Behavior Modeling Explained -

Forum for JUPITER-80
realtrance
Posts: 128
Joined: 16:39, 8 April 2005

Re: Behavior Modeling Explained -

Post by realtrance »

It's clearly a difficult thing to explain without getting technical; I applaud Vince working on it, and I look forward to the finished draft of his effort.

It sounds like this is really an evolution specifically of what Roland was doing with AP synthesis on the V-Synths, and with ARX as well. It's a form of modelling, which they try to get at with "Behaviour Modelling," but very different from COSM or physical modelling of the past. What physical modelling and COSM were trying to do was model the real-time interactions, basically, of oscillators with resonators -- string with body of guitar or violin, reed with body of flute or oboe, amp distortion with body of speaker type, etc.

What AP synthesis does is model one element of sound generation that in the history of electronic synthesis has been the most difficult one for electronic instruments to capture, in competition with acoustic/"natural" instruments: the phrasings that occur with differences of attack on note-on, the differences in noise in the body/resonator, say, of a flute that come with blowing a note harder, in other words, emulation, using a keyboard interface, of changes in attack and tonality that occur as a result of the varied performance techniques associated with different instruments.

Even with a piano, what happens when you strike a note harder/faster? It changes the timbre of the note as it increases the sympathetic resonance of strings nearby the struck note, all amplified/structured by the resonance of the piano's resonant body. With a guitar, different kinds of strums, at different speeds, don't just play the same notes in different ways, they interact with the body of the guitar and create timbral changes. And so on. Most of the expressiveness of various non-electronic instruments comes out of the mastery of very specific techniques associated with controlling the attack and resonance and timbre generated by the instrument. It's basically an "analogue" phenomenon that is far more complex than, historically, most electronic instruments have been able to capture.

I theorize with "behaviour modelling" that Roland is basically baking in a lot of what they first experimented with with AP synthesis and ARX into a full-scale suite of tones, partials and samples. Anyone who's used Roland instruments over the past fifteen years gets what the whole logic of sample/tone/partial is in Roland architecture, yes? Well here, basically, you get all of that but integrated with "gestural modelling" appropriate for the type of instrument you're working with, using the basic categories that have been familiar since the advent of physical modelling: string, wind, struck string/mallet, piano/e.piano, clavinet. These categories in some ways go even further back, even to pre-electronic organs, where the very earliest attempts to "model" certain aspects of non-organ instruments were present.

"Gestural modelling" is really one of the last, big frontiers in the development of electronic synthesis. It's not that you just want to get an electronic instrument to "perfectly imitate" an electro-acoustic one; it's that taking on the complexities of what happens in the creation and articulation of a note on non-electronic instruments is worth doing, because it draws upon the science of understanding what's going on in those instruments, and applies that science to being able to model complex phenomena electronically that have never before been captured with such modelling.

Once that's available, and you, as a musician, have a chance to use more complex performative technique, and along with that, _combine_ these different kinds of performance and gestural techniques in your performance, be it of a solo voice or a combination of instruments, you have the chance to engage in expressive musical performance with the last boundaries and constraints present in electronic synthesis basically, completely removed!

At that point, with an instrument that gives you such gestural subtlety, you're playing an electronic instrument with every bit as much expressive capability as any of the non-electronic instruments that came before. At that point, it's up to you, as a musician, whether you want to "limit" your performance just to doing an expressive reed, string or electric piano solo, in a traditional, well-understood way, or whether you want to take apart and recombine what was possible with acoustic instruments in new ways, and provide an expressive performance that makes use of previously-understood gestural elements for entirely new types of sound.

I think that's what Roland is doing here, with Jupiter-80. Fortunately, Vince is here to correct me if I'm wrong. :)
vladuca
Posts: 28
Joined: 08:21, 7 April 2011

Re: Behavior Modeling Explained -

Post by vladuca »

I think that's what Roland is doing here, with Jupiter-80. Fortunately, Vince is here to correct me if I'm wrong. :)
Pretty much right on, realtrance!

THANKS!
realtrance
Posts: 128
Joined: 16:39, 8 April 2005

Re: Behavior Modeling Explained -

Post by realtrance »

No thank you, sir, and Roland; glad to be of any help. Good luck with the release of this instrument! It strikes me as a very decisive move for the company.
vladuca
Posts: 28
Joined: 08:21, 7 April 2011

Re: Behavior Modeling Explained -

Post by vladuca »

Leh173 wrote:The specs on Roland's site lists this: Harmony Intelligence, 17 types



Somone can explain better what is it? I didn't understand it.
I have yet to try it on the JP, but from my experience with this feature in other stuff, the upper part will have a melodic harmony above what you play solo based on the chord voiced in the lower part.

I think Scott demo'ed this in one of the JP videos out.
Mystic38
Posts: 1105
Joined: 14:04, 24 August 2009

Re: Behavior Modeling Explained -

Post by Mystic38 »

Actually, i read this as

"so people who play keyboards can sound like they know how to play the guitar, trumpet, violin and zamboni" :D
kenchan wrote:so we can all sound like we know how to play keyboards now? :D
Post Reply