Future Features - KEYBOARD INTELLIGENCE?

[galaxiesmerge]

Hi,

I know I have touched on this elsewhere, but, I have put together some thoughts in a more structured way:

There are two parts here (some of it revisits some earlier posts that were in some areas fragmented):

PART-I: Keyboard Intelligence
PART-II: Oscillator and Filters

Okay, so here goes.

PART-I: Keyboard Intelligence

This is about having smarter performance code for detection and mapping of whether the player is going *up* a note or *down* a note and whether the user is pounding on a single note. A feature that would allow sound-switching (sort of like what the rotors do) would allow you to have, for example, a violin up-stroke going up and a violin down-stroke of the bow going down, and, in repeating notes, alternating between the up/down stroke.

Keyboard intelligence is *way* under represented in performance instruments although ROLAND corporation did make an effort with (monophonic) "Articulative Phrase Synthesis". Keyboard intelligence would be a far better solution. For example, when hitting a chord, automatically creating a "strum" (which is really just a super-fast arpeggio!). I am sure that this can be done, and, I would be happy to throw in whatever knowledge I have in helping achieve that. I think that SOLARIS would be benefit enormously in showing off "Keyboard Intelligence" for synthesized instruments also. As an example of keyboard-intelligence I will list a variety of other elements:

1) Intelligence Portamento ---- left hand going down a scale versus right hand going up a scale provides two portamento effects in opposing direction
2) Crossing Portamento --- the Portamento algorithm always chooses the *largest* distance between notes to execute the portamento --- conversely, user can set it to the smallest distance between notes.
3) Intelligent Percussion --- automatic drum-rolls, speed controlled with the modulation wheel
4) Modulation Effects assignable to note interval changes: for example, portamento for semitone shifts and glissando for octave or fifth shifts
5) Keyboard Zones --- each keyboard zone has its own sound

Well, I could go on and on but I think this is sufficient - comments welcome. I am eagerly awaiting my own SOLARIS and looking forward to it as is, but, I wanted to memorialize my thoughts and contributions to the idea pool here. Now, onto part-II.

PART-II: Oscillator and Filters

I would also say that having more oscillator types and algorithms (such as the DX7, SYNERGY type stuff is higher priority than other things here). In terms of oscillators, I would love to see a waveshaping oscillator of the kind found in the DWSII called the Pearl Polysensor. I played with the keyboard and it produced extremely beautiful sounds since the waveshaping functionality was modulated.

It would also be useful to have physically modelled sounds:

1) Percussive types based on a membrane model; and,
2) Resonator types (for woodwind, sax etc...)

In terms of models:

1) CS80 Oscillators
2) CS80 Filters

I can contribute real CS80 information since I own a CS80. The best vst *effort* is the ME80 by Memorymoon but there are several areas it falls short in getting the sound. Since I own a CS80, I am quite willing to do *extensive samples* or to put together a convolution profile. Due to the way the CS80 filters and voices are built, the wave shape changes as you go from end of the keyboard up to the other end: this is why Arturia CS80 does *NOT* sound like a CS80. There is also that specific way the filter responds.

I would say that a decent emulation is the ME80 (by memorymoon) - much better than the CS80 by Arturia. The ARP 2600 by WayOutWare is indistinguishable from a hardware ARP 2600 but again the Arturia is easy to identify.

The magic of the CS80 came about through three things (for your consideration):

- The two oscillators, *and* importantly, each with their own *sine* wave subharmonic generator (over which you had level control)
- the way the wave shapes change as you go from the low end of the keyboard to the upper end
- the way the filter response changes: it does not
- AND HERE IS THE BIG ONE: polyphonic aftertouch that could control amplitude, filter cutoff, the pulse widths, and LFO's speed and depth
- ALSO: each time I look that wave forms and responses: each oscillator is slightly different in waveshape - about 5-10% in the case of pulse-widths, shape of saw etc.. so that some form of randomizing element will be needed to give timbral *movement* which you get in a very subtle way on the CS80 due to its old style construction.

The best way is to model the original presets of the CS80 as accurately as possible and that will lead to everything else.

The CS80 is really about performance and with a poly keyboard like the Kurzweil MidiBoard, the Prophet T8, the Haken Continuum or the VAX77 the SOLARIS should sound awesome too (even without a CS80 model). The Continuum has a somewhat different poly mode (and that would need to be attended to - I can send you the technical write-up of it).

[marzzz]

One of the great things about the Solaris is that we will be able to set up a basic CS-80 architecture: two parallel 8-voice synths with two filters in series, and being able to add in a sine wave post-filter, then mixing everything at the end. As a former CS-80 owner, I would love to see CS-80 oscillator and filter emulations (and my VAX77 is waiting to control it). The Arturia softsynth was vastly disappointing- I have had better luck with OmniSphere (which uses actual CS-80 samples) as well as U-He's Zebra2. Haven't tried the ME80 because it is PC-only.

[John Bowen]

I’ve mentioned before that we have looked into the CS-80 filter design, and the coder told me about some interesting characteristics that he found. Hopefully the model will not be too difficult to code when the time comes.

[galaxiesmerge]

I had written the following:

"the way the filter response changes: it does not"

And I meant to write:

"the way the filter response changes: it does not change in the same way as other, for example, Moog or Arp filters --- it has a neat quality in keyboard tracking unlike others I have tried"

In the meantime, I look forward to receiving my SOLARIS when it is ready and I am excited about the future of the product.

[galaxiesmerge]

Just continuing:

1) A function to select emulated sample/bit-crushing from 24 down to 4 bits (I don't see much point in one or two bits!!!) and that bit=crush can be velocity or pressure modulated (or modwheel or ribbon or ... ) --- and that the bit-resolution be applied to "filter" (i.e. an 8 bit resolution filter working on a 24 bit resolution sound) or to "resolution" of sound samples.
2) ARP-2500 and 2600 Filters (have a look at CMS and some of the online ARP 2500 Demos) - of course, in addition to the CS80 and/or Yamaha GX1 filters and oscillators
3) Roland Jupiter-8 filters
4) Diode-Filter emulation
5) Keyboard tracking with invertibility (i.e. traditional tracking opens up a filter, invert it and it closes as you go up the keyboard)
6) Portamento that goes up but not down or downwards is glissando or, conversely, Portamento that goes down but not up, or add a split point to the keyboard where portamento on either side of the split is different
7) Dynamic program switching (i.e. the decay of the previous notes continues while you load the next patch).
Any thoughts on these are welcome --- they are just notes and ideas ... I have *never* seen anyone do an ARP 2500 filter and so this would be a *first* and it would definitely draw attention to SOLARISp (compared to say, another French manufacturer the runs on a different but related dsp platform at half the rate of the SOLARIS).

[John Bowen]

Point #1 is already available in the Solaris functionality.
Point #5 can be done, but you have to do it via Cutoff modulation by Note.
Point #7 is something that should be possible once we have Multimode done.
Point #6 - can you tell me which keyboards do this kind of glide/portamentos so I can check out these functions? These would be pretty complicated.

You have a lot of requests for additional Filter emulations!! How much would you be willing to pay for such an OS update, because it’s going to be expensive...

john b.

[galaxiesmerge]

Hi John,

The DK-Synergy had tracking portamento and I have written some Smalltalk code in Kyma (not a fair comparison) to do this. I am, first and foremost, a *player* in that it is in playing and most specifically in *performance* that music comes alive and breathes. So most of my requests are for things that enhance sound quality and performance. I am willing to spend money in modular increments, in the same way the modular synthesizer companies charge for their stuff. At the minimum, I would be willing to pay around $250 for just the CS80 and GX1 emulations, I would pay $250 for the Arp 2500/2600 emulations, and $250 for the DIode Filter/Jupiter-8 emulations (which emulates the Jupiter-4 and other of their synths). I would pay $1000.00 for all the filter emulations *and* Keyboard intelligence. In fact, for Keyboard intelligence alone, with no filters, I would pay $400.00 right now. I own the Roland-V-Synth GT, the V-Synth XT, and I tried out their new Jupiter-80 synth --- I was very, very disappointed with the Jupiter-80 performance: a performance keyboard that caters to the sound-designer and player (that's me) should focus on making ease of use (Solaris), synthesis (Solaris) and playability (Yamaha CS80 in its heyday) --- The SOLARIS can be what the CS80 was in its day, if it integrates more in the way of performance and keyboard-intelligence ---- in fact, you are welcome and free to trademark the phrase "Keyboard-Intelligence" for your product line (if you do a Google Search, you will see that your site is the first to pop up and to have that term and also, I was I believe, the first to use it, though a near neighbor on a forum used the term once to describe the failure of the Jupiter-80). So feel free to use that term for an update package.

I think that the SOLARIS will find its way to the top by offering what it does offer and by innovating in an area that is just a wide gaping market hole that eventually will converge --- I bought the V-Synth GT because of the AP, for one feature alone, even though I owