The Official Sequential/DSI Forum

UDO Super 6 - 12 Voice Polyphonic Binaural Analog-Hybrid Synth

Razmo

  • ***
  • 2168
  • I am shadow...
    • Kaleidoscopic Artworks
Re: UDO Super 6 - 12 Voice Polyphonic Binaural Analog-Hybrid Synth
« Reply #60 on: February 17, 2020, 10:30:20 AM »
Ugh...I hate to be "that guy" but honestly...I think I'd like this more if it was VCO based. Digital oscillators are nice...but the layout and UI just feels like it should be 100% analog...almost feels like Arturia should have made it and called it the PolyBrute.
At 24 MHz, your ears can’t tell the difference between FPGA digital and Analogue — no one can. 
And FPGA OSCs are far more versatile. I love my analogue synths, but I like to see innovation like this.
I’d love to see Dave do an FPGA synth.  He has the knowledge (and ears) to do it well.

Though, again, I expect to be in the minority.

Yes, but you would still require, that the perfect algorithm for mimicking analog oscillators completely, existed... Which to my knowledge, it does not... I for one can easily hear that PEAK do not have analog oscillators... Yes, it is without aliasing and all, but it does not sound 100% analog to me... But maybe that was intentional, I do not know.

With that said, I still would prefer the UDO exactly as it is... It sounds amazing, even if the oscillators do not trick me...in fact, this "bineural" thing is basically what the Evolvers were as well, so there is not much new in it, except that so few synths actually use it.

I actually could see myself getting this, but not solely because of the FPGA, but also because having a poly with SSI analog filters are high on my list... At least until Dave does an SSI poly with much deeper engine than the UDO because that which is its advantage is also its weakness: hands on, no menus equals a shallow engine.
If you need me, follow the shadows...

OceanMachine

Re: UDO Super 6 - 12 Voice Polyphonic Binaural Analog-Hybrid Synth
« Reply #61 on: February 17, 2020, 03:46:27 PM »
At least until Dave does an SSI poly with much deeper engine than the UDO because that which is its advantage is also its weakness: hands on, no menus equals a shallow engine.
Well technically it already exists.   ;)

P.S. I already know what you're going to say next. It doesn't have FPGA oscillators, etc. 

Razmo

  • ***
  • 2168
  • I am shadow...
    • Kaleidoscopic Artworks
Re: UDO Super 6 - 12 Voice Polyphonic Binaural Analog-Hybrid Synth
« Reply #62 on: February 17, 2020, 07:47:47 PM »
At least until Dave does an SSI poly with much deeper engine than the UDO because that which is its advantage is also its weakness: hands on, no menus equals a shallow engine.
Well technically it already exists.   ;)

P.S. I already know what you're going to say next. It doesn't have FPGA oscillators, etc. 

Analog vcos please!  ;D... Or at least a mixture... I could easily see a Poly Evolver type binaural synth with these filters...

But yes... If nothing else comes, then PX would cover it, though I am not in the market for another deep sampler.
If you need me, follow the shadows...

Razmo

  • ***
  • 2168
  • I am shadow...
    • Kaleidoscopic Artworks
Re: UDO Super 6 - 12 Voice Polyphonic Binaural Analog-Hybrid Synth
« Reply #63 on: February 17, 2020, 07:59:14 PM »
Btw... There already is kind of a Sequential SSI polysynth with analog vcos, it is called Prophet 6... Its filters are modelled on the SSM filters which is basically what the SSI is cloning.

Thus it should be a logical step to make a chip version of the P6, with more voices and deeper engine comparable to a REV2/P12 in complexity. Give me such a device with 16 voices, and the PX way of handling the voices in 8 stereo or 16 mono modes (let each oscillator be pan'able in stereo mode so you choose yoursekf if you want a bineural setup) . Dual timbral with two FX per layer. That would be a synth hard to get around for me... Throw in an extra two wavetable oscillators (two analog and two digital in total) and I am sold.
« Last Edit: February 17, 2020, 08:03:01 PM by Razmo »
If you need me, follow the shadows...


Re: UDO Super 6 - 12 Voice Polyphonic Binaural Analog-Hybrid Synth
« Reply #65 on: February 18, 2020, 01:21:57 AM »
With that said, I still would prefer the UDO exactly as it is... It sounds amazing, even if the oscillators do not trick me...in fact, this "bineural" thing is basically what the Evolvers were as well, so there is not much new in it, except that so few synths actually use it.

The Evolvers can't really do what the Super 6 does, though. Binaural means more than just a stereo signal path. You can't de-phase or modulate the phasing between the left and right channels on an Evolver. You can, of course, offset the filter's response to the filter envelope between the left and right channels. That would be the closest feature of the Evolvers in this regard.

Btw... There already is kind of a Sequential SSI polysynth with analog vcos, it is called Prophet 6... Its filters are modelled on the SSM filters which is basically what the SSI is cloning.

There are two types of SSM style filters around in Sequential synths right now: One being modelled after the SSM2040 (Pro 2, Prophet-6, Pro 3 - not identical circuits in all cases), which was used in the Prophet-5 Rev 1&2 and the Kobol Expander; and the other one being modelled after the SSM2044, which was used in various E-MU products (SP 1200, Emulator I, etc.), the Korg Mono/Poly and Polysix, the PPG Wave 2.2, and the Simmons SDS5 drum modules. The latter one is used in the Prophet X and the Super 6.
« Last Edit: February 18, 2020, 05:35:46 AM by Paul Dither »

OceanMachine

Re: UDO Super 6 - 12 Voice Polyphonic Binaural Analog-Hybrid Synth
« Reply #66 on: February 18, 2020, 09:15:46 AM »
Dave says it's "loosely" based off the SSM (2040), which was in the P5 rev 1&2. As we all know it's a discrete design instead, but I'm wondering if there are perhaps other differences he's alluding to... My previous post shows the datasheet for the SSI chip version, which is closer to the original. Here's their 2044 based chip datasheet. Clearly a different topology and amount of pins.

Razmo

  • ***
  • 2168
  • I am shadow...
    • Kaleidoscopic Artworks
Re: UDO Super 6 - 12 Voice Polyphonic Binaural Analog-Hybrid Synth
« Reply #67 on: February 18, 2020, 10:19:14 AM »
Well, the tone of the Polysix filter I loved very very much when I had a Polysix... it sounds really vintage to my ears, but ofcourse the oscillators also play an important role in the end character... a 16 voice analog with REV2 depth would certainly catch my attention... but only with the right vintage vibe to it.
If you need me, follow the shadows...

Re: UDO Super 6 - 12 Voice Polyphonic Binaural Analog-Hybrid Synth
« Reply #68 on: February 23, 2020, 12:29:33 PM »
Excuse my ignorance, but what is Binaural Synthesis? I've search for an answer but with no satisfactory explanation.  I understand it may refer to some form of 3D sound, but that's all I've found.

I assume in this context it will probably mean that when you play a note, different voices will be triggered on the left and the right channel, for example a regularly tuned saw wave patch on the left channel and a slightly detuned saw wave patch on the right channel. It's something one would usually achieve with multitracking: recording the same track twice with the same sound, but slightly differently tuned, and then hardpanning it to the left and the right. That might also explain the numbers 6 and 12. So maybe you're left with 6 voices in binaural mode while you get 12 voices in normal mode.

If you look closely at the upper photo, you can see a section called "Voice Assign" and below there are two buttons, one with a sort of stereo field icon above it.

Yeah, I can understand if by binaural here they mean a frequency differential between left and right, although I think that's an odd term to use since in other contexts it carries some psychoacoustic trickery implications - either recreating sounds more accurately in the sound field by capturing the differences your head makes in perceiving a sound, or by offsetting the frequencies between each each to induce various brain states ala relaxation, daydreaming, alertness etc.

If we could get a synth that did the latter without the need for headphones... we might have to include heavy machinery warnings at our live shows.

LoboLives

Re: UDO Super 6 - 12 Voice Polyphonic Binaural Analog-Hybrid Synth
« Reply #69 on: February 24, 2020, 11:38:27 AM »
After hearing a few more demos...I'm not really overly impressed to be honest. Also the lack of splits/layers is really odd for a digital/analog hybrid....I feel the same way about the HydraSynth.

Re: UDO Super 6 - 12 Voice Polyphonic Binaural Analog-Hybrid Synth
« Reply #70 on: February 19, 2021, 12:01:30 PM »
I've purchased a UDO Super 6. Theoretically it's arriving in a few weeks, but they're backlogged and who can tell if one can trust predictions re: when next shipments arrive.

I'm in a band that makes indie, alternative, and synth rock/pop. I've owned the following: Prophet 6, Prophet Rev2, OB6, Juno 60, Juno 106, Alpha Juno 2, Nord Lead 2x, Nord Stage 3, Microkorg, Model D, Poly D.  Of those I kept the Ob6, Juno 106, Model D and Stage 3. 

I also own many of the big VST's (Omnisphere, Arturia V, Tal, Serum, etc.).

I simultaneously love the OB6 and Juno 106 sounds yet wish I could have those sounds and a synth capable of more/other types of sounds.  I sometimes question if I should've sold my P6, I bought it toward the beginning of this synth journey I've been on the past several years and after getting to know the OB6, I might get a lot more out of the P6 than I did at the time. In a mix my VSTs would probably be all I ever need, but the heart wants what it wants, and there's something to the workflow and joy of playing real synths and that something extra in their sound.

I went to Perfect Circuit a couple months ago for several hours and played with the Prophet 10 rev4, Roland System 8, Hydrasynth,  Moog One, Novation Summit, Prophet 6, OB6, Nord Wave 2, and the Super 6.  The Super 6 was the one that blew me away sonically and in terms of interface/workflow.  As you can probably tell, classic Roland synths and tones speak to me, and the Super 6 felt like a Jupiter 8/6 on steroids. I wish it had 61 keys and a simple screen (not for menu diving, but for patch names/bpm, etc.).  But the binaural feature isn't just a gimmick.  And even without it it's a 12 voice FPGA synth with 2.5 oscillators (one of the LFO's can be put into audible range, and the two main oscillators are different and capable of many tricks in their respective rights).  One moment it sounded like a Jupiter, the next it could go into FM and wavetable territory, or it could be dialed into a one osc + sub setup and pull off great Juno-like tones, albeit with the binaural feature bringing pads to life.

As you can tell, I was a big fan.  This is all super subjective, and if it's not others' liking that's great. It is a wonderful time for poly synths. What made me happy about the Super 6 is that it felt like something both new, and something that could pull off/harken back to my favorite classic Roland synths.  I suspect Sequential is going to do a new rev of the OB-X next, and I'll check that out. But I'm most excited to see if Dave has any more new designs in him, and I really do think he should take a long look at the Super 6 for inspiration.  A Sequential synth with an FPGA osc, an analog osc, analog filters, etc. could be a new direction, as I feel with the P6 and P10rev4 it feels a bit like another pure analog poly would be a re-tread unless it had some new fundamental feature to distinguish itself.

Who knows, I'm out of my depth. I love these things, but I'm not engineer or master.  I'll just say that the FPGA oscillators stand toe-to-toe with analog.  Whereas playing that System 8 etc. you can hear the difference.  I didn't miss my 'analog warmth' at all when playing the Super 6.

Anyway, enough of a ramble.  Hopefully it arrives soon and I can report back with the pros and cons.  Likely cons are just how small of an outfit UDO is and it seems they're still ironing out bugs.

timboréale

  • **
  • 166
  • MIDI nerd
Re: UDO Super 6 - 12 Voice Polyphonic Binaural Analog-Hybrid Synth
« Reply #71 on: February 19, 2021, 12:48:19 PM »
Interesting summary, thanks for that. I feel obliged, however, to point out as a DSP engineer that there's no difference whatsoever between an FPGA oscillator and any other traditional DSP oscillators in the end result. People make a big deal about them, but there's just no justification for it. One of the supposed "differences" is that "FPGAs run faster" which is total bull. FPGAs might have a higher core clock speed than ye olde DSP chips of yore like the SHARCs and 56k series, but you can get ARM cores that run circles round most of the non-military-grade FPGAs out there. And in the end, digital math is digital math, no matter what silicon you compute it on. If you do the math right, and you do it fast enough, you end up with the same results going into the DAC one way or the other.

So, I'm sure you like the sound of your UDO, but it's not because it's got FPGA oscs that give it that sound. It's just that the designers of that synth were better able to execute their programming vision on that hardware than on something else. But the math underlying it can be crunched on any chip with the guts to do it using the same principles identified by Fourier, Nyquist, etc. all the way forward. Whether any future Sequential products use FPGAs or not* for their hybrid oscillator section will have absolutely zero bearing on their ability to produce beautiful sounds.

*And the ones they currently have that don't use FPGAs have no faults due to their particular DSP implementation either - though some people may not like the particular algorithms they're running, since it's the algorithm that creates the waveform, ultimately. Running faster and oversampling is only one way to solve problems of aliasing, etc. More complex maths can also do this at lower sample rates (and those chips run at hundreds of MHz so internally they can be oversampling even if the DAC isn't running at double or quad rates), and it's always a tradeoff of development and mathematical complexity versus cost versus available hardware in the end. FPGAs have some serious limitations too, that no doubt we'll be seeing more of as we see more of them in synths.

Anyways, enjoy your UDO and it'll likely be a fine addition to your already prolific arsenal. Just don't think it's because they're FPGAs. :)
Rev2-16, Prophet 12 Keyboard, Nords, etc...

Re: UDO Super 6 - 12 Voice Polyphonic Binaural Analog-Hybrid Synth
« Reply #72 on: February 19, 2021, 08:37:18 PM »
Whether any future Sequential products use FPGAs or not* for their hybrid oscillator section will have absolutely zero bearing on their ability to produce beautiful sounds.

A lot of people still think the ESQ-1 is one of the most amazing sounding synths, and I'm pretty sure it ran on potato chips.  It's also loved for it's crunchy sound, so FPGAs would be entirely lost on it I would imagine.

There was even Atari software that would let you not just randomize patches for the ESQ-1, but control the amount of randomization, way back in the 80s!  I just saw a video of Cevin Key doing that recently.
« Last Edit: February 19, 2021, 08:39:23 PM by proteus-ix »

Re: UDO Super 6 - 12 Voice Polyphonic Binaural Analog-Hybrid Synth
« Reply #73 on: February 20, 2021, 12:48:42 AM »
Interesting summary, thanks for that. I feel obliged, however, to point out as a DSP engineer that there's no difference whatsoever between an FPGA oscillator and any other traditional DSP oscillators in the end result. People make a big deal about them, but there's just no justification for it. One of the supposed "differences" is that "FPGAs run faster" which is total bull. FPGAs might have a higher core clock speed than ye olde DSP chips of yore like the SHARCs and 56k series, but you can get ARM cores that run circles round most of the non-military-grade FPGAs out there. And in the end, digital math is digital math, no matter what silicon you compute it on. If you do the math right, and you do it fast enough, you end up with the same results going into the DAC one way or the other.

So, I'm sure you like the sound of your UDO, but it's not because it's got FPGA oscs that give it that sound. It's just that the designers of that synth were better able to execute their programming vision on that hardware than on something else. But the math underlying it can be crunched on any chip with the guts to do it using the same principles identified by Fourier, Nyquist, etc. all the way forward. Whether any future Sequential products use FPGAs or not* for their hybrid oscillator section will have absolutely zero bearing on their ability to produce beautiful sounds.

*And the ones they currently have that don't use FPGAs have no faults due to their particular DSP implementation either - though some people may not like the particular algorithms they're running, since it's the algorithm that creates the waveform, ultimately. Running faster and oversampling is only one way to solve problems of aliasing, etc. More complex maths can also do this at lower sample rates (and those chips run at hundreds of MHz so internally they can be oversampling even if the DAC isn't running at double or quad rates), and it's always a tradeoff of development and mathematical complexity versus cost versus available hardware in the end. FPGAs have some serious limitations too, that no doubt we'll be seeing more of as we see more of them in synths.

