nammday1 - 04

The Prophet 12 from Dave Smith Instruments is a landmark synth: packed with polyphony and sound features, it’s the latest demonstration that “new” and “synth keyboard” can go together. So, for our first-look tour through the new Prophet, we get a special treat – writer, electronic musician, and mathematician Gina Collecchia got a hands-on look with the engineers who designed the instrument. Gina, who demystifies signal processing for all the rest of us in her book Numbers and Notes, get to talk with Chris, one technologist to another. -Ed.

Meet the newest member of the Dave Smith Instruments family: the Prophet 12! A hybrid synthesizer that mixes analog circuits with digital synthesis, this instrument represents a lot of “firsts” for the pioneering, San Francisco-based company. I sat down with three of Dave Smith’s engineers at this year’s NAMM in Anaheim, CA, and got the scoop on what’s different about the Prophet 12.


Following the naming convention of the Prophet 5 and the Prophet 8, the newest Prophet has 12 voices. Signals begin in the digital domain with digital oscillators. Every voice can have four of them, PLUS a “sub-oscillator,” humming an octave below the fundamental frequency.

Sixty oscillators. This alone should indicate the magnificent range of sounds you can get out of this beast. From classic synth tones to cinematic soundscapes, the Prophet 12 is a marvel. Interesting, then, that I consider it the most transparent synthesizer I’ve ever seen. Its design has an emphasis on clarity for musicians to really understand how their sound is crafted and choose knobs with confidence.

The display on the Prophet 12 is clean and the updated controls make it easier to use than ever. (Click for a closer look.)

Diagram of the signal flow in the DSI Prophet 12. The digital portion is colored blue while analog is orange-tinted.

The digital effects, accessed in the “Character” region of knobs, have charming names like Hack, Decimation, Drive, Girth, and Air. The first two could only be done in the digital domain: Hack crunches bits and Decimation does sample rate reduction. Drive is the familiar effect of overdrive and Air/Girth are two high and low shelf filters that have, you know, airy and girthsome properties.

Beyond the effects unit, the remainder of the synthesizer is analog. The signal is passed through resonant, 2- or 4-pole, low pass and high pass filters with a 14 octave frequency response–a whole octave greater than the Prophet 8.

The VCAs control panning and ADSR envelopes, and then the signal reaches a tuned feedback loop (“Dave’s Fave!”) that’s wired up with the same technology as that from the DSI Poly Evolver. Cranking up the resonance of the low and high pass filters gives the feedback more prominence. Multitap delay lines (4 per voice!) are placed in the Prophet 12 next to the analog path to simulate effects like flanging and reverb.ย All of these voices are then summed and routed to a pair of stereo analog distortion units with wet/dry hardware mix.

The user interface of the Prophet 12 is perhaps the most dramatic element of the makeover.

The display on the Prophet 12 is clean and the updated controls make it easier to use than ever. (Click for a closer look.)

The display on the Prophet 12 is clean and the updated controls make it easier to use than ever. (Click for a closer look.)

Menu selection is performed via tabs and buttons instead of arrows. Assigning sources and destinations is a breeze in the Modulation region, mapping controllers and knobs intuitively to anything else. A clever sort algorithm provides graphical display of what’s affecting what in the 16-deep modulation array. Using the “Show” button, any parameter can be viewed without changing the sound as knobs are touched.

Presets can be stored in “playlists” and accessed with the program buttons, a feature designed with keyboardists in mind.

Four LFOs per voice can also be piled on, and since they go up to 500 Hz, these are essentially 48 more oscillators! ONE HUNDRED AND EIGHT OSCILLATORS. Since these too are digital, there’s much less voltage control to take care of in the Prophet 12. The voltage control that does occur is updated at a rate of 11 kHz. These LFOs are syncable to the internal MIDI clock or the new 32-step programable arpeggiator in the place of a step sequencer. The high update rate provides a buttery smooth texture to the modulation, and makes the synth’s sound pretty much indistinguishable from that of any analog synth.

There were many other advantages to going (partially) digital in the Prophet 12. “Digital is way more flexible,” says engineer Chris Hector. “Analog oscillators are not really stable enough for FM synthesis. The oscillators have lower than ever aliasing to clean up even more digital artifacts.ย And, the digital effects are where it’s at for the characteristic DSI sound.”

Oscillators are stacked in the Prophet 12 such that Osc 2 modulates Osc 1, Osc 3 modulates Osc 2, and Osc 4 modulates the third. Therefore, this mean machine can do subtractive, wavetable, FM, and AM synthesis, and any combination thereof. You wouldn’t think from just the look of it that the Prophet 12 had such range, but its design maximizes both simplicity and possibility. And to me, that’s the number one goal in new instrument design.

