Virtual Vintage Analog

[Al Coda]

4 hours ago, Tusker said:

It's just that even in my not-new 8 core Intel I9 MacBook Pro with 32 Gigs of Ram,

 

I´m on a Intel i7 2.5GHz quad core, 6MB L3 cache w/ 32GB RAM laptop ...

My main DAW rackmount is Intel i7 4.0Ghz quad core, 8MB L3 cache w/ 32GB RAM too.

Both are gen6 i7 processors.

It´s "old" already, I know,- and I don´t complain.

All I said is, the best plugins are CPU hungry,- always,- on Intel or AMD CPUs or DSP chips !

And it won´t change in future because faster machines allow developers coding more accurate component modelling emulations,- and they will do.

 

And as I said in former posts,- I´ll have a look for Mac M1 or M2, possibly for a "X" model too,- WHEN Native Instruments will be native M silicone ready.

 

But because of interest:

 

AFAIK, Uhe plugins are core locked.

So, what speed (Ghz) is your Intel i9 proc and how´s the CPU load (%) of Repro 5 when you load just only this plugin w/ a 6-voice poly patch ?

(I like to have at least 1 voice more than a original Prophet-5 offered)

I mentioned Repro and JP-8 because they are really good, but I didn´t buy because on my i7 processors, each consumates up to 50% CPU on a single core while most other plugins are average under 10% CPU,- Reaper as the host and w/ Thomas Mundt´s "LoudMax" master limiter inserted in the main inserts included.

 

☺️

 

A.C.

[Mike Martin]

Yesterday, I got out my good 'ol Korg Polysix which I've owned since 1984 and compared it to the virtual Korg PolySix software. There is no question that the real-deal was more FUN to play. The immediacy of the controls is just fun and since I've owned that instrument for 38 years I know exactly how to get every sound it can make out of it. 

 

The virtual Polysix, honestly was quite good. In fact I did a side by side comparison recording of the two of them and two people guessed wrong about which instrument was with - I left the "running noise" of my real PolySix in the background of the virtual tracks and people couldn't tell the difference. The UNISON mode is the one place where I noticed a difference by default but with a little tweaking of detune amounts in the software it was even closer.  I'd expect you'd find that much variation between two different real analog models. 

Playing the virtual instrument - without having dedicated knobs every parameter, just isn't as inspiring in comparison. I don't see myself getting carried away for hours on the plug-in as I might on the real thing but I know that if I need something specific it can probably do it and it won't have the noise of the original. 

So I'm not rushing out to buy new or used analog. The nearly endless supply of virtual instruments keeps me mostly happy. What I want is a control surface that makes using these software tools more engaging. 

 

[Tusker]

9 hours ago, Al Coda said:

I´ll have a look for Mac M1 or M2, possibly for a "X" model too,- WHEN Native Instruments will be native M silicone ready.

Not to add to anybody's gas but I was in the Apple store today and those Mac Studios have M1Max processors and RAM up to 64 gigs . They are cheaper than the MacBook pros and they look like they can be racked. I can't think of a better performance to price option.

 

9 hours ago, Al Coda said:

So, what speed (Ghz) is your Intel i9 proc and how´s the CPU load (%) of Repro 5 when you load just only this plugin w/ a 6-voice poly patch ?

The speed is 2.3 GHz.

 

With your requested six voices and the multicore (MCORE) on, the CPU was just below 50%. This is Repro 5 with one set of six voices:

 

 

So I thought, that is interesting. It still looks like one core is doing all the work. So I duplicated the track five times and I noticed that other cores were sharing the burden. This is Repro 5 with six sets of six voices (36 voices):

 

 

Finally, I pushed it a bit further by duplicating tracks until there were 11 tracks. At this point the I was getting close to 50% CPU so I decided to stop. This is Repro 5 with 11 sets of six voices each (66 voices):

 

 

Hope this helps Al. Best,

 

Jerry

 

[AROIOS]

10 hours ago, Mike Martin said:

Yesterday, I got out my good 'ol Korg Polysix which I've owned since 1984 and compared it to the virtual Korg PolySix software. There is no question that the real-deal was more FUN to play. The immediacy of the controls is just fun and since I've owned that instrument for 38 years I know exactly how to get every sound it can make out of it. 

