VSL Synchron Bösendorfer Imperial

I'm not convinced what Pianoteq do with their models is really emulating the pianos they mention. I believe they just pay the corresponding piano company to use their brand and then tweak their generic model to fit the stereotypes, like e.g. Yamaha is "bright and punchy for jazz", Steinway is "dark and soft for classical", etc. You don't believe they get a brand new Steinway which they dissect to measure or burn to make a spectral analysis of the lacquer, do you? 🤣 We've hypothesized that they take audio samples of a piano and the synthesis method of Pianoteq is actually a clever compression that does some Furrier stuff and sums sine waves and coefficients and whatnot (as described in their old patent) but they have been slippery and said that's not the case anymore. So what is it now? Well, as I said, they just have some generic synthetic engine that can be tweaked to be brighter, duller, etc., then they say: OK, we will call this one a Petrof since "Petrof" ends in "f" which in musical notation means forte, and we think this current tweaks sounds forte, let's pay them a few thousand dollars and see how it goes 😉 I'm just speculating of course. But prove me wrong!

CyberGene You don't believe they get a brand new Steinway which they dissect to measure or burn to make a spectral analysis of the lacquer, do you? 🤣

Yes, I do, CG!

Unless you’re saying they’re lying, are you?

Yes, they actually got a ‘used’ Steingraeber for a decent price; which they then dissected and further deconstructed all the way down to the smell and the lacquer!

Once they’re done, the piano in question is put up for sale; however, modartt now keeps the essence of that instrument in their secret vault (for further improvements).

CyberGene 😉 I'm just speculating of course. But prove me wrong!

I say we send some of our European members to the Modartt facilities in France for some deep espionage dumpster diving, to look for traces of the pianos they have burned. Only then will we know for sure.

Ralphiano I say we send some of our European members to the Modartt facilities in France for some deep espionage dumpster diving

I promote our dear @Frédéric L for that mission since he must be closer to those dumpsters and is the only one who can understand French. I tried to study French for some time and gave up. Has any of you tried to learn French? It's incomprehensible. You have something like beaucoup which is not short for beautiful cupcake. And everything is a variation on that theme.

Pete14 I wonder if VSL has some ‘exclusivity’ deal with Bosendorfer (owned by Yamaha of all things); otherwise, why don’t we see a Bosendorfer VST out there from the big guns (Pianoteq, VI Labs, etc)?

Perhaps Bosendorfer doesn't want its name sullied by the likes of Pianoteq?

CyberGene I tried to study French for some time and gave up. Has any of you tried to learn French? It's incomprehensible.

I started learning it quite young, though that was so long ago that my skills have diminished.

The fundamental language idioms are not unlike those of English, and the two languages share many words. Indeed, the Norman invasion instilled French into the English language. So I find it not so difficult to handle French.

The hardest part: verb conjugation. English has an utterly different way of handling verbs. For that reason French, Spanish, Italian, and Portuguese are all difficult for an English speaker.

CyberGene 583km from here… But I guess the travel is not the hardest part. This depends of their security crew. If they are well organised, they use paper grinders for important papers before trowing them in a trash.

CyberGene One clue : one of their model is named Bechstein Digital Grand. They tried to emulate a Kontakt library. This does not imply calculating mechanical impedance or such thing. Then trying to match a sound by tweaking their model parameters seems the most probable way.

Pete14 I wonder if VSL has some ‘exclusivity’ deal with Bosendorfer (owned by Yamaha of all things); otherwise, why don’t we see a Bosendorfer VST out there from the big guns (Pianoteq, VI Labs, etc)?

Both the Ivory and Galaxy collections include Bösendorfers. I think Modartt may have trouble modelling the effect of the rim in Bösendorfers, because it is not layered like the other makes (as far as I know).

CyberGene I'm not convinced what Pianoteq do with their models is really emulating the pianos they mention

They ought to at least measure and use the various distances that distinguish the different models they simulated. If not, I would consider their alleged diversification dishonest. But I can't see how they would even be able to build their models without these data (unless they were guessing at them, but that would be silly, given how easy it is to take a tape measure and use it).

QuasiUnaFantasia They ought to at least measure and use the various distances that distinguish the different models they simulated.

Then can do something simple : measure how fast the sound is decayed, partial per partial. Identify multiple phases of the sound (prompt sound and aftersound)*. Identify beats. Then adjust the sound engine parameters to match these measures : no measure of the piano itself (distances, impedance…)

* See https://www.speech.kth.se/music/5_lectures/weinreic/motion.html

QuasiUnaFantasia Both the Ivory and Galaxy collections include Bösendorfers

Yes, but these -Ivory and Galaxy- are minor leagues and not the big guns 💪

I’m surprised they still hang in there, but who knows, with the right attitude and hard work they might make into the big leagues some day!🤣

Now, about this rim you talk about, I don’t think the king of VSTs (Pianoteq) would have any problem modeling it.