Gina, like CDM’s editor-in-chief, is a previous resident of Louisville, Kentucky. Maybe there really is something in that water the bourbon and racehorses like so much. She is now at Stanford University. -Ed.

  • Not R Kelly

    My bank account’s telling me nooooo. But my bodyyyy… my bodyyyyyyyyy’s telling me yeee-heeeessssss.

  • Andreas

    I hope this monster ( Schmidt Polyphonic Analog Synthesizer) gets released some day.

    • KBSoundSmith

      It has been released. You have to order it, then they make it. It will just set you back $25000.

    • Phil

      And you’ll also need some friends to carry the 150 lbs the Schmidt weights.

    • Miles Bader

      Of course, when adjusted for inflation, the Prophet 5 originally cost like $15000 in 2012 dollars!

      I suppose the sort of people who could buy a P5 back then could pick up a Schmidt today… :]

      [We’re totally living in a wonder world of cheap synths these days… :) ]

  • markLouis

    Sixty oscillators, forty-eight more oscillators, etc. In the thread here when this was introduced, someone commented something like, “Who has time to program that many oscillators?” and someone replied something like, “You’re doing it wrong.” Would anyone elaborate, in general, how a musician would use resources like this to “Do it right”?

    • Peter Kirn

      Well, remember, you’re not programming them independently. You have that many in order to get 12-voice polyphony. Think about playing two triads (left and right hand) – you’re already at six-note polyphony. Start stacking and this isn’t all that crazy. And the sub-oscillator is counting as an oscillator, too.

      We’re definitely talking some layered sounds, not *just* stripped-down bass lines; that’s clear.

      On the other hand, this is also an answer to the question someone asked about monosynths of why sometimes you might be fine with just one-note polyphony. It depends on how you’re assembling and layering your timbres, and what you’re trying to accomplish.

    • markLouis

      Thanks, Peter. I didn’t mean to imply the numbers were “crazy” but Gina sounded so breathless–“ONE HUNDRED AND EIGHT OSCILLATORS”–that I wondered how pros think about these things. I guessed you mentally block-out groups of oscillators for leads and groups for harmony, but if you start thinking like an arranger keyboard, with percussion voices and mid-range voices and pads all going at the same time, then these numbers seem to be numbers that can create good complex leads, but not really wildly large numbers of interacting sounds at all. It seems to be like a good lead and some harmony, if I’m understanding that blocks of oscillators get used up reasonably quickly.

    • Peter Kirn

      Well, because those numbers add up so quickly, there’s an advantage to using digital oscillators. (ducks)

      I mean, to me, the beauty of something like this – or those monosynths – is that they aren’t arranger synths; you’re building up these oscillators to create sounds that make them feel like a single instrument (rather than a whole computer). But there are definitely sounds that call for this sort of architecture. And again, in treating these as different instruments, I think people are going after different instrumental roles for the monosynths than something like the P12.

    • A

      The proof is in the interface. :) it’s not so much the time as being able to manage and keep track of all the parameters.

      I’m 50/50….

      On the positive side, the performance controls on the Tempest look awesome, say they took that concept, but made them assignable groups that you could also modulate with other modulators. That gives you a lot of rhythmic and evolving sound possibilities. Also, it seems the controls go into audio rates now, so that gives you the option to do like, I dunno, micro-harmonics within sounds and just get really in depth programming expressive sounds.

      Also positive, the evolver had assignable modulation that let you re-use modulators on different targets, but in different and negative amounts, this helped in programming relationships within sounds. Personally, I’d like to see more of this, but with less fixed relationships. Things like inverse, half, 2x, 1/5, etc. Again, the evolver had a nice feature where you could specify a modulation in step amount and this helped to work quickly. EMU had a cool virtual patch cord system that I wish came back where you throw in lag processors, scalers, all kinds of neat stuff.

      On the negative side, in the polyevolver, another synth with lots of options going on, the 4 part combo mode didn’t let you save the individual parts of the combo, making it useless as a way to group controls. If you want 4 different voices to sit together, you have to hear them together and adjust to taste, but you can’t, so, you have to slowly jump back and forth and edit blind.

      It’s really depressing. It’s still one of my favorite synths, but I know it could be so much more fun to work with and I could push it harder, if the interface wasn’t fighting me. :(

      The p12 has only 2 sides, but it appears they’re independent this time, at least. I dunno, I could see up to 6 being useful! The prophet 12 is going to be amazing for synthesizing percussion, but if like the polyevolver you have to mash all your elements into one patch, it could be hard to manage.