Anyways, enjoy your UDO and it'll likely be a fine addition to your already prolific arsenal. Just don't think it's because they're FPGAs. :)

What an awesome response.  I'm no engineer and ya I get it, math is math and computer power is computer power.  I didn't mean to come off as some kind of FPGA evangelist. I don't know the tech at all. I've heard the creator of the Super 6 talk about why he prefers FPGA's to DSP in terms of higher sample rate and the way you have to build/code for it, but that's all lost on me.  What I should really say is just, for whatever it's worth to others reading this thread, this is the first non-analog synth I've ever played where I didn't go 'oh that's nice it can do all the things, but where's the sound man?'  I.e. I'm drawn to analog synths (and even then you're right, I didn't like some analogs - e.g. I didn't really like the way the Deepmind or Rev2 sounded') for the sound, hence why I've stuck with the Ob6 and Juno 106, they have very wide sweet spots and are very easy to make sound gorgeous.  When I tried the System 8, the Hydra Synth, and the Summit, they each had their moments, but they didn't have that sound that makes you go 'Woooahhh!' I've had that reaction with a Juno, an Ob6, and now the Super 6.

Thanks for the interesting response.  If you're curious here's one of the interviews where the guy talks about FPGA's.  Whether it's marketing gobblegook or not, it seemed like he at least believed what he was saying re: the pros over DSPs.

Here's one where he talks about it in detail:  https://youtu.be/4Rr4TaQCK90?t=3290


(Other vids of him talking about it)
https://youtu.be/u_L4rOwoUho?t=689

https://youtu.be/4Rr4TaQCK90?t=1285


Thanks!
« Last Edit: February 20, 2021, 12:53:17 AM by Samtheon47 »

timboréale

  • **
  • 166
  • MIDI nerd
Re: UDO Super 6 - 12 Voice Polyphonic Binaural Analog-Hybrid Synth
« Reply #74 on: February 20, 2021, 06:56:44 AM »
Thanks for the interesting response.  If you're curious here's one of the interviews where the guy talks about FPGA's.  Whether it's marketing gobblegook or not, it seemed like he at least believed what he was saying re: the pros over DSPs.

Here's one where he talks about it in detail:  https://youtu.be/4Rr4TaQCK90?t=3290

All I got from this is that he personally prefers the building block editors of FPGAs over actually having to write code, he even admits the math is the same and thus the end result.

His comments about parallelism are correct, when speaking about a single chip, but irrelevant and solvable by simply using a few more dsp chips, like Dave Smith did with the Prophet 12 > same end result, different means of accomplishing it, but as for high sample rates, there's no advantage to them for audio processing - it's the same snake oil as boutique audio cables, etc. People can't hear MHz frequencies or even 100's of KHz and anybody who says they can has in actual double blind tests been proven wrong, so I have no idea what he's going on about that. Not to mention that while all the pros of FPGAs are easily attainable by good DSPs and microprocessors, the cost of the DSP architecture is far lower AND the development of them is far easier, for a competent programmer. He even admits to this too. So he's taking a hard, expensive, route all to brag about some MHz sample rate that is ultimately completely pointless. The controls on a Roland don't even operate at audio rates - the little CPU in a Juno 106 for instance is a Z80 running at something like 1MHz and polling the interface (knobs, switches, etc) at likely far less than even 1KHz (I don't have the exact specs handy, but those are typical values for that era). And those are DCOs controlled by that tiny slow little thing, but nobody says they sound terrible... I digress.

After hearing this interview I feel even less like buying an UDO now that I know he's tossed 10% of the build costs into that FPGA... lol!

Still, if you like how it sounds, I wish you years of happy playing!

Edit: Oh, I almost forgot, if you want to know why it really sounds the way it does, look at the filter, not the oscillators. Oscillators contribute a much smaller fraction (for traditional subtractive waveforms, I am saying) of "the sound" than do the filters. Roland has a very distinctive filter architecture and that's a huge part of the sound.
« Last Edit: February 20, 2021, 07:01:43 AM by timboréale »
Rev2-16, Prophet 12 Keyboard, Nords, etc...