Pete14 Now, about this rim you talk about, I don’t think the king of VSTs (Pianoteq) would have any problem modeling it.

I don't think so either, as far as the mathematics is concerned, but with a solid rim the entire surface of the piano starts vibrating along with the soundboard (whereas with a laminated construction there should be less vibration of the piano surface), and this increases the amount of calculations considerably, which would probably necessitate beefier CPU's.

There are EWQL and UVI which have a Bösendorfer too. It is quite common !

QuasiUnaFantasia and this increases the amount of calculations considerably, which would probably necessitate beefier CPU's.

If they model the sound and not the piano itself, this would just mean different parameters (longer decay by example), but no more computation.

Hey, I'm French (was born there), but at this point have spent more than half my life in the good old US of A. But Moddart is near Toulouse and whenever I am in France (every 3-4 years or so), I tend to not go by there, although of course I have been to the "pink city", a lovely part of the country. While you're there, go to Bordeaux following the Garonne river, stopping as many times as you can among the various wineries! 🍷

Why do I answer in a thread about Pianoteq and Bosendorfer? I don't know: I still don't do VSTs at this point, so please resume your debate: don't mind me. 😅

QuasiUnaFantasia I don't think so either, as far as the mathematics is concerned, but with a solid rim the entire surface of the piano starts vibrating along with the soundboard (whereas with a laminated construction there should be less vibration of the piano surface), and this increases the amount of calculations considerably, which would probably necessitate beefier CPU's.

I am not so sure that the allegedly "solid" rim of the Bosendorfer is any better than a high quality rim. It is my impression (one that I don't mind having corrected) that the bending of the Bosendorfer "solid" rim is achieved by cutting notches in the rim on the inside of any curve. As the wood is bent, the notches (absence of wood) allow the bending without imposing destructive stretching on the outer surface of the bend. It would be much like automotive drive belts that have notches on the inside surface to allow the belt to bend around the small radius pulleys of the AC compressor, alternator, and other belt driven devices.

It seems that the numerous breaks in the wood fiber created by these notches would substantially change the nature of the wood and its resonance characteristics, perhaps similarly to the resonance damping effect that the glue layers in a laminated soundboard/guitar top do.

I think I got the impression of the notching process from a YouTube video, perhaps from the Bosendorfer company.

Ralphiano I believe the main point here is that the sound propagation is not the same from a solid structure as it is from a laminated one (where there are reflections from each layer boundary). I have personal experience with my familiy's 19th century Hornung & Møller grand, which also has a solid rim, and the outside of this rim vibrates intensely, when the piano is being played. Therefore you could not model the sound propagation correctly if you disregarded the rim. But I suspect (without having any concrete evidence) that Modartt has always disregarded it.

QuasiUnaFantasia I have personal experience with my familiy's 19th century Hornung & Møller grand, which also has a solid rim, and the outside of this rim vibrates intensely, when the piano is being played.

To satisfy my curiosity, have you ever investigated how the bending was accomplished on your family's Hornung & Moller? I'm very curious. I once built my own mandolin, which required some bending. However, those woods were only about 1/8 of an inch thick, and I do not think those techniques (steam softening of the resins) would work on wood as thick as a piano rim. So, I'm really curious how wood whose thickness might be an inch or more, gets bent without suffering destruction.

Ralphiano I'll have to check next time I get the chance (fairly soon).

Ralphiano Okay, I've checked now, and it was quite a surprise, because the Hornung & Møller is different from both the modern laminated ones, and the Bösendorfer style.

The rim is in fact layered, but the layers are stacked on top of one another, each layer being a couple of inches thick. The curve in the rim seems to be sawed, so basically a number of similar pieces of wood are cut to fit the curved shape, and then they are stacked onto each other. No bending of wood necessary!

Moreover, the rim is something like 4 inches thick!

By the way, the Hornung & Møller grands have a very distinct sound. To my ears they are unique, with all other brands sounding quite similar to one another.

QuasiUnaFantasia The rim is in fact layered, but the layers are stacked on top of one another, each layer being a couple of inches thick. The curve in the rim seems to be sawed, so basically a number of similar pieces of wood are cut to fit the curved shape, and then they are stacked onto each other. No bending of wood necessary!

How interesting! This would have been the least expected possible answer!

The construction technique you described would result in a rim that is completely, or nearly completely, tension free. There would be no forces trying to straighten the bends back to the wood's natural state.

This might explain the unique sound. I remember a demonstration of how tension in the wood forming the sound transducer increases amplitude as compared to a tension free sound transducer. I cannot remember if the demonstration tried to explain tonal differences. But, its overall point was that when a soundboard is built with tension tending to oppose the force of the strings acting upon it, it produces much more sound. Thus, the arched design of many string instrument tops, including the piano.

Although the piano rim is not acted upon directly by the strings, it's unification with the soundboard where they meet would make it a part of the overall transducer. And, perhaps the Hornung & Moller differs from the absence of sound contribution from the rim that other pianos exhibit.