 

The virtual Polysix, honestly was quite good. In fact I did a side by side comparison recording of the two of them and two people guessed wrong about which instrument was with - I left the "running noise" of my real PolySix in the background of the virtual tracks and people couldn't tell the difference...

And that's 18 year old code that ran smoothly on Pentium III/IV CPUs, not to mention its ported version that even the puny Atom CPU in KRONOS can handle without breaking a sweat.

As Tusker mentioned, plugin programmers can get a lot of mileage out of current computing hardware. It's a matter of allocating computing power wisely. This is also evident in the performance Virus managed to pull out of the Motorola 56300 chips.

[Al Coda]

7 hours ago, Tusker said:

Not to add to anybody's gas but I was in the Apple store today and those Mac Studios have M1Max processors and RAM up to 64 gigs . They are cheaper than the MacBook pros and they look like they can be racked. I can't think of a better performance to price option.

 

I googled and investigated a bit more in the "Max" and "Ultra" versions.

Very promising and I didn´t expect these being so compact

 

7 hours ago, Tusker said:

The speed is 2.3 GHz.

 

With your requested six voices and the multicore (MCORE) on, the CPU was just below 50%. This is Repro 5 with one set of six voices:

 

Now I´m surprised about your results being so close to mine.

We have Intel i7 2.5 GHz vs Intel i9 2.3 GHz and almost identical results w/ Repro 6 voices on a single core.

I said "up to 50%" on my "ancient" Lenovo W540 laptop (Win10 Pro, 16 GB RAM) and precisely it fluctuates between 45 and 48 %.

On my i7 rackmount DAW,- 4.0GHz and 32GB RAM,- it´s better.

In fact, I expected more performance boost from a Intel i9 vs 6th gen Intel i7.

 

The 4 GHz i7 is a 4790K stock speed,- a "Haswell",- and the i7 in my laptop is a "Haswell" too.

I bought these because they were confirmed to run well for audio and especially also for S|C Scope/XITE.

 

7 hours ago, Tusker said:

So I thought, that is interesting. It still looks like one core is doing all the work. So I duplicated the track five times and I noticed that other cores were sharing the burden. This is Repro 5 with six sets of six voices (36 voices):

 

That confirms Uhe plugins are core locked.

Multi core operation becomes active not before multiple instances are in use.

 

Thx a lot for the test and all your effort !

 

But now there´s a lot of GAS for the new Mac Studio Max here.

 

☺️

 

A.C.

 

P.S.:

 

In my post above I mentioned 32GB of RAM for my laptop,- but that was wrong,- 16GB RAM only.

[Tusker]

2 hours ago, Al Coda said:

Now I´m surprised about your results being so close to mine.

We have Intel i7 2.5 GHz vs Intel i9 2.3 GHz and almost identical results w/ Repro 6 voices on a single core.

I said "up to 50%" on my "ancient" Lenovo W540 laptop (Win10 Pro, 16 GB RAM) and precisely it fluctuates between 45 and 48 %.

 

Yes, that's a great point! I was curious how much "better" the I9 would be. It's just an incremental difference. This laptop is great for me though, probably till Zebra 3 shows up. I mostly rely on Zebra HZ which brings a lot more interesting sound-design than Repro to the table, while not wasting power on vintage faithfulness. In Zebra, the comb filters, oscillator effects, resonators, DX-FM, msegs and modulation maps are really delicious. Maybe I just like the modern sound? I do like Repro's effects though. They are buttery smooth, like a good analog output stage.

 

As I mentioned, 6-10 Zebra instances and three dozen orchestral library instances is perfectly available on this I9. I dare say an I7 would be similar...

 

Wouldn't it be neat (hint) to see this same test performed on a silicon Mac for comparison?     💡😎

 

[CyberGene]

12 minutes ago, Tusker said:

Wouldn't it be neat (hint) to see this same test performed on a silicon Mac for comparison?     💡😎

I can do that however I just sold my MacBook Air M1 and am currently waiting for a heavily back-ordered MacBook Air M2 to arrive which could take up to 3 months 😕 But there are certainly many other people on the forum who can test it and I'll be more than interested to see a comparison between it and an Intel-based Mac.

 

I use a maxed-out company-provided MacBook Pro 16" with M1 Pro but I'm not sure if I am allowed to install personal-related software on it, so I will have to check that first.

[Stokely]

Agreed that Repro's effects are top notch, and could probably make a very nice fx suite on their own.   Not that Zebra's or Diva's are terrible,  but IMO Repro's are a step up.

I get worse performance when I use Multicore in Repro, I think I read this may be due to the fact that Logic is also doing things multi-core and there is a some kind of conflict?  In any event, I am able to use a fairly large number of tracks with my 2016 MBP before running into problems so I don't worry too much about it.   I'll freeze tracks if needed and ultimately render them out as audio if I want to do a serious mix (removes the urge to tweak performances, I have to kind of impose these kinds of restrictions on myself or nothing will ever be "finished"!)

I might go ahead and pick up Zebra HZ at some point.  All my u-he stuff I got during one of their rare sales and it's tough for me to buy things not on sale!   The flip side is that I've bought a lot of stuff because it was a great bargain, but I end up going back to a few of my favorites all the time, u-he's being among them.

[Anderton]

23 hours ago, Mike Martin said:

 There is no question that the real-deal was more FUN to play. The immediacy of the controls is just fun and since I've owned that instrument for 38 years I know exactly how to get every sound it can make out of it. 

 

I can't help but wonder if that's why rack units never sold as well as companies expected.

 

BTW, don't forget to check the Polysix battery periodically! I caught mine just before it starting leaking past the battery. Talk about a close call...

 

Finally:

 

23 hours ago, Mike Martin said:

What I want is a control surface that makes using these software tools more engaging. 

 

When MIDI 2.0 profiles get nailed down, a MIDI 2.0 controller would be able to talk with whatever keyboard you wanted, and you wouldn't have to relearn assignments for all your instruments. Seems like maybe it's a product opportunity for...oh, I dunno...maybe some successful Japanese company that makes keyboards. Whoever did that cool XW-P1 might be a good choice

[ProfD]

32 minutes ago, Anderton said:

I can't help but wonder if that's why rack units never sold as well as companies expected.

I think both the lack of immediate access to controls/programming and increased polyphony in KBs stifled rack unit sells and ultimately led to their demise. 

 

Back in the day, when a new KB was released, it was almost automatic that a rack unit would follow. 

 

Nowadays, they make KBs small and cheap enough that it is rack unit with keys attached.😁😎

[Tusker]

I'm quite excited by MIDI 2.0, the Espresso touchscreen display, NI's NKS standard, Bitwig's CLAP standard and the Expressive E Osmose. They suggest that musical interfaces will continue to improve for keyboardists. None of these will feel as natural as sitting at an instrument I have spent years with. How could they? But it seems we will continue to have more and more expressive choices, whether we want a simple rig or a complicated one. There's also a pile of relatively low cost analog. It's a great to time to be a keyboardist!

[Al Coda]

20 hours ago, Tusker said:

Wouldn't it be neat (hint) to see this same test performed on a silicon Mac for comparison?

 

 

 

A.C.

[ElmerJFudd]

 

 

https://www.perfectcircuit.com/catalogsearch/result/?q=soundforce software controller
 

 

[Mike Martin]

On 8/23/2022 at 2:44 PM, Anderton said:

I can't help but wonder if that's why rack units never sold as well as companies expected.

 

BTW, don't forget to check the Polysix battery periodically! I caught mine just before it starting leaking past the battery. Talk about a close call...

 

Mostly true about rack units but no one has been making those for quite a while. 

 

My original Polysix battery (the kind that leaked) was replaced long ago when I had a MIDI kit added. The lithium battery that is in there is currently dead and needs replaced.

 

On 8/23/2022 at 2:44 PM, Anderton said:

When MIDI 2.0 profiles get nailed down, a MIDI 2.0 controller would be able to talk with whatever keyboard you wanted, and you wouldn't have to relearn assignments for all your instruments. Seems like maybe it's a product opportunity for...oh, I dunno...maybe some successful Japanese company that makes keyboards. Whoever did that cool XW-P1 might be a good choice

 

Casio was very early with MIDI 1.0, I hope that we'll see some 2.0 features soon.

[Tusker]

On 8/22/2022 at 2:07 PM, Al Coda said:

So, what speed (Ghz) is your Intel i9 proc and how´s the CPU load (%) of Repro 5 when you load just only this plugin w/ a 6-voice poly patch ?

(I like to have at least 1 voice more than a original Prophet-5 offered)

 

It looks like Woody has been having the same CPU-Load questions we've had.  He's done a very interesting test with several of the leading synth plugins ...

 

Sylenth1 

Diva

Kontakt

DCO-106

Serum

Massive

Massive X

 

and provided conclusions and  graphs (at about the 12th minute). 👍👍

 

 

 

 

 

 

 

 

 

 

[AnotherScott]

 

On 8/22/2022 at 3:40 PM, Mike Martin said:

Yesterday, I got out my good 'ol Korg Polysix...Playing the virtual instrument - without having dedicated knobs every parameter, just isn't as inspiring in comparison.

 

Maybe a well-placed 12.9" iPad could scratch that itch? 

 

 

 

[Theo Verelst]

The load on your Intel Cpu is affected by a variety of factors, even by how your Dsp program has been compiled. Depending on version, the load can vary, based on how you set up the performance influencing parameters of your Cpu and operating system. If you have some sort of Mobo accompanying control window, for the best result it should be set to governing mode "performance" instead of efficiency or such.

 

This should make your Cpu hotter, and onky if it in thus setting becomes significantly warm  like 60 degrees celcius for instance, your Cpu is actually performing work a significant portion if the time. Governers on the latest I's influence a lot of things, including bus speed, memory speed and how many cpu are actually non-sleeping. Accelerators like Sse and others are important and the number of them per core and their clock speed can influence heavy computations a lot!

 

You would like a tool to show actual processor clock speeds while looking at tjose Cpu use graphs, which can be normalized per processor or thread different than what you might think. I don't recall proper windows tools for that in general.

 

From a Dsp system programmer's point of view, it should mske cores hotter if you set a somewhat longer buffer for audio compute purposes, because Cpu architecture is sensitive to pipelining.

 

T

[David Emm]

On 8/21/2022 at 9:26 AM, CyberGene said:

Get an Apple Silicon equipped Mac and you would be able to stack as many instances of U-He plugins as you wish and play them in real-time without a single glitch. 
 

Here’s an arrangement of an excerpt from Bruckner’s Eighth Symphony I made entirely with U-He Diva in Logic Pro. I used close to 30 instance of Diva for each orchestral instrument there (even string section was made of multiple instances) and it could play in real-time with the maximum selected quality:

 

 

 Damn, I hereby vote for a "Switched-On CyberGene" album. I'll even pre-order! I know its a hellacious amount of work, but I'd forgotten how I bolted down chunks of Carlos and Tomita like a hungry wolf goes for a beef brisket. Using synth for classical works has to walk a thin line between enrapturing and unbearably cheesy, but it hits a welcome spot for me. I'm one of those weirdos who whistles bits of "The Rite of Spring" in public. I fondly recall the day a woman turned, pointed at me and said "Stravinsky!" I didn't have the heart to tell her I wasn't him 'cause he had long since snuffed it. 💀     

[Konnector]

Ghz speed makes a performance difference for VST and VSTi plugins from what I've always understood.  I would consider 2.3Ghz quite slow. You can buy custom laptops these days that run @ 5Ghz. Of course a rackmount/desktop PC will smoke a laptop, but it involves a few extra bits of hardware if you're using one live. (keyboard, mouse, monitor.) 

[CyberGene]

8 hours ago, David Emm said:

Damn, I hereby vote for a "Switched-On CyberGene" album. I'll even pre-order!

Thank you! 😀❤️🍻

[CyberGene]

8 hours ago, David Emm said:

I'm one of those weirdos who whistles bits of "The Rite of Spring" in public

I also do that! 🤣 One of my favorite classical works! And it’s a hell of an idea to make a synth arrangement of (some of) it. Might do some day. 

[Tusker]

3 hours ago, CyberGene said:

I also do that! 🤣 One of my favorite classical works! And it’s a hell of an idea to make a synth arrangement of (some of) it. Might do some day. 

 

Hey CG, sign me up for buying your CD also. Another ROS fan here. ❤️ Can't wait to hear what you do with the bassoon solos. Double reed on synths!! 👍 😊

[Al Coda]

On 10/21/2022 at 5:31 PM, Tusker said:

It looks like Woody has been having the same CPU-Load questions we've had.  He's done a very interesting test with several of the leading synth plugins ...

 

Sylenth1 

Diva

Kontakt

DCO-106

Serum

Massive

Massive X

 

and provided conclusions and  graphs (at about the 12th minute).

 

I just checked ...

 

16% CPU load and occasional peaks of 17% w/ C.A. DCO-106 on my Lenovo 2.5Ghz quad core i7 laptop.

ASIO4ALL, Win10 Pro x64, 16GB RAM, 2 SATA-600 SSDs.

Patch "Big Strings",- 8 voices play 8-voice block chords w/ long(er) release times as they are programmed into the patch.

So some overlap of voices when changing chords.

 

I think it´s all patch dependend how much CPU/DSP cycles softsynths consumate.

TAL JP-8 consumates everything between 5 and 98% depending on patch and voicecount.

It´s seems to be the most CPU hungry device I tried on that laptop.

 

[Iconoclast]

Along the lines of re-creating a similar and relatively simple sound out of several different analog and virtual analog instruments, I ran into this video which does a similar thing with the Summit, Hydrasynth, Cobalt 8, System 8, Montage and GAIA.

 

[CyberGene]

I have some recent stuff for this thread, although I'm not sure how many people would really be interested, but hey, it's on-topic and I'm the author of the thread, so why not 😀

 

In a recent thread @Tuskerasked if someone has combined a synth to a piano. The idea entered my head and I coincidentally discovered a Tchaikovsky song for soprano and piano and so I decided to overdub myself and play the voice part on a synth. My Behringer Model D had been sitting dusty for at least a year, which is why it was my choice. I turned it on and luckily what I last dialed on it a year ago was perfect: a single oscillator with the shortest pulse (rightmost position) and a slight filter envelope. But it's the typical unmistakable Moog sound that reminds oboe/clarinet/flute. I connected my Hydrasynth via MIDI but also programmed the aftertouch to send control voltage output to open the filter and increase the modulation (vibrato), as well as the ribbon to act as a volume control (a silly idea but I was in a hurry and didn't know where my expression pedal was in the mess). It turned out very expressive:

 

I liked the result a lot. Do you? Anyway, I decided to try to recreate that patch on the Hydrasynth itself because I've always said on this forum how the Hydra can do any synth and especially a Moog should be e piece of cake, right. Well, wrong! I lost hours trying to recreate the same patch 100%. Turns out there's a big omission on the Hydrasynth. There are only 3 pulse waves with narrow width but none sounded like what I was looking for. So, I then used a square wave and the "PWM Orig" mutant which is exactly what is needed, right? Well, not exactly 😉 While I was able to approach the patch pretty close, it wasn't exactly the same, no matter what but what is more important: on fast legato playing there was some clicking between notes. And no, it's not the "fast digital attack/release", I know what I'm doing and it's not that. It turned out the problem is just the way mutants work. They need to buffer the incoming oscillator wave, so that they can apply the DSP that shortens the pulse width but that causes clicks. It was confirmed by Glen Darcey on the ASM Facebook group. And here's where this thread comes to mind again. The Hydrasynth is great and is my favorite synth but it just does the PWM in a rather over-engineered way. I don't need actual pulse width modulation. I need a fixed width. But they don't give you lots of fixed pulse widths, instead they give you 2-3 of them and the rest is you using a square wave going through a PWM mutant which is just too awkward for such a simple need. Here's a short video that I recorded to demonstrate the Hydra PWM artifact, I'm switching between Model D and Hydrasynth, see the captions:

 

So, I was rather disappointed with the Hydrasynth at that point. I like it for what it is, but in the context of this thread, it's not the perfect Virtual Vintage Analog synth 😢 Well, yes, it's not marketed as such but still...

 

And then I decided to recreate the same patch on my favorite Virtual VIntage Analog synth, the U-He Diva. Well, I spent more than 30 minutes to nail it down but it became 100% the same! I love Diva, I don't know of any better software!

 

I didn't record any direct comparison between the Model D and Diva unfortunately, however I recorded an entirely new piano and synth video, a Romance for horn and piano by Scriabin where I play the horn part with the Diva patch in question. OK, I upped it a but by adding shape modulation through the mod wheel and additional filter control through the pitch bend (no actual pitch change, only filter mod), so take it as it is, but you should trust me: Diva is simply fantastic! 😀 Here's the Scriabin video:

 

[David Emm]

First, your piano/synth duets are excellent. That romantic realm translates well, like Tomita's Debussy choices.

You're far ahead of me in piano chops, you bastard!  

The woodwind-like aspects are a perfect blend, well-tweaked for each piece. 

 

My piano days are gone, but love for the sound and muscle memory persist, thank Crom. IMO, piano and synths make good partners because there is a subtle similarity going on in some mystical Venn spot, like the natural partnership between AC guitar and human voice.

 

As to the Hydrasynth not being a primary choice for analog, its not for lack of trying. I've heard a couple of programmers come respectably close, so its a matter of the primary design goal rather than any real lacks. Its a perfect case for two tiers from the far ends of the pool. In Cretaceous lingo, my old Korg DW8000 was a very early wavetable synth that was great at DX bells, but thick string sections, not so much. Blending that with the strings or pipe organ from a Mirage sampler led to an 80s burst of audio 3-D. Virtually no one has just one synth, so MIDI the Hydra up with your next best analog/VA and let 'er rip.      

[NewImprov]

On 10/23/2022 at 1:00 AM, CyberGene said:

I also do that! 🤣 One of my favorite classical works! And it’s a hell of an idea to make a synth arrangement of (some of) it. Might do some day. 

I would totally buy your synth version of the Rite! I listened to your Bruckner piece, and loved it, it gets that otherworldly but expressive thing that Wendy Carlos and Tomita got. Great work!

[Tusker]

17 hours ago, CyberGene said:

And no, it's not the "fast digital attack/release", I know what I'm doing and it's not that. It turned out the problem is just the way mutants work. They need to buffer the incoming oscillator wave, so that they can apply the DSP that shortens the pulse width but that causes clicks.

 

I like both the Scriabin and the Tchaikovsky! Very very nicely done!

 

Love what you are doing with the ribbon and am curious why you chose the pitch bender for tonal control on Diva? It's beautiful. I love the release/reverb of the note off with Diva. However would you not get finer control with the long ribbon?

 

For me the crux of the matter is that in addition to the note-on rise and fall, (a short envelope) instruments such as the voice, violin and soprano saxophone have a rise and fall (a long envelope) which is mandated by the constraints of breathing and bow length. This second rise and fall crosses as many notes as the player can support. Usually it's a musical phrase, so it carries meaning. A long ribbon would be a good surrogate for this motion, (say sweeping from left to right and back as the phrase opens and closes). I could imagine a day when this type of "breathing" would become part of the muscle memory of the synthesist, just as it is with other musicians. There are organists for whom the pedal motion is roughly similar. A long breath. A sigh. A pedal is just not as subtle however. Neither is the Touche which is good for short envelopes but doesn't have the long phrasing capability. One of my favorite long phrase performances was at the Polar Music Awards, when Barbara Hendricks sang Deborah's Theme to composer Ennio Morricone. Her long breath support and precise control is key to the beautiful phrases, I think. I remember using a BC2 breath controller and having to choose between the ability to control dynamics partially from the keyboard or wholly from breath. The former was better for short phrases and the latter was better for long phrases. 

 

BTW CG, thanks for clarifying that it's not just the "restart from zero" envelope in the Hydra which is causing the artifacts. That's the first thought which crossed my mind, so it's good to cross that off. I'd love to hear you do this with something other than the mutant narrow pulse. As a synth, the Hydra really has it's own voice and you play it very well, so it would be a shame not to let it shine. Looking forward to hearing more. 👍 👍

 

 

 

[Jim Alfredson]

On 8/20/2022 at 5:51 PM, Iconoclast said:

The point being, not to hang yourself after listening to the Jump intro 7 times, but that even though these parameters are pretty simple, you can definitely hear the difference.

Would anyone care? eh, I dont know. But I know that once you get used to hearing the OB-6 do it, you can't unhear the difference.  

 

 

Missed this thread the first time around. That's a fun comparison. The OB6 definitely sounds like an Oberheim. The Forte patch wasn't too bad but I think I could program a better one. The VA engine in the Kurzweil's is pretty damn good, all things considered. I've been able to recreate sounds originally made on my ARP Odyssey MkII on the Forte and they are damn close. One trick is to use a separate LPF to attenuate the high-end a bit and to use the drive feature in one of the filters (don't recall which one at the moment). 

Regarding the overall point, the thing that most manufacturers miss (including Arturia and their plug-ins) is emulating the sloppiness of the envelopes. That randomness in each note due to envelope offsets is subtle but key to recreating a true vintage analog sound.

[CyberGene]

1 hour ago, Tusker said:

Love what you are doing with the ribbon and am curious why you chose the pitch bender for tonal control on Diva? It's beautiful. I love the release/reverb of the note off with Diva. However would you not get finer control with the long ribbon?

Well, this is trivial: because I am using both the pitch bend and modulation for different things (the pitch bend does a global filter cutoff control + volume, whereas the mod wheel changes the shape from narrow pulse towards square). Since they are next to each other it is easy to control them together which is what I do but I would have loved to have two mod wheels as on an OB rather than one self-centering pitch bend wheel... 

 

1 hour ago, Tusker said:

BTW CG, thanks for clarifying that it's not just the "restart from zero" envelope in the Hydra which is causing the artifacts. That's the first thought which crossed my mind, so it's good to cross that off. I'd love to hear you do this with something other than the mutant narrow pulse. As a synth, the Hydra really has it's own voice and you play it very well, so it would be a shame not to let it shine. Looking forward to hearing more. 👍 👍

 

Well, here's some geeky stuff I learned on the ASM Facebook group: if you use the Phazdiff mutant with a saw wave, then that would actually result in a pulse wave with different width depending on the amount but what is more important, the way Phazdiff mutant works is without introducing clicking artifacts (why - because phase diff subtracts an inverted signal with a phase difference from the main signal. If you draw it on a paper to visualize it, a saw wave + another saw wave that is inverted and with different phase and subtract one from another, it would indeed result in a pulse wave. The good news is you can start calculating the result within the same cycle hence no need for buffering and no artifacts/clicking). For the life of me I wouldn't have thought that to produce proper pulse waves with different width I should use a saw with PhazDiff rather than the most obvious square + PWM Orig. But there's a catch: for some reason it doesn't sound exactly like what I was seeking for, although the clicking is eliminated. I suggested to ASM to just provide 10-20 new precooked waves such as 5%, 10%, 15%, 20%... pulse. Anyway, I love the Hydra and I haven't bought it to emulate Moogs 😀 As a matter of fact it's great at emulating Moogs. But not everything.