https://www.youtube.com/watch?v=lF4mU64MtVs

The other thing that just perhaps might be a factor is the "smoothness" in the attacks in that demo - the way the "clunk" of the hammer very smoothly and beautifully morphs into the sustain - I'm not sure Pianoteq does this as well yet.

Greg

Gosh, you're all talking in such delicate metaphors.  Eran's piano is simply a little bit out of tune!

You can put Pianoteq out of tune too if you want.  Use the unison width slider (top left) or the condition slider (just under the picture of the pedals).

Philippe Guillaume has himself expressed something which just might support everyone else’s singing ideas —metaphor or other— about Unison width and Direct sound duration parameters combined:

See forum topic: Getting Started Modeling.  Although it is basically about mallet instruments, Guillaume makes no exception in it to the U4 upright piano model.

If everybody can already agree initial note singing or ringing is sometime affected by Unison width, perhaps everyone will eventually see enough to learn that that parameter in initial stages at least depends upon the Direct sound duration of the voiced note.

You can put Pianoteq out of tune too if you want.  Use the unison width slider (top left) or the condition slider (just under the picture of the pedals).

Another thing to consider is that with an upright piano, you have a piece of wood between the microphone/your ears and the strings.  You're not getting the direct sound that you get from a grand piano, so it's slightly more muffled, and this might contribute to your perception of it as more "singing".  (Some upright pianos have harder hammers to compensate for this.  But Eran's video sounds quite nice and mellow.)  On Pianoteq, you can close the piano lid, or you can change the microphone position to get different effects.  Try the microphones on the "wrong" side of the lid (to the left of the piano from the player perspective), or behind the player, or underneath the piano.

Una corda pedal might help too.  As well as the pedal, use the soft pedal slider to change the amount of pedal effect.

Did you actually measure it?  Unfortunately it looks as though we can't attach images in this forum.  But I'm looking right now at a spectrogram of a single note recorded from Pianoteq.  (To get the sound of a single string, instead of three strings, set unison width to zero, or use a low bass note.  Or record a harpsichord instead of piano: it's the same physical modelling engine, just different attack.)  It clearly shows the different harmonics decaying at different rates, including some beating (you might call it wavering: some of the harmonics drop in volume, then rise and drop again over time).  There's plenty of free spectrogram plugins out there, so it's possible for anyone to repeat this experiment.

Check Balázs Bank's papers at http://home.mit.bme.hu/~bank/publist/ if you want to learn what goes into a physical model of a piano.  I don't know if Pianoteq uses these exact methods, but I expect it would be something similar.  Certainly what I'm seeing in the spectrograms lines up with the descriptions in those papers.  A good starting point is http://home.mit.bme.hu/~bank/publist/taslp10.pdf  It describes something called modal synthesis, where each harmonic gets its own volume and its own decay rate.  And yes, it includes different directions of vibration.

Very nice sound Etalmor - like milk and honey. Lovely.
Greg

Thanks a lot Greg, I appreciate it :-).

--Eran

I guess I've grown accustomed to this imperfect sound ... I do remember once playing a Fazioli grand in a store, the lid was closed, but it sounded so "perfect" that it was almost synthetic ...

Greg

Hi dear forum,

<...>

Here in my video "When I Fall In Love" played on my old Yamaha U1 I can identify this singing quality, even though the recording quality is low.

<...>

Any thoughts?

Thanks, Eran

I viewed and listened to your video of playing "When I Fall In Love" on your Yamaha Upright piano.  The singing quality I believe you are hearing is a "chorusing" effect of unison strings being out of tune with each other.  This chorusing effect seems to bloom or unfold as the strings' vibration patterns shift from moving parallel to the strike of the hammers to perpendicular to the strike of the hammers, courtesy of the agraffe on the bridge.  Actually, it's not a 100% shift from parallel to perpendicular vibrational movements, it's closer to the strings vibrations occupying a sort of oval cross section (partially parallel and partially perpendicular to the hammers' blows).  When the strings (in free vibration) have switched to this more complex pattern, the ear perceives that as
blooming or singing.  I believe that if your upright Yamaha piano were freshly tuned, the effect you described would probably be diminished.

Cheers,

Joe

Hi dear forum,

As much as I adore Pianoteq, (company, product, project, and community),
I still find it missing some "singing" quality. To put it in other words, I feel that in an acoustic piano the raw tone (of a single note) has more energy and it keeps "evolving" immediately after the attack.

Here in my video "When I Fall In Love" played on my old Yamaha U1 I can identify this singing quality, even though the recording quality is low.

I find that increasing the "direct sound duration" gives some of this effect, but it's not a solution for me, because I'd like to hear this quality only immediately after the attack, and not affect the entire decay ...

Perhaps an enhancement to the "direct sound duration" could help: A "direct sound envelope" ... i.e. control this energy as a function of time (time is relative to the natural decay)

Any thoughts?

Thanks, Eran

I would argue if I may, Unison width, Duplex scale, and Damping duration combined with the Direct sound duration are together what might make your piano sing.

However Soundboard parameters Impedance and especially Cutoff and Q factor affect the decay factors in any singing.

Although few if any of the parameters (above) were adjusted in my preset, I've a possible singing example too:
https://forum.modartt.com/download.php?id=3871
https://forum.modartt.com/file/1rv9licw
Hungarian Rhapsody No.2 MIDI file

If you need to do an A/B comparison of the files from DonSmith and me, I suggest raise the Volume level on my preset to -8dB about almost equal to Don's.

If you have the Pro version try editing the harmonics in the Spectrum Profile note edit section. I've uploaded an FXP.
https://forum.modartt.com/file/40rmg1jr

Now if anyone is an end user qualified to give advice on making presets, certainly DonSmith is the man!  He has consistently been making presets and posting them freely and unselfishly whereas others can benefit from them.  So, anything he adds about them comes tried and true.

(a) What is the actual behavior of a real piano which humans are able to perceive.
(b) Our knowledge of the behavior of the piano
(c) Our knowledge of psycho-acoustics
(d) Physical/Mathematical models for (b) and (c)
(e) Efficient computational models for (d)

I would like to see an additional spectrum adjustment feature that would cause certain and chosen harmonics to vary in amplitude (and even frequency - maybe). ... I think that an actual acoustic piano single string note emits the individual harmonics in varying amplitude/frequency/stereo field.

I think that one thing that causes the varying amplitude of the various vibrating full and fractional lengths of a string is that, unlike a two-dimensional representation of the vibration on an oscilloscope, a struck string vibrates in different directions perpendicular to the string at different times (anyone can see that a bass string on a grand piano, when struck, is not vibrating in just one direction, not just up and down or sideways, but in all directions), which it seems would cause a variation in amplitude from any particular listening or microphone position.

In addition, there are two contrary forces at work which interact in complex ways regarding a musical-instrument string: the tension of the tuned string itself, and the force with which it is struck or plucked. Greater force can momentarily increase or sharpen the frequency of a note, while the tension of the string acting against that force and pulling the string back to a static state can decrease or flatten the frequency (which led to stretch tuning--the highly-tensioned treble strings of a piano tuned progressively a little sharper, because they would decay a little flat (and sound more in tune as they decayed), whille the laxly-tensioned bass strings would be tuned progressively flatter, because the harmonics are more prominent than the fundamental frequency, and the harmonics would normally sound a little sharp when the string is struck, unless the string's fundamental frequency is tuned a little flat).

So there seems to exist a complex variation and interaction from moment to moment of both amplitude and frequency in the physics of a struck physical string, a kind of natural tremolo, and a variation in frequency, slightly sharpened when they are struck, and a little flattened as they decay, of both the fundamental tone to which a string is tuned, and its harmonics.

Now if you limit a given note to one string, the individual harmonics will fall/decay in a steady but unwavering  rate as observed on a spectrum scope display.

I would like to see an additional spectrum adjustment feature that would cause certain and chosen harmonics to vary in amplitude (and even frequency - maybe).  One might say that adding chorus effects that … not quite. 

I think that an actual acoustic piano single string note emits the individual harmonics in varying amplitude/frequency/stereo field.  Artificial chorus added does not render the same affect.

As an analogy to liquids - chorus would be like mixing four ingredients - for example - then gently shaking them as a whole.  But I am suggesting that if one could continuously add and remove the four ingredients individually and cyclically the result would not be the same.  I think a struck acoustic piano string magically emits its individual harmonics in such a manner - whether or not there are unisons. They certainly don’t emit themselves from a fixed point on the string - each harmonic floats and decays independently along the length of the string.

Lanny

Thanks Don!
Can you please explain how the specific ranges of note harmonics you changed + temperament, help to this effect?

It's hard, to explain, [Spectrum Profile - Note Edit] working left to right, the main characteristics seem to be the ability to create a prominent sound when you raise, or lower sections sharply and a more blended one when you use to smooth tool to transition the change. After a while, you get used to what happens when you make adjustments to it. If you open the FXP file and look at the edit profile you'll see where I've tweaked it.

After you get the tone and feeling in the keys you like, you could tweak the behavior of the sound with the other features: velocity, mics, tone, pitch, sound speed, tuning: scala-files, etc.

I believe if you knew the behavior of a piano sound and could tweak the corresponding features in Pianoteq, then you should get the same sound. I think most of what we hear is about the behavior of the sound. That's sometimes hard to pinpoint. Pianoteq is so well constructed to features of a real piano, it's hard to know what other features could be added to the software.

etalmor wrote:

Hi dear forum,

As much as I adore Pianoteq, (company, product, project, and community),
I still find it missing some "singing" quality. To put it in other words, I feel that in an acoustic piano the raw tone (of a single note) has more energy and it keeps "evolving" immediately after the attack.

Here in my video "When I Fall In Love" played on my old Yamaha U1 I can identify this singing quality, even though the recording quality is low.

I find that increasing the "direct sound duration" gives some of this effect, but it's not a solution for me, because I'd like to hear this quality only immediately after the attack, and not affect the entire decay ...

Perhaps an enhancement to the "direct sound duration" could help: A "direct sound envelope" ... i.e. control this energy as a function of time (time is relative to the natural decay)

Any thoughts?

Thanks, Eran

If you have the Pro version try editing the harmonics in the Spectrum Profile note edit section. I've uploaded an FXP.
https://forum.modartt.com/file/40rmg1jr

Thanks Don!
Can you please explain how the specific ranges of note harmonics you changed + temperament, help to this effect?

Hi dear forum,

As much as I adore Pianoteq, (company, product, project, and community),
I still find it missing some "singing" quality. To put it in other words, I feel that in an acoustic piano the raw tone (of a single note) has more energy and it keeps "evolving" immediately after the attack.

Here in my video "When I Fall In Love" played on my old Yamaha U1 I can identify this singing quality, even though the recording quality is low.

I find that increasing the "direct sound duration" gives some of this effect, but it's not a solution for me, because I'd like to hear this quality only immediately after the attack, and not affect the entire decay ...

Perhaps an enhancement to the "direct sound duration" could help: A "direct sound envelope" ... i.e. control this energy as a function of time (time is relative to the natural decay)

Any thoughts?

Thanks, Eran

If you have the Pro version try editing the harmonics in the Spectrum Profile note edit section. I've uploaded an FXP.
https://forum.modartt.com/file/40rmg1jr