1. A slight unison detuning, making the sound a bit richer (like the well known "chorus" effect)

2. The overall tuning giving it an extra bit of richness - i.e - the way it has been stretch tuned. But I suppose the stretch tuning if done properly is supposed to sound like that, and maybe my piano tuner didn't see any reason to comment on it.

I think I'll just sit back and enjoy it.

Greg.

Greg.

And then there's the resonance/body of the sound.

Greg.

Gilles, it' not pinching the nose I refer. I refer to allow the nose participate more in the sound we produce, like more air coming from it, like trying to use the nose to imitate a resonance.
    We are accustomed to the expression "nasal voice", like some secretary, but this nasal voice, from the popular expression, is cause some people do not use the nose while talking, getting like pinching the nose.

Oh, I see what you mean now. I sing in a choir and after some warmup exercises to open up the head resonators, I always feel air (and so, sound) coming out of my nose. Of course this is needed to alleviate the nasal sound (that we get when having a cold for instance), not encourage it.

Sorry for my misunderstanding.

Greg.

(Is it the harp sound? Process of elimination, here...)

My suggestion for pianoteq implemantation is to add new ressonance algorithms to the small pieces on piano, and "modelled noises" able to interact creating resonances.
  Sure it would take a lot CPU power, but it could have a option to disable, returning to sampled noises as is now.
Consumers with high velocity dual processor would have no problem to use this new functions.

  Sometimes I imagine pianoteq could do anything if ours computers had enough power.   Perhaps Modartt work, about create a complete natural sound to run on any mediun computer, it's quite like try  to put elephants in a volkswagen.

They do seem to detune strongly in the recording. Possibly in a way that differs from how Pianoteq handles the detuning.  In a real piano, if I recall correctly, two of the unison strings in each set of three are tuned to pitch and one is detuned slightly. I imagine Pianoteq assumes this arrangement. But over time on a real piano, one of the two "in tune" strings" can and does go off-pitch, too, causing a bit more twang. That may be what we hear in the JLL recording. Maybe in some future edition, it would be nice to be able to control this (technically bad) third string detuning?

But I'm writing from memory--there may be other "correct" ways of tuning the unisons when there are three strings.

  My Roland F-90 sampled piano have not that, and I don't think I never heard something similar to that in most piano library, or not that well in any one.

  The example I gave was to try yourself imitate the sound of a note, but not speaking the notes (Do, Re Mi...)

Gilles, it' not pinching the nose I refer. I refer to allow the nose participate more in the sound we produce, like more air coming from it, like trying to use the nose to imitate a resonance.
    We are accustomed to the expression "nasal voice", like some secretary, but this nasal voice, from the popular expression, is cause some people do not use the nose while talking, getting like pinching the nose.

I emphasize also that the design intentions for the YC5 were very well done. It is just me who listened for something else than the intended goal.

Philippe:
I ran the Fourier analysis starting at the beginning of one note up till the sound dies away. I was not very careful though, and samples are not of the exact same length actually.

I used the largest FFT size of 65536 samples and the largest overlap (99%) so as to get the most frequency resolution.

I agree that the Yamaha is drier than the C3 which is more "Steinway"-like I could say. I also remember not liking very much the middle octave on my Kawai upright which sounded a lot like a Yamaha (although brighter).

Maybe my ears are internally connected to my nose... 

Have you tried using a binaural setting if you're playing over headphones? It may reduce what you hear as a nasal sound.

http://picasaweb.google.ca/gllsprs/Spec...onMn8bdeQ#

Maybe Philippe can have a look if he feels like it, and tell me if I'm wrong. I am not an expert in signal processing, so there may be something obvious that I miss.

These were done with an old version of SoundForge (4.5)

Note that the second harmonic is a little bit stronger than the fundamental at velocity 60 for the YC5. We can also see that the YC5 has a lot more high frequency content at velocity 127 than the C3, like it was designed to have.

I might explain the importance I give to separate note timbre by the fact that I like to play polyphonic music. Of course, chord resonances are much more important in actual use and pianoteq has always been outstanding in this much more difficult to attain goal.

I will try explain in words again.

Nose...  I think a nose is missing.

Try yourself imitate the sound of a middle piano note. Ok, now try to imitate this same sound, but try to not use your nose anyway, no sound, no air from it.  Now close the mouth and try to use only the nose.

Well, it's a little like that in kmy opinion. The sounds with nose and the sound withou are different.
In a 100% natural piano sound It's like the note have a kind nose emitting a "sister sound" together.  In pianoteq I think the note progression still lacks a bit of the "nose effect".  Specially from some strong chords like we hear in the linlk I posted above.

That is really strange since the "nasal" quality you get from emitting a sound and pinching your nose is precisely what I disliked in the YC5 middle register (when played softly) and also in the earlier C1 and C2.

Now I understand the YC5 was analysed at high velocity so as to sound best when played hard as a rock piano must be, and that goal is very well obtained. No nasal sound when playing FF and a lot of credible harmonics.

I tried to pinpoint what annoyed me in the YC5 vs the C3 by running some spectrum analysis with a middle C at velocity 60 and 127 on both, but no conclusive result emerged. All harmonics are present with relatively equivalent amplitudes.

The only thing I notice (but can't say for certain that is not due to noise components) is there seems to be a lot more inharmonic components around the fundamental with the C3 at lower velocity.

Maybe this is the explanation. To my ears, single notes on the YC5 at low velocity in the middle range sound too much like simple additive synthesis. Chords are more realistic though.

If anybody is interested, I could post the graphs.