Uneven partials

The relative height of each partial follows from the physics laws, namely:
- the inharmonicity of the string, which depends on the string length,
- the soundboard impedance.
There are no other controls for the partials height. I didn't mention unison width as this parameter acts directly on the fundamental, and the rest follows from the 2 laws above.

That means that you do have (an indirect) control on the partial height via those two parameters.

Hey.
Still, I am very impressed with the program Pianoteq. But when I play fast dry and jerky parallel thirds or too noticeable unnatural sound (in the range of A4-A5). Notes in the attack too similar. I noticed that a live tool has more fake at the beginning of the sound. This is particularly noticeable in the room with a long reverberation. Returns attack superimposed on sounding purely in his tone duration.
I saw again in some small discrepancies location overtones in the strings and choir is not clear according to a mathematical formula. Perhaps the reason for this is a small string of corrosion over time, settled dust (or a fly sat))). Possibly, if the string is longitudinally twisted a little, it shifts the location of overtones relative to each other. Then the disharmony of the string does not change, but there is a slight discrepancy between the height of the overtones from a mathematical formula. Some overstate overtones, some underestimate. The program Pianoteq location overtones sound almost perfectly coincides with the formula Tunelab program. In the real instrument, differences arise with the formula. In Pianoteq using slider "width unison" Ideally located ideal overtone too divergent. It evokes a feeling of similarity of sound (the attack). The auditory tract is a certain group of neurons, which is responsible for the perception of innovation in sound. Perception of timbre depends to a very large extent on the attack. I think the nature of falsity of the string at the beginning of the sound (the first milliseconds) might be something to do with the location of the overtones. Accordingly, the three strings of the choir, with some minimal differences from each other can give different falsity in the attack, which overlap each other and in the end will be beautiful, individual, velvety beginning of the note. And that would be different from note to note.
But I push the pedal - and oh !!! I'm drooling with pleasure)))
I wonder, am I right or mistaken? Your thoughts on this subject ???
Thank Sche)

The overtones in Pianoteq are very close to those in real instruments, because they are computed via a complex model including not only the string inharmonicity but also the influence of the soundboard impedance. I never tried to measure it with the Tunelab program but maybe this is not so easy to measure?

I am happy for you

When I tried to compare the timbre of my Erard model2 (2.48m) with the Erard model1 1922 (2.12m) from the Kivir collection, I (until now) got my best results by adjusting pianoteq the length of the strings to try to match (more closely) the frequencies of the partials with those given in the frequency table for each partial of a note of the spectrum profile of pianoteq pro. I was surprised to find an optimum at string lengths very different from the ratio between 2.48 / 2.12m between the 2 instruments. On the other hand, I did not modify the ptq parameter "soundboard impedance".
Q1? Is there a way (without calculation too complex) to approach the soundboard impedance value of pianoteq note by note, bringing it closer to spectrum or envelope parameters of a recording of the acoustic piano for a given note?
Q2? Is it advisable to play first on the impedance and then (or not) on the string length or should it play only on the length of string and not on the impedance of the soundboard?)
Q3? For residual spectrum profile adjustments (which I am trying to do manually now on this Erard model2) note by note, today the frequency of the abnormal partial either is obvious and it is possible for me to adjust it directly,
or (often) the impression of abnormal timbre is (for me - not a professional) diffused and I try to remove (-60db) progressively the partials going from the highest to the smallest, stopping at the first attenuation of the elements of stamps "abnormal" to decrease the smallest partial from which I notice a noticeable difference. (iterative process after each elementary modification of a partial in this context).
Q3.1?) Is this approach correct? (should we go in the other direction from the smallest partial to the highest?)
Q3.2?) Sometimes, by reducing (even very little) the level of one of the partials, I notice the appearance of abnormal "ghost" levels at frequencies of other partials whose level was initially correct. I suppose there is a perception of differential frequency between several partial frequencies, perhaps intermodulation effects. Which method can be used to identify them in order to control their level?

Q4) Which parameters ptq to adjust the duration of the phases of the sound envelope? soundboard impedance & Q factor do not seem sufficient (global action only, not the relative duration of the phases between them)

Thanking you in advance for some elements of answer,

Bruno

After different tests I did not (yet ?) find any parameter to correctly adjust the amplitude of the sound envelope during the first 0.2 seconds compared to the persistent level after 0.5 seconds of notes such as A3. This is detrimental to the quality of the attack of the note.
Maybe we need in Pianoteq pro a envelope editor NOTE by NOTE in the way of the spectrum profile? (necessarily separate editor for Rights and Left channels, and perhaps "ideally" as separate for piano / mezzo / forte levels.
(Would the power of today's laptops allow it ...)

link: https://lh3.googleusercontent.com/GOM1K...64-h813-no

Bruno

@bm - maybe I am naive to offer the suggestion of Direct Sound Duration, in balance to other worked factors such as energy and inertia (right-click them for humanize function) - It to me, gives a different 'sense' of depth to the stereo field which feels more natural like I hear in a room (reflecting on my impression of this, I could say it seemed to have less phase issues in the field), attack and early tail.. worth mentioning in case it's overlooked or your actual solution, although sometimes I wonder if you're attempting something else in greater depth of course.

@Amen, I like that list but having trouble seeing which item you mean.

I like the idea of per chanel volume out for each mic (after we balance the L&R volumes in the interface, being able to sum the mics in total would save many clicks) that way we could probably more quickly adjust things for phase free stage. I think of Zooming in for a L&R volume per note - could be powerful for placing any note anywhere in the image (maybe in the moment I'm forgetting some way we can already do this - and it doesn't really relate to the thread really).

This kind of thread is always interesting. Thanks to those digging in and always exploring.

I thank you for these suggestions.
I had already tried direct sound duration, with no real change of the envelope. With the impedance setting the action is mostly on the sustain phase. For the phase just after the attack (mainly for A3 between 0.1 and 0.5 seconds after the start of the note, (see link) I still have not found a solution, I tried the blooming parameters the day before yesterday. actually have a small influence on the envelope, but not enough for this particular phase.
I have the impression that the shape of the envelope obtained with Pianoteq is a composite obtained with the combination of sounds obtained with several microphones, hence certainly the interest of being able "one day" to control the envelope note by note , for each microphone input (if the power of laptops allows it without having to use fans making a plane noise ..)
If any of you have an idea to "thin" the envelope right after the attack I am a taker ... even if I'm not sure of the improvement of the sound that we could get with that.
https://photos.google.com/share/AF1QipP...RzNEs5clNn

Bruno

Bruno, if you want a thinner envelope in the sustain part, you should increase the direct sound duration, not reduce it (your picture mentions 0.35, I suggest you to try instead values beteween 1 and 5). Reducing the impedance was correct though.