How is sympathetic resonance made ?

[maxpiano]

Nord pianos (acoustic and electric) use no modeling whatsoever. Everything you hear is a recording. It may be stretched, it may be looped, it may have processing, but you are always hearing a recording.

String ressonance is modelled.

Ramon

I don't think so, on the Nord Piano library pages themselves you can actually read about the different versions that (example):

The Large versions have String Resonance samples all across the keyboard, providing you with a very full sound.

Don't confuse algorithms like "intelligent Sample processing and scripting" with pure Physical modeling, where 100% (every single bit) of the resulting waveform is calculated by algorithms (such as in Pianoteq or Viscount Physis piano)

It is like confusing a movie editing and processing tool (e.g. Adobe Premiere) with a RayTracing or 3D modeling software.

[mon8169]

The Large versions have String Resonance samples all across the keyboard, providing you with a very full sound.

This is pedal down samples, not sympathetic string resonance. The stage EX has this algorythm but not the original stage, which can play the pedal down samples.

Ramon

[maxpiano]

The Large versions have String Resonance samples all across the keyboard, providing you with a very full sound.

This is pedal down samples, not sympathetic string resonance. The stage EX has this algorythm but not the original stage, which can play the pedal down samples.

Ramon

True but the same samples may be used by the algorithm that implements the Sympathetic Resonance, anyway unless you have a different and official information from Nord programmers about this we both are just speculating. Overall the Piano engine of Nord is sample-based rather than modeled.

[mon8169]

I'm not speculating. The Piano engine is sample based and the Sympathetic Resonance is modelled.

Sympathetic Resonance cannot be sampled because it depends on the already held down notes when you press a new note down, so it's impossible to sample so much possible combinations.

For this reason the original classic Stage can play all the samples (even de pedal down ones) but it doesn't have the Sympathetic Resonance built in (it doens't have enough power/memory to run the Sympathetic Resonance algoryhtm)

Ramon

[maxpiano]

I agree that the reason why the Classic cannot run Sympathetic Resonance algorithms is due to the processing capacity of its DSPs (older than the EX's).

I would like to be sure instead that Sympathetic resonance is 100% "physical modeled" and not instead also using the same String resonance samples + additional processing (you don't sample the combinations, but "assemble" them)
Would be nice to get a hint from a Clavia programmer, anyone passing by here?

But we are going pretty OT now

[Frantz]

& Interesting.

Dunno the answer and I'm not sure we will know what's done inside our red keyboards, but maybe more generally.
Let's but the question into it own topic, have a look on the web, and hope someone who knows will bring more knowledge.

In sympathetic resonance computation, I'm not sure that the combinations could be simplified to trigger a few samples containing the proper harmonics.
I guess that it depends on the number of overtones taken into account.
For quick fun, let's take only the fourth of each pitch, that makes ( 12 choose 2 ) = 12! / 2! (12−2)! = 66 possibilities per octave.
If we work the frequencies and factorize them, maybe there is only a few samples left, I don't really believe it.

The question remains, how is sympathetic resonance made ?

[juhana]

& Interesting.

...

The question remains, how is sympathetic resonance made ?

I know as much as anybody else, but to me Nord having the special NPNO format for piano sounds suggests that there is something else than just (multilayer) samples going on, i.e. DSP code that would model some aspects of the instrument. That would also explain the different versions of the sounds in piano library (hardly necessary to version plain samples?) as well as the really small size of the older EPs (could you really make those instruments with that small samples?)

Anyway, just speculating, but I believe there is some circumstantial evidence for the NPNO sounds being more than just samples.

[Frantz]

Just found the following article about "Kontakt PMI Pianos" and sympathetic resonance : http://www.postpiano.com/products/K2.htm
Simply :
1- have an algorithm to decide what sample to trigger, according to what is played.
2- process and trigger them !

That could be : "trigger some E 330 Hz when A 440 Hz is played, because they share an overtone of 1320 Hz (3rd harmonic of A and 4th harmonic of E)".
Just guessing
Frantz.

[mon8169]

& Interesting.

...

The question remains, how is sympathetic resonance made ?

I know as much as anybody else, but to me Nord having the special NPNO format for piano sounds suggests that there is something else than just (multilayer) samples going on, i.e. DSP code that would model some aspects of the instrument. That would also explain the different versions of the sounds in piano library (hardly necessary to version plain samples?) as well as the really small size of the older EPs (could you really make those instruments with that small samples?)

Anyway, just speculating, but I believe there is some circumstantial evidence for the NPNO sounds being more than just samples.

It's also my guess, my speculation. I wrote some time ago that it could be that the Nord pianos only had one velocity layer and that all velocity variations were programed with filters and other methods. The reason I think so is the size of the samples (apart from compression) and that I still haven't heard any velocity switch in any piano (but maybe I'm deaf )

I'm just speculating here...

Ramon

[wartaler]

This is what I think;

Small size of Pianos in the library: -sheer speculation of course; the samples aren't regular samples, but something like MBE, 'vocoder'-like (with enough bands so that it doesn't sound 'robotic')
Such codecs are quite memory efficient I think. With the simple 16 band vocoder of my little R3 I can get quite a nice EP like sound already..

Symph. SR: easiest accomplished with synthetic sounds/wavelets, if you want to capture the characteristics of a real piano, you could record an individual resonating string - by swiftly dampening all other strings, including the one played, directy after playing. The Nord board requires a bit of processing power to apply rules for when, how loud, & which resonating notes to output.

Pedal down: the entire piano sampled again. but now with the pedal down

Albert