Synth designer Jürgen Michaelis has announced preorders for a limited run of the Jaymesonic Resonator Neuronium synthesizer, a synthesizer that offers a unique approach to synthesis:
Because we get periodically inquiries on the Resonator Neuronium, we have opened a new preordering cycle until December 31, 2010. If more than 20 preorders come in until then, we at Jomox can manufacture a new run of the R.N.
The Resonator Neuronium has 6 analog resonant neurons. The net parameters can be programmed by the left hand structure in form of a hexagon by use of knobs, buttons and LEDs.
The cost is 2499.00 €.
i'm having a lot of trouble understanding what makes this synth so expensive. anyone care to explain? http://www.youtube.com/watch?v=URPDalPo44M&fe…
Yes. No synthesizer can make trampled mouse or cat-in-a-dumpster noises like this one. Well, okay, my home-made cacophonator can, but this looks at least 2479 euros nicer.
Yeah, I'm really not seeing the appeal of this. The vast majority of the demos sounded like what would happen if you wiped a window really fast.
As far as I can deduce (keeping in mind that I've never encountered one in person, and haven't even read the manual yet), part of the appeal of the RN is that it's completely unlike anything else in terms of sound or approach to producing sound. It's unpredictable in the same way organics are unpredictable. Unfortunately, that also makes it a serious bitch to program, it's finicky as all hell, and getting a usable sound from it is more a happy accident then the result of planning.
To paraphrase an old car commercial featuring James Garner, this is not your fathers synthesizer. This isn't even your crazy uncle Steve's synthesizer. This is a beast from a completely different planet of sound. You don't so much program it as interact and attempt to negotiate with it.
From skimming the description at http://www.jayemsonic.de/2l4-resonanteneuronen.ht… I get the impression that the basic idea is less about neural nets, but about a certain view of the Moog transistor ladder circuit. The philosophical and mystical stuff on that page made me skip reading the text after a few sentences and just look at the figures which make a lot more sense than anything I've read about Michaelis "neuronal" stuff before. Someone with a little free time and a soldering iron should look into building an analog version of the Resonator Neuronium modulo the genetic algorithm stuff.
Bottom line: It not only seems to be a combination of two theoretical concepts that are often used to fool the uninformed public into thinking a product is rocket science: Neural nets (almost always in their second most basic incarnation) and genetic algorithms (preferably with copious references to biological terms). It's like giving a 15 year old (or thereabouts) some popular scientific magazines to study and having him come up with a few product ideas afterwards. 😉
Cool! Just $2.5K. Dude, you've got my order. Not! If they made a software ver. for $25.00 I still wouldn't buy it. Cuz it jus don't sound gud.
Yeah, I picked up on the references to clever design concepts and programming techniques too. Unfortunately, if the end result still sounds like a circuit-bent Casio toy, that is bound to limit the appeal somewhat, especially at this price. However, it's an attractive-looking gadget and I'm sure that plenty of overpaid DJs and noise musicians will have to have one.
Actually, one of the YouTube demos had some interesting sounds; the rest were of the "look what I've got I'll just turn the knobs aimlessly for six minutes and you will be impressed" style. It's possible that if someone took the time to learn how to use it properly, the vids could be more impressive. But I won't be finding out.
Having now read the manual… it's actually really interesting from an experimental sound design point of view. Let me see if I can boil it down, stripped of the bio references:
You can think of each "neuron" as a little stack of synthesizer functions: DCO, VCF, VCA, plus microcomputer controller that has a little sequencer for the VCF and a program that controls the DCO waveform. Then you take these six little synths and cross wire the hell out of them. You can mix the audio outputs of the six little synths; you can fiddle the filter FM network layer; you can fiddle with the programming of the microcomputer in each little synth controlling the DCO waveform so that it reacts to the conditions/output of the other synths in the network. And then there's an overall micro-controller and VCA for the whole box of tricks after the mix layer (before the main outs).
It's deep as heck, and the manual is about as informative as the DSI Evolver manual: tech specs, not programming suggestions or explanations. But it sure sounds like fun… I tried to imagine an equivalent built in a modular, and it just got huge. And the amount of inter-node interaction and feedback loops would make it difficult to implement in software without the risk of locking up the host computer is eternal recursive function calls.
This is probably best though of as a sound design lab set, rather than a "synthesizer" as the term is commonly understood.