Upcoming Synapse OB-Xa: Obsession

[perfumer]

I would muchly prefer discussions on a subjective level ("I prefer these over those") than turning comparisons into death matches ("Synth/Dev X is better than Y").

I was eternally grateful to Roland that they made analogue emulations which used a hell of a lot more CPU than ours. Of course I praised them

Makes sense!

[Ingonator]

15 years ago i owned a real Jupiter-8 and later used all available emulations where i currently own Roland Cloud Jupiter-8, Arturia Jup-8V and Diva.

While Diva is great on it's own (with lots of options also due to the differnt synth modules) and i also like the Arturia plugin the Roland Cloud Jupiter-8 is my favorite emulation so far and it has a few advanced features compared to the real thing where my favorite is having an identical set of waveforms for both Oscs.

CPU use of the Roland plugin is quite high but with the latest version it is much lower than with the very first version that was released. In the OPTIONS menu there is also a feature for using less CPU which results in around 40% less CPU use.
Besides that reducing the voice limit /default is also reduces the CPU use of the plugin. The CPU use is not independent from the voice limit even if only a single voice is played.

After having used the Roland Cloud subscription for more tha 12 months i had cancelled the subscription and was allowed to keep 2 plugins for an unlimited time. Those were the Jupiter-8 and System-8 plugins. FWW with one of the updates the System-8 plugin now also has a Jupiter-8 and a Juno-106 filter model, each with 2, 3 or 4 poles.

[Cinebient]

Like Funkybot´s Evil Twin says...
"But you'd have to almost have to post examples because my issue isn't with any one static setting, but rather the sum of the whole. How the oscs, interact with and drive the filters, how the filters interacts with the envelope, how all that interacts with the amp."
This is what i care about as well. Here i really doubt any "exact emulation" will bring anything new here really compared to what i already have. Because some of my favorites here already cover several filters from this area in one plug-in. Having the same, no actually even a much bigger range of sweetspot actually. F.e. my favorite Minimoog emulation is actually no Minimoog emulation and i do not care if its a 100% or 95% accurate emulation. Well, it might even be a 250% better emulation than the original

[PAK]

There's no weight at all to the sound of the Roland model. There's no chunk.

Really meaningful descriptions, these. I know you think you're communicating something, but I have no idea what you actually mean. Are you saying the bass gets cut off? Lack of actual low end? Anything that could be, you know, actually measured?

Hmm. Well, I kinda agree with his opinion, so I'd say this..

Part of the reason for differences is voice tuning is more perfect on Roland's emulation than you'll generally find on any hardware examples, and the "condition knob" it provides is a silly joke which does nothing to fix that. Whereas Diva provides a proper basic ability to tune the voices, where the only real control Roland provide is the VCO fine tune. However it's on the user to tune, and guess what? Most don't and won't. Would you like to describe how you measure this (easily audible) difference on a polyphonic sound? If everything could be broken down into "and this is how you measure it" the process of emulation coding would be a lot easier..

In terms of other factors - Since the earliest days, a consistent way I can tell virtual analogue apart from hardware is it fails to emulate the influence of the VCA on the sound. If you listen to enough Jupiter comparisons with the original hardware, even on Youtube, you may eventually notice that, on some sounds, the perception of volume is consisently "softer" on the versions running software. That means both plugins and the Boutique / System 8 (Meaning it's not just input chain and/or converters, there's something different about how the analogue synth is outputting the sound.. )

This should translate to a slightly lower volume, at the same measured peak, and/or reduced perception of transients on the attack portion which may, or may not, also be accompanied by less perception of certain frequencies - not necessarily always low end. I say perception because, whilst these differences can likely be measured, sometimes other factors change what is heard in a way not reflected by measurement alone.

EG Someone posted an interesting test of DCO Vs VCO on their DSI hardware to Youtube, which was basically about how the voltage/frequency variation on the VCO might influence user perception of "less harsh" higher frequencies, which they then emulated on the mod matrix of the DSI DCO synth to give the DCO the same VCO-like quality.

