Roland cloud Juno 106

[zerocrossing]

One thing possibly overlooked, on the Roland plugins, is some of the controls are bi-directional. Filter LFO, ENV-1 and KYBD etc go from -128 to 128, and you have to ignore anything below 0 to stay within hardware range.

Things like high / maximum resonance vary between hardware, and Roland’s approach does seem more towards ideal calibration specs rather than any individual unit. But differences shouldn’t be huge..

If anyone can name a specific Factory preset (Edit: I see A-72 and A-46 were mentioned) )
A-72 aka "6-4 1984 Pluck Sweep" - Roland requires you turn the volume up towards the mid 50's as it starts quieter than Softube. It's also duller - turn the VCA Tone control up a bit to compensate. With those tweaked VCF Res (set a bit higher on Roland) and Kybd adjustments should definitely get you fairly ballpark. Softube's used a slightly different shaped KYBD curve which, if you match the upper keys, means Softube's tends to let a bit more bass through on the lower notes. The release holds a little more on Softube's too, so you'd have to raise the release on Rolands slightly to try to match.

Based on A-72 I'd still say you can get 'em pretty close and where neither is vastly better. That said I prefer Softube's, on that one, slightly

I think a lot of people try and match settings on emulations and when they hear the differences, decide that something is either a good or a bad emulation, but that’s often a mistake. When I’m matching sounds, the first thing I do is make sure levels and filter settings are as identical as I can get them. I use cranked resonance to and a spectrum analyzer to make sure I have things as tuned as I can get them. Often I find that the differences that seemed huge end up being tiny or nonexistent.

I also don’t totally care about accuracy. It’s not like I’m going to buy a Juno 106. Been there, done that. To me, there’s nothing interesting about a vintage synth beyond some nostalgia. Something like Roland’s or Softube’s software is perfect for me when I want a little bit of that classic sound but also something that goes beyond. Frankly, neither go far enough, IMO, but they’re still useful. Not every part requires a modulation monster. As long as a synth sounds good to me, I’m in, regardless as to how accurate it is.

[PAK]

I think a lot of people try and match settings on emulations and when they hear the differences, decide that something is either a good or a bad emulation

You're 100% correct about that. Though emulations do tend towards a simpler, more ideal (and easier to implement) side of expectations, versus the real world.

EG Tracking variations across a keyboard. They aren’t unusual when you start involving old hardware and voltages. If you've got a sound with differences (commonly bass or high end) between one end and the other, with no way to fully adjust a keyboard curve to replicate, perhaps it's not hugely unreasonable for someone to conclude this is due to some sort of fundamental emulation failing, rather than simple reasons and a dev with maybe too much focus on specific aspects rather than the overall picture.

I also don’t totally care about accuracy. It’s not like I’m going to buy a Juno 106.

I lean more towards that side of things too, though I’m glad Softube's there doing their approach - even if I think they could maybe do the whole free-running (Vs per voice CPU management) thing better for slower systems. For some people nailing a 106 is important. There's an entire genre of music (Spacesynth) where the 106 Bass is the equivalent of the Supersaw in trance music.

Frankly, neither go far enough, IMO, but they’re still useful. Not every part requires a modulation monster.

I think (beyond their role in music) many of the “classics” maintained popularity because there’s a good balance between interface complexity and the sounds they offer. A "true" emulation would have to offer all calibration possibilities - which would also translate to more flexibility in sound design too. But I also accept most users don't want to deal with too many options, and it would require a considered implementation. But it does mean there's still room to improve on what's already been done - if anyone was ever inclined to do it

[Arashi]

Based on A-72 I'd still say you can get 'em pretty close and where neither is vastly better. That said I prefer Softube's, on that one, slightly

Yep, that's what I found too. The Softube one has a touch of extra texture in the resonance (it's subtle, but nice), but they are otherwise very close. A couple pages back, I posted a tweaked version of the Roland preset that matches the Softube version as closely as I could manage.

And then there's A46 Dark Pluck (aka 1984 Dark Pluck), where the Softube and Roland emulations are basically bang-on after volume matching (and lowering the cutoff ever so slightly on the Roland version), while DCO-106 is thin and weak (regardless of volume), because they didn't model the Juno-106 bass boost when the HPF is at 0, and there's nothing you can do to make it sound right.

[zerocrossing]

I think a lot of people try and match settings on emulations and when they hear the differences, decide that something is either a good or a bad emulation

You're 100% correct about that. Though emulations do tend towards a simpler, more ideal (and easier to implement) side of expectations, versus the real world.

EG Tracking variations across a keyboard. They aren’t unusual when you start involving old hardware and voltages. If you've got a sound with differences (commonly bass or high end) between one end and the other, with no way to fully adjust a keyboard curve to replicate, perhaps it's not hugely unreasonable for someone to conclude this is due to some sort of fundamental emulation failing, rather than simple reasons and a dev with maybe too much focus on specific aspects rather than the overall picture.

I also don’t totally care about accuracy. It’s not like I’m going to buy a Juno 106.

I lean more towards that side of things too, though I’m glad Softube's there doing their approach - even if I think they could maybe do the whole free-running (Vs per voice CPU management) thing better for slower systems. For some people nailing a 106 is important. There's an entire genre of music (Spacesynth) where the 106 Bass is the equivalent of the Supersaw in trance music.

Oh, I know. I often will use emulations like the 106 when I’m writing something for a game that is in a style that needs it and a lot of complexity or expression is not necessary.

Frankly, neither go far enough, IMO, but they’re still useful. Not every part requires a modulation monster.

I think (beyond their role in music) many of the “classics” maintained popularity because there’s a good balance between interface complexity and the sounds they offer. A "true" emulation would have to offer all calibration possibilities - which would also translate to more flexibility in sound design too. But I also accept most users don't want to deal with too many options, and it would require a considered implementation. But it does mean there's still room to improve on what's already been done - if anyone was ever inclined to do it

Usually when I say something hasn’t gone “far enough,” I’m talking about things like the Roland Cloud synths that don’t support things like aftertouch or MPE. I can add my own complexity in the performance if I have enough hooks to latch onto with expression. This is where TAL is great, though I don’t own their synths yet, just the Sampler.