Jupiter 80 VA engine - new or same as GAIA?

Forum for JUPITER-80
User avatar
PauloF
Posts: 4201
Joined: 02:35, 16 January 2006
Location: Lisbon, Portugal
Contact:

Re: Jupiter 80 VA engine - new or same as GAIA?

Post by PauloF »

kokocalamar wrote:
Mystic38 wrote:The JP80 has clearly been stated in this thread many times to include a VA section.. move along huh?
No, I'm staying. I like it here!
Lol ;-p
ozy
Posts: 169
Joined: 13:32, 1 December 2010

Re: Jupiter 80 VA engine - new or same as GAIA?

Post by ozy »

aron wrote:In any case, here is how you can tell if a sound generator is sample-based:
1. Detune two waveforms of the same type: you will hear that their phases restart with each keypress, because all samples start playing back from the beginning; on a real VA the oscillator always oscillates so each keypress will make a slightly different sound
2. Hold a note, and do a pitch-bend down an octave or more: you will hear that the sound looses brightness, because the sample's frequencies are all lowered together; a real VA oscillator is always generating new harmonics
I beg to partially disagree.

Some, second-tier, VAs (most of the current VAs, in fact) are known NOT to have free-running oscillators.

(hence their thin sound, in my real-analogue-fanatic opinion)

Your sentence
aron wrote:on a real VA the oscillator always oscillates
should be corrected as follows:

a) on a REAL ANALOGUE the oscillator always oscillates [free running OSC],

(albeit it is a known fact that two VCOs under the same power source tend, in time, to sinc)

b) the same free-running is simulated in SOME very good VAs

So, the aural effect you correctly describe, and which depends on "digital" alignment of the two osc waves,

can happen both

a) on a rompler
b) on a cheap VA
c) even on some good VAs
d) and even on a extremely tight and heavily controlled contemporary Real Analogue.
RonF
Posts: 180
Joined: 17:48, 10 February 2010
Location: San Diego, CA USA

Re: Jupiter 80 VA engine - new or same as GAIA?

Post by RonF »

Great post ozy!!!
Mystic38
Posts: 1105
Joined: 14:04, 24 August 2009

Re: Jupiter 80 VA engine - new or same as GAIA?

Post by Mystic38 »

agreed..

i didnt want to go there..lol..

there is no such thing as a "real" VA.. fake is fake. period...how well you fake it is another matter.

many VA use "random start position" for their "oscillators"
RonF wrote:Great post ozy!!!
realtrance
Posts: 128
Joined: 16:39, 8 April 2005

Re: Jupiter 80 VA engine - new or same as GAIA?

Post by realtrance »

I always find the uses of the term "real" and "fake" as strange as the uses of the term "natural" and "artificial" when discussing electronic synthesis.

It really boils down to what you're after. There are artifacts introduced by digital synthesis that weren't there with analogue synthesis, some of which have in later times been embraced as features, rather than bugs (down-sampling, for instance).

If you're after perfect emulation, so perfect it's indistinguishable from the "real" thing..... it'll never happen. However, the search for such perfect EMUlation has been a serious driving force in the R&D of electronic synthesis, and continues to be. Detailed science on the details of "analogue" reality so that it can be emulated digitally continues to drive new discoveries. Those new discoveries become all the more fine-grained in their results. I consider SuperNATURAL synthesis to be one of these. It captures something in electronic synthesis that hadn't been captured previously.

I think the sustained concern over whether the virtual analogue portion of JP-80 is analogue enough really boils down to whether or not Roland has spent enough time furthering their pursuit of the science of digital emulation of analogue electronic synthesizers from the pre-digital age. I think there's a popular perception, whether valid or not, that Roland stopped paying attention to further advances on that front after the release of the JP-8000. They've since done a lot with COSM, etc. on V-Synth, and provided a lot more built-in variety of VA oscillator options, again, say, on V-Synth, than they had with JP-8000. But one marketing issue they haven't resolved is a blurring of the line between full-on mathematical simulation of analogue oscillator waveforms, and simply using samples of same as substitutes.

I think many of us feel using samples is cheating, and precisely because Roland has chosen _not_ to make the distinction clear, we end up being suspicious, and fearful that certain artifacts induced by sample-based oscillators will be present in instruments we use with the intention of having the best current science of analogue oscillator possible, to hand. This is something, in my opinion, I'd love Roland to clear up some day.

The catch is, if they _have_ been advancing their VA modelling technologies, it's probably in a way that is proprietary research, and the more they discuss that, or disclose it even to their own marketing folks, the more of a risk of revealing the details of those technologies to competitors they run. And Roland has, traditionally, been pretty closed-mouthed about going much into the technical details of what they're up to; they're mainly focused (some of us would say appropriately) on the musical results.

And it could be as simple as: sample-based electronic synthesis is far more worthy of investing in, in terms of R&D. It's far more possible to get a wider variety of more interesting results, for a variety of musical purposes, than it is by going down the build-a-better-VA-oscillator-and-non-steppy-digital-filter path. It could well be that more Roland R&D is concentrated on sampling than on digital VA modeling, because they've left that part of the market to others -- mainly in Europe, Access, Novation, Waldorf, in particular, in alphabetical order -- to inhabit, and don't consider it worth their while to compete in that area of R&D.
ozy
Posts: 169
Joined: 13:32, 1 December 2010

Re: Jupiter 80 VA engine - new or same as GAIA?

Post by ozy »

realtrance wrote:I always find the uses of the term "real" and "fake"
who used those terms?

I said "real" and "virtual",

and there is a MATERIAL, OBJECTIVE difference between a voltage-controlled real analogue and a computer+software based virtual analogue.

They generate sound in materially different ways.

It's not a matter of "convention" or "definition".

They are different, as much as a cello is different from a flute.

And, WHEN the "virtual analogue" is patently an attempt to emulate a real analogue, it can be judged on this merit.
RonF
Posts: 180
Joined: 17:48, 10 February 2010
Location: San Diego, CA USA

Re: Jupiter 80 VA engine - new or same as GAIA?

Post by RonF »

100% +1 to ozy's comments. Virtual Analog means ANYTHING goes.....some seek to emulate specific historic real world circuits and components....some seek to realize a circuit that has never existed....or is physically impossible or impractical. Some seek to get the best emulation with the least process load...while others seek to create the greatest realism or quality by pushing the outer limits of the process load. There IS NO such thing as a "real" or "proper" VA. Each is entirely developer specific....and therefore....naturally, the results vary, and then some are just better than others. Talking about, for example, the JP80 NOT being a "real" VA simply because it doesn't have a dedicated sync parameter is entirely missing the point. Its a VIRTUAL creation. There are no "rules". Nothing says it has to be a virtual RE-creation of anything, in fact. The developer might add or omit "features" and "parameters" at will, for any reason. The software development is just another form of art. What makes it special is its results. The JP80 VA sounds marvelous! What makes it different than a PCM or sample based machine.....is science.....not art. We should not confuse the two paradigms.
realtrance
Posts: 128
Joined: 16:39, 8 April 2005

Re: Jupiter 80 VA engine - new or same as GAIA?

Post by realtrance »

I never thought the term VA was all that useful to begin with; it initiated a war over which synth "most accurately modeled analogue" that will never die. Guess as a marketing ploy that's successful, but I never cared when judging any synth. How good it sounds to me, and how inspiring it is to play and sound design on, are the only criteria I ultimately care about.

Have most digital synths made since 1997 sounded great? Yes! Can any of them do "everything?" No! Would I want any of them to? No. It's that simple, for me at least. :)
RonF
Posts: 180
Joined: 17:48, 10 February 2010
Location: San Diego, CA USA

Re: Jupiter 80 VA engine - new or same as GAIA?

Post by RonF »

I agree with you. What is important in the VA moniker is the V.....Virtural. The "spin" and "marketing" is the A.....Analog. Analog is analog, and Digital is digital. A better description for this technology, IMO, is "Modeled". A software modeled synthesizer component is what separates this type of tech from PCM sampling. Its the continuous real time nuances that make it "special" (when well implemented, that is) as compared to the more static limitations of PCM tech, by its nature. So perhaps a better moniker might be VM (virtually modeled) synthesis. I guess the point of VA....is that the software virtually models the continuous real time nature of an analog component. How 'analog" it sounds in the end is very subjective and case specific. But it is a different approach to sound output than PCM, and perhaps that is what really matters.
User avatar
solitud
Posts: 25
Joined: 20:48, 24 September 2011
Location: Berlin

Re: Jupiter 80 VA engine - new or same as GAIA?