BTW If someone knows you're listening for the level differences then it's easy to mislead in either direction. But it doesn't mean it's not there Unless a virtual instrument goes out of its way to do anything with the VCA code (and Diva does, to a degree, where Roland doesn't) then you might get perceptions like "thin", or "no weight" being mentioned etc. This is likely part of the source of his perception difference between Diva and Roland's plugin, and why you might favour Diva. But to do that you'd first have to know the difference between it and the output volume. I doubt most users get that far.

Even with the one big thing going for Roland's version, which is it's easier to dial in to the hardware ranges then (filter aside) you're limited to a lowly 7 bit resolution, and it tends to only be 8 bit when they've added a minus parameter (-127 through +127 etc). 8 bit resolution - which is very low for a VST - and, if bi-directional controls were added, any parameter has its effective resolution of that control reduced to 7 bit (because it uses 128 in each direction for a total of 8 bits.)

For that you're also stuck with something which is Roland's idea of timings and tunings, which is unlikely to tie in with the variations found in the real world instrument which, btw, has a ton of calibration control. The whole calibration thing is arguably even more relevant to Oberheims, and I know that Synapse knows that. So it will be interesting to see where and how they proceed there.

If I turn off the FX I'd say I can get Diva to do what Roland's plugin does very well. And whenever I made attempts, at recreating a fairly well known Jupiter 8 sound, (to me) Diva tends to win quite easily. EG Which d'you think will do a subjectively better "Crockett's Theme" pad with no help from external FX? A lot of this is simply to do with the fact that, overall, Diva gives more control over the sound. Any Oberheim emulation would face the same competition from Diva. EG Even though Diva lacks some of OP-X's voice tune features, overall it'll generally produce something closer to the hardware if tweaked correctly.

And did you try the thing that I said? Velocity sensitivity is crucial in the original JP-8, because it allows you to extend the modulation ranges for envelopes.

I don't know if he did, but I'm not noticing any big difference. If you're sure can you please post presets of whatever you're doing on Diva Vs Roland so I can take a look? If it's over 14-15kHz then maybe I'm just not hearing it, but I'd like to check anyway The only setting which makes any substantial difference here is Roland's VCA Tone control. Or, as it might be better known elsewhere, an EQ knob AKA the same basic optional trick Arturia offered to reduce differences too ..

[Echoes in the Attic]

I was eternally grateful to Roland that they made analogue emulations which used a hell of a lot more CPU than ours. Of course I praised them - But I think I only ever checked out the SH-2 and the System 100. Never tried the ones discussed here.

To be fair to Roland (as much as I hate their Cloud), the cpu is now actually quite low, Significantly lower than Diva for polyphonic patches actually.

[PAK]

To be fair to Roland the cpu is now actually quite low, Significantly lower than Diva for polyphonic patches actually.

Using Divine mode? Great is just fine in the vast majority of scenarios, and even "fast" is ok for a lot of stuff..

Also, no, low, or high resonance? Diva intelligently reduces CPU load based on resonance usage. That's why blanket statements simply can't apply.

The lower quality mode on Roland's plugin is pretty unusable (IMO)

[Echoes in the Attic]

There's no weight at all to the sound of the Roland model. There's no chunk.

Really meaningful descriptions, these. I know you think you're communicating something, but I have no idea what you actually mean. Are you saying the bass gets cut off? Lack of actual low end? Anything that could be, you know, actually measured?

Hmm. Well, I kinda agree with his opinion, so I'd say this..

Part of the reason for differences is voice tuning is more perfect on Roland's emulation than you'll generally find on any hardware examples, and the "condition knob" it provides is a silly joke which does nothing to fix that. Whereas Diva provides a proper basic ability to tune the voices, where the only real control Roland provide is the VCO fine tune. However it's on the user to tune, and guess what? Most don't and won't. Would you like to describe how you measure this (easily audible) difference on a polyphonic sound? If everything could be broken down into "and this is how you measure it" the process of emulation coding would be a lot easier..

In terms of other factors - Since the earliest days, a consistent way I can tell virtual analogue apart from hardware is it fails to emulate the influence of the VCA on the sound. If you listen to enough Jupiter comparisons with the original hardware, even on Youtube, you may eventually notice that, on some sounds, the perception of volume is consisently "softer" on the versions running software. That means both plugins and the Boutique / System 8 (Meaning it's not just input chain and/or converters, there's something different about how the analogue synth is outputting the sound.. )

This should translate to a slightly lower volume, at the same measured peak, and/or reduced perception of transients on the attack portion which may, or may not, also be accompanied by less perception of certain frequencies - not necessarily always low end. I say perception because, whilst these differences can likely be measured, sometimes other factors change what is heard in a way not reflected by measurement alone.

EG Someone posted an interesting test of DCO Vs VCO on their DSI hardware to Youtube, which was basically about how the voltage/frequency variation on the VCO might influence user perception of "less harsh" higher frequencies, which they then emulated on the mod matrix of the DSI DCO synth to give the DCO the same VCO-like quality.

BTW If someone knows you're listening for the level differences then it's easy to mislead in either direction. But it doesn't mean it's not there Unless a virtual instrument goes out of its way to do anything with the VCA code (and Diva does, to a degree, where Roland doesn't) then you might get perceptions like "thin", or "no weight" being mentioned etc. This is likely part of the source of his perception difference between Diva and Roland's plugin, and why you might favour Diva. But to do that you'd first have to know the difference between it and the output volume. I doubt most users get that far.

Even with the one big thing going for Roland's version, which is it's easier to dial in to the hardware ranges then (filter aside) you're limited to a lowly 7 bit resolution, and it tends to only be 8 bit when they've added a minus parameter (-127 through +127 etc). For that you're also stuck with something which is Roland's idea of timings and tunings, which is unlikely to tie in with the variations found in the real world instrument which, btw, has a ton of calibration control. The whole calibration thing is arguably even more relevant to Oberheims, and I know that Synapse knows that. So it will be interesting to see where and how they proceed there.

If I turn off the FX I'd say I can get Diva to do what Roland's plugin does very well. And whenever I made attempts, at recreating a fairly well known Jupiter 8 sound, (to me) Diva tends to win quite easily. EG Which d'you think will do a subjectively better "Crockett's Theme" pad with no help from external FX? A lot of this is simply to do with the fact that, overall, Diva gives more control over the sound. Any Oberheim emulation would face the same competition from Diva. EG Even though Diva lacks some of OP-X's voice tune features, overall it'll generally produce something closer to the hardware if tweaked correctly.

And did you try the thing that I said? Velocity sensitivity is crucial in the original JP-8, because it allows you to extend the modulation ranges for envelopes.

I don't know if he did, but I'm not noticing any big difference. If you're sure can you please post presets of whatever you're doing on Diva Vs Roland so I can take a look? If it's over 14-15kHz then maybe I'm just not hearing it, but I'd like to check anyway The only setting which makes any substantial difference here is Roland's VCA Tone control. Or, as it might be better known elsewhere, an EQ knob AKA the same basic optional trick Arturia offered to reduce differences too ..

I can understand why some might prefer the Roland version over Diva for a Jupiter sounds since out of the box, the Roland version has more of that shimmery top end that the jupiter did. So for brighter sounds like stuff with sync or just an open filter, I think cloud version sounds great and it's pretty mix ready. But I also agree that Diva has more "weight" to it, as is the case if you compare it to most synths actually. Comparing the two will usually lead someone to see Diva as being better for bass and those heavier sounds but is also darker. Funny thing though, if you throw a high shelf EQ on Diva and boost from around 8kHz on wards by 5db or so (adjust both to taste), it adds so much of that sheen that I personally (and others I've read) found missing when comparing to analogs or even other emulations. And it works on most sounds. Of course a simple coarse EQ bump like that may bring out some harmonics that are not desired sometimes so it may not work, but I found it to be so much better that my default preset in my host has an EQ in the chain so it loads by default. I don't know why they made it so dark but it's like a whole new synth when you bump up that top end.

Even after that tweak I still think the Roland version has some aspects to the sound that is quite pleasing in the higher harmonics that works well for some sounds even better than Diva, but the reverse is probably more common.

[Echoes in the Attic]

To be fair to Roland the cpu is now actually quite low, Significantly lower than Diva for polyphonic patches actually.

Using Divine mode? Great is just fine in the vast majority of scenarios, and even "fast" is ok for a lot of stuff..

Also, no, low, or high resonance? Diva intelligently reduces CPU load based on resonance usage. That's why blanket statements simply can't apply.

The lower quality mode on Roland's plugin is pretty unusable (IMO)

If you mean the "Optimize for lower cpu usage", no not with that. And no not Divine mode in Diva, just Great. Talking about low or moderate resonance usually. Your typical polysynth and pads sounds, things where you'd get a buildup of voices. The latest Diva update did help narrow the gap to be closer to what the Roland Cloud cpu level.

[Echoes in the Attic]

Hey I started a new thread to bring the off topic OB-Xa talk out of the Legend thread and now here we are talking about Jupiter comparisons? Wha happened??? Let get back on topic!

[Ingonator]

Hey I started a new thread to bring the off topic OB-Xa talk out of the Legend thread and now here we are talking about Jupiter comparisons? Wha happened??? Let get back on topic!

Well i am currently sitting here doing patches with the Beta of the OB-Xa plugin...

[Echoes in the Attic]

Hey I started a new thread to bring the off topic OB-Xa talk out of the Legend thread and now here we are talking about Jupiter comparisons? Wha happened??? Let get back on topic!

Well i am currently sitting here doing patches with the Beta of the OB-Xa plugin...

I'm looking forward to joining the party! (any minute now....)

By the way, I still use your Poly-Ana GUI. Even though it is very cartoonish, I like the Valhalla type aesthetic of it. So much easier than any other skins. Too bad the dev doesn't include it on the site. That synth still stands up amazingly to all these newer emulations too...

[PAK]

the Roland version has more of that shimmery top end that the jupiter did.

Can you point me to a specific Roland preset (must be a specific preset, even your own) where you feel this is true? IMO a lot of the incorrect perception of this stems from the fact that you HAVE to use the analog2 shape to produce equivalent brightness with saw sounds, otherwise it will sound more dull. Also pilot error because there are multiple factors on Diva which influence various things.

Some people might also use Roland's VCA Tone control, not understanding it's (effectively) an EQ boost. Thus Diva would have to provide an EQ to replicate that (the closest it gets is the HPF:Bite module)

I don't hear any big difference re:shimmery top. BTW, I *do* hear a difference with TAL Uno, where I like how it handles the upper frequencies..

Hey I started a new thread to bring the off topic OB-Xa talk out of the Legend thread and now here we are talking about Jupiter comparisons? Wha happened?

Haha. Though I think maybe it's more on topic than you might think - Diva will also cover a lot of Oberheim ground too

[Cinebient]

Hey I started a new thread to bring the off topic OB-Xa talk out of the Legend thread and now here we are talking about Jupiter comparisons? Wha happened??? Let get back on topic!

Are you new at KVR?

[Ingonator]

At the moment i am using the OB-Xa plugin together with Arturia Matrix-12 V2 (with a 2nd keyboard controller) so i am now close to an "Oberheim overdose"...

For the Matrix-12 V2 i am using the Waves OneKnob Phatter plugin to increase the low end which is not necessary with the OB-Xa plugin.

While the Arturia plugin is nice a Matrix-12 plugin by Synapse Audio or U-He is still one of my "wet dreams"...

[chk071]

For the Matrix-12 V2 i am using the Waves OneKnob Phatter plugin to increase the low end which is not necessary with the OB-Xa plugin.

Dunno why the Arturia plugs always suffer from a lack of bottom end... I can definitely confirm that from their other plugins.