings are wonderful. Thanks again for posting them.

Nice job Piet !

But you forgot Roland V-piano and Roland Supernatural piano.

It must cost quite a bit to have this collection.

But you forgot Roland V-piano and Roland Supernatural piano.

Interesting that even between only samplers we can hear that some are more clear than others, and some have just a little bit of muffed sound. One critic to older versions of pianoteq was about some muffed sound, but the actual version it's more clear than many sampled pianos.