Post by solitud »

I think some of you may have a slightly elevated opinion of what constitutes a model.

x = time();
y = sin(x);

This model describes a perfect FREE FLOATING sinus oscillator. You can implement this on nearly every platform with very high polyphony (estimated > 300000).
Of course we don´t want the sinus to be perfect because this would not be "analog". There are many factors that influences the analog oscillator like the temperature of the system, resistance and tolerances in components.
Because it would be too expensive to consider all factors they are grouped as a variable influence constant OFFSET. The model now looks like this:

y = sin(x) + rnd(OFFSET);

Our marketing department can now distribute this model as H.Y.P.E.R. Realistic™ Synthesis.
aron
Posts: 88
Joined: 23:03, 11 September 2007

Re: Jupiter 80 VA engine - new or same as GAIA?

Post by aron »

I didn't write that post that was attributed to me. I was quoting someone else who was asking for a test. (Artemiy)
User avatar
solitud
Posts: 25
Joined: 20:48, 24 September 2011
Location: Berlin

Re: Jupiter 80 VA engine - new or same as GAIA?

Post by solitud »

aron wrote:I didn't write that post that was attributed to me. I was quoting someone else who was asking for a test. (Artemiy)
Hey Aron, the test is completely valid, I was just making a little fun not to be too serious about technical terms and marketing strategies ;)
ozy
Posts: 169
Joined: 13:32, 1 December 2010

Re: Jupiter 80 VA engine - new or same as GAIA?

Post by ozy »

realtrance wrote:the term VA initiated a war over which synth "most accurately modeled analogue" that will never die. Guess as a marketing ploy that's successful, but I never cared when judging any synth.
+1

I've got a "prophet 5 sync" replica from a "EMU 4" sampler: next to the original, it's thinner but fairly playable

The same sound, "remodeled" on the Novation KS VA, was bad and purposeless.

Would I ditch the prophet for the emu? No. Is the emu useful? yes.

Or: is the Korg Radias a "modeled analogue"? No way. Does it sound like a analogue? Not even remotely. Is it a bad synth? No.

Analogue has its virtues and exploits them. fat, warm, and infinitely nuanced. It has low polyphony, limited routings and limits on the upper frequencies' definition, which make it not suitable for acoustic emulations or some glassy, sharp, well-defined sounds (pads, pianos, pads, vocals). Filter above 36 db or formant filters become a luxury for few.

Digital should exploit ITS OWN virtues (almost unlimited filter flexibility,, waveforms galore, definition, polyphony)

and stop selling bad "prophet horns" and asthmatic "moog basses" to the masses who want to "gneeewww" and won't pay for it.
thunderkyss
Posts: 837
Joined: 07:38, 21 May 2003
Location: Port Arthur,TX

Re: Jupiter 80 VA engine - new or same as GAIA?

Post by thunderkyss »

RonF wrote:100% +1 to ozy's comments. Virtual Analog means ANYTHING goes.....some seek to emulate specific historic real world circuits and components....some seek to realize a circuit that has never existed....or is physically impossible or impractical. Some seek to get the best emulation with the least process load...while others seek to create the greatest realism or quality by pushing the outer limits of the process load. There IS NO such thing as a "real" or "proper" VA. Each is entirely developer specific....and therefore....naturally, the results vary, and then some are just better than others. Talking about, for example, the JP80 NOT being a "real" VA simply because it doesn't have a dedicated sync parameter is entirely missing the point. Its a VIRTUAL creation. There are no "rules". Nothing says it has to be a virtual RE-creation of anything, in fact. The developer might add or omit "features" and "parameters" at will, for any reason. The software development is just another form of art. What makes it special is its results. The JP80 VA sounds marvelous! What makes it different than a PCM or sample based machine.....is science.....not art. We should not confuse the two paradigms.
By this definition, the Fantoms are virtual analog
aron
Posts: 88
Joined: 23:03, 11 September 2007

Re: Jupiter 80 VA engine - new or same as GAIA?

Post by aron »

Straight from Roland site - Gaia:
Huge sound with three virtual analog engines onboard, each with a dedicated oscillator, filter, amplifier, envelope, and LFO.


Virtual analog synthesizer sound generator (Number of part: 1)
(Oscillator + Filter + Amp + Envelopes + LFO) x 3

Why is everyone saying this is not a VA when Roland said it is?

http://www.rolandus.com/products/produc ... arentId=83
Post Reply