      So hopefully, the p12 has some group edit and copy functions to re-use parameters. It’s not hard to use that many parameters, it’s hard to keep track of and integrate them all, without editing the wrong thing!

    • gina

      I think the Prophet 12 does have those sorts of improvements to its program banks and voice controls. I know there’s a “Unison” mode which allows you to make adjustments to every active voice together (pretty sure that’s the group edit you’re desiring), and I would assume that voices can also be soloed, not designed together. You can absolutely save programs and presets much easier and more quickly than before; I think it was a big focus for the engineers in this synth.

    • Miles Bader

      I’m totally with you on the PolyEvolver complexity… I loved the sound of the PE (and the keyboard action, nicer than the P’08 I think… TT), but despite the knobby interface, found the interface quite overwhelming, and tons of stuff (like splits etc) required a lot of fiddly fiddling… Even just the sheer number of knobs on the PE was pretty daunting.

      The P’08 is a simpler architecture, but I think the UI is better thought out as well. It seemed pretty clear to me that DSI learned a lot from making the PolyEvolver, and that the P’08 reflects that.

      It will be interesting to see what the P12 is like in practice, as many of the PE’s features have returned… :]

  • tempestfail!

    hopefully better than the tempest that has been abandoned.

    • carson

      The Tempest has not been abandoned by any means. We are still working on it’s OS and functionality

    • markLouis

      “We” — You work at Dave Smith? I really love DSI products, but I wonder if you can comment on an impression of mine: So many low cost products these days do quite a bit–the Sub Phatty, Minibrute etc analog; endless digital things; and the Kaossilator range etc. in percussion and playing around. My impression is that people who have followed Roger Linn/Dave Smith for decades, the real cognescenti, can look at DSI products and get a good gut feel, a mental and emotional feel, for the capabilities that separate DSI products from the low cost stuff that does similar things but less-deep things. 1) Do you think more people would buy DSI products if people simply had a deeper understanding of what the technical resources DSI products provide can accomplish? 2) Is the market encroachment from the low-cost, possibly less-deep technical things threatening to DSI itself, I mean, is there a chance DSI will have to re-think product design for the future?

    • Peter Kirn

      Wait, but —

      * Some of those specific products, and products in those price ranges have been around for many years, too.
      * DSI sells products in that price range.
      * I’m not sure what’s more or less “deep” about the different products you’re mentioning here, generally, and
      * Customers of music products don’t always use the complete functionality of their tools. Ever. Hell, I don’t always. ๐Ÿ˜‰

      In other words, I’m not sure DSI can answer what you’re asking. What are you asking?

    • markLouis

      “many years”–the Minibrute came out last year, sub phatty this year, Kaossilator 2 pad last year. The Tempest costs $2k, the Kaoss range is one to four hundred. When products have less functionality it is easier for customers to use the FULL functionality because they can SEE the full functionality so I wonder if more more more (A HUNDRED AND EIGHT OSCILLATORS). On the other hand, more more more can do cool sophisticated stuff, if there is a market/need/desire for “sophistication” or cool or people can spend money for functionality they’ll never use. I was curious if such issues are impacting DSI. I certainly didn’t mean to offend anybody or ask unanswerable questions. I was just curious in a conversational way about these issues and DSI as a particular company. Sorry. I’ll shut up now.

    • Charles

      DSI has the MoPho, Tetra, and MoPho x4. Until the Tempest and Prophet 12 came along, some people were accusing them of focusing too much on the low-range. I think they have the balance right.

    • Charles

      (Oh, and the Evolver which started it all can be had for under $500 of course.)

    • Pym

      We are constantly re-thinking our specific product design approach, while adhering to a very simple overall goal: making cool, fun synths that are easy to use

      You make a good point where too many features mean they’re not as easily accessible (without a REALLY well thought out UI) and people end up not using them because of it. With the P12, there are very few things that you can’t access with a single button push. For example: the OLED was re-thought out to use tabs rather than arrows to more easily get to parameters so even the “hidden” stuff is no more than a single button push away. The mod section was completely re-thought out so it’s easy to edit existing and assign new mods, which people often overlooked because they’re complicated and slightly hidden in our previous UIs. We learned a lot from the Tempest and tried to apply the design philosophy it to a more simple product… all in all, we’re all thrilled with the way it’s taken shape. I honestly have more fun using this synth than anything else we’ve made… it’s all straightforward and simple so you never have to scratch your head and wonder “what the hell does this do?”

      We don’t really cater to “market demands.” We make stuff that we have fun making and get both us and our users excited.

    • HaroldZeb

      Tetra owner here. I feel your pain. It took something like 3 years to even have the ability to save patches in multimode, on a synth largely marketed as multitimbral. LFOs still don’t stay in sync, noticably clicky envelopes, NRPN transmission issues, and various lesser bugs. How does the father of MIDI screw even that up on his own gear, and leave it unfixed for years after release? I can’t sit down and tweak the knobs for more than an hour without it freezing up, requiring a hard reset. I keep it because we’re told these bugs will be taken care of year after year, and I really want to believe, because it has the potential to be an incredible synthesizer.

      Keep this in mind- if you’re considering buying a Dave Smith instrument, expect a plethora of bugs from the start, and pray the worst will be fixed before their next product offering.

  • Eric Schlappi

    So it’s got digital oscillators and analog filters. Wasn’t that the big deal in the 90s? Then it sort of died a deserved death and people started playing with real analog again or using a computer?

    That’s how I remember it anyway. Analog filters do not an analog synth make, or even much of a “hybrid.”

    • Peter Kirn

      An earlier draft of this story made more of the “analog” circuitry than it should have, so I take part responsibility for this.

      I thought we might be getting to the stage at which people consider the architecture that makes the most sense.

      I’m trying to imagine that somehow combining the versatility of digital oscillators with the sonic performance of analog filters and DCAs *doesn’t* make sense as a design. That you’d actually want it to “die a deserved death.” Maybe you can explain why that’s a bad idea?

      The essential feature of these keyboards to me ultimately comes down to sound and playability. What’s analog and what’s digital in the circuitry is a sonic issue, not a religious one, and I hope that us describing the portions that are analog as what they are doesn’t cloud that.

    • Pym

      Back in the 90’s the digital oscillators had a “sound” to them because of the limited processor power you had available. In large part this was due to aliasing and bit/sample rate limitations. This isn’t the case anymore and you can get nearly alias-free oscillators even when doing complex functions (like sync, glide, and modulations) that were places the digital oscillators definitely showed their weakness in the 90s synths. Without the aliasing, it’s very, very difficult to tell the difference between a very well designed modern analog and digital oscillator in a double blind test… especially when it’s going through an all analog signal path that warms them up. In our tests really they just sounded ‘cleaner’ than the analog ones. Some of the artifacts that the analog oscillators have could also give them away, but if we really wanted to we could re-insert them in the digital realm without much trouble. This advantage of digital is huge for things like FM: you get no phase problems since the oscillators are not free running, no click or delay from starting them up if you do want them to reset per note, no drifting or noise (unless you WANT it there) and far, far more variety in the sound than you can get from a similarly priced analog circuit.

      We’re also very quickly reaching a point where not just the analog vs. digital debate but even the VCO vs. DCO debate is going to become invalid (or VCF vs. DCF or whatever)… chips are coming out that can do ALL the CV updates at full audio rates, so there’s no audible aliasing on the control voltages. That takes away the big advantage VCOs had.

      Where digital still is catching up is the non-linear effects… filter, VCAs (to some extent), distortions and some compressor/limiter type effects (although the modelling . You can do phenomenal models nowadays, but to get a similarly amazing sound you need a lot of processing power… so we chose to add them in the analog domain. With the higher resolution CVs, you have nearly the best of both worlds. Sonically it’s just… delicious!

      At some level what the debate is about is the simplicity of getting a great sound. One of the less obvious advantages of analog is that you are confined to a very limited set of options and those options have to be set up in a way that sounds great across the whole range… so no matter WHAT you do, it sounds good (at least that’s the goal). Digital stuff is less likely to have as much attention put on that sort of detail, you are often given access to the whole range of a parameter rather than a nicely thought out curve that limits you in a way that sound great but has a good amount of range. More options gives you more control, but sometimes that’s not a good thing. We try and focus as much on our digital side as the analog to limit it to a range that always sounds good rather than giving people access to lots of options and different algorithms (not mutually exclusive of course, but with limited man power it’s one place you have to choose what to focus on).

    • Charles

      Which 90s are you remembering? Analog synthesis was dead in the 90s, with the exception of a few holdouts and the people raiding the pawnshops.

      Digital technology can do oscillators really well – better than analog in some respects (FM, wavetables). Digital models of filters still have a ways to go. So a hybrid synth with digital oscillators and analog filters makes a *lot* of sense. (And the Prophet 12 doesn’t just have analog filters, the VCAs and distortion are analog as well.)

      Other hybrid synths with digital oscillators and analog filters: the Waldorf Microwave and its ancestors, the Prophet VS, the Korg DW-8000, Ensoniq ESQ-1 and cousins, the DSI Evolver, the Mutable Instruments Ambika and Shruthi, etc.

      These are all pretty highly regarded synths, and with good reason. Frankly I’m surprised there aren’t more – but most of the hybrids got caught by the rush to ROMplers at the end of the 80s and never had much commercial success, and Dave Smith is just about the only one to revisit the approach. The P12 is very good news.

    • Eric Schlappi

      I was thinking of the Waldorf Q+ and the synths you are mentioning. I feel like there were a good number more but I confess I’m having trouble trying to pull them out of my brain. The only one of the synths you mentioned I think sounds good is the Evolver, and that has DCOs along with the digital oscillators, and even so you would not want listen to one voice without all the fancy modulation it does.

      I mean, if it sounds good it is good. I just feel like all the new software synths are doing is reminding me how good the old stuff sounds and it’s ludicrous to dress something like that up with some analog filters and charge this much for it.

      If it gets you excited though…

    • Charles

      None of the synths we’ve mentioned was released in the 90s (the Q+ came out in 2002). If you don’t care for digital oscillators that’s your prerogative – luckily you have more VCO and DCO options than ever right now. But your ears are definitely not mine and there are lots of people like me who really like the sound of hybrid synths, even the old ones with aliasing, so this is pretty exciting news. Dismissing it as some kind of sales gimmick is just silly.

    • A

      More the 80s. Analog was still “worthless” because of lack of MIDI and hybrids got overlooked because people wanted realistic presets and lots of them.

      I see hybrids as where synths started to head somewhere super interesting, but the interfaces were mostly crippled and the computers inside them were underpowered.

      The time is right and the p12 is super exciting to me.

      For one, it sounds great.

      Two, it might be the best FM interface yet. FM sounds most awesome when there are of modulators and the p12 has those in spades. FM also generates loads of harmonics, perfect for filtering. I’m curious to know how the 4 oscs interact. I hope there is the ability to resample your waveform a la the Prophet VS (according to the manual , at least). I hope they came up with some interesting and smart ways to manage all this modulation, like using the performance controls from the Tempest. ๐Ÿ˜‰

      Three, feedback delays, DSI make some demos doing waveguide synthesis! :) The feedback is analog sines,yeah? Run that through the analog filters and you have a completely analog synth. ๐Ÿ˜‰

      Only thing is, I hope the weighted keys are adding a fair amount to the price. I don’t see myself being able to raise 3k. :(

      …and I don’t want any more keyboards. Hope to see a 6 voice split-able 3/3 module, or something along those lines. It be neat to see like a red and black evil twin to the the Tempest too

  • Ben

    It seems to be a really nice synth. I already have an evolver desktop and a mopho … time for me for a poly synth? I just would like to see if it’s possible to have more parts in the future, with voice reservation, perhaps? From the “inventor of midi”; i would expect more midi channels to be used. That surely can be changed in the microcontroler part?

  • gesslr

    Aw hell. I gotta figure out a way to buy one… :-)

    I have not seen a recently released keyboard that offers something my arsenal of virtual instruments does not.

  • scottrod

    Want. You will be mine. Oh yes, you will be mine.

  • scottrod

    Saw Dave on the Past, Present, And Future Of Midi video where he was talking about poly AT. He mentioned that synth manufacturers are at the mercy of the keyboard manufacturers (Fatar, et al.) for keybeds.
    No. You’re the customer. YOU demand what you want. As long as you give in and buy simple channel AT units, that’s what they’re going to make. I would EASILY pay an additional $350 for a nice semi-weighted (Prophet T8 feel) poly AT keyboard attached to the Prophet 12. And while you’re at it, make sure it has a full127 midi resolution. The day I discovered my T8 was only 64 resolution was the day it went into the corner. And, it doesn’t transmit poly AT out the midi port.
    Dave also said people don’t want to spend the ime to really learn how to play the keyboard with the requisite muscle control. If I had the right controller, yes, I WOULD spend the time (years) to achieve the muscle memory required to weild it properly.
    Infinite Response has a nice unit, but the price with the magnesium body is just too much. Maybe you could buy/license keybeds from them. NOW you’ve got a synthesizer for the ages.

  • Niklas wallin

    Best 12voice monophonic possibly.. It has a very special way of doing voice stealing so you cant play slow attack&release in multimode . It’s a shame really, and the customer care has their heads stuck waaay deep down in the sand about it. Still love it but it could have been so much better if they were’nt such assholes about it and force faul sharp attack sound on me. it’s not a polyphonic instrument, it’s a 12voice monophonic one, keep it in unuson and you’ll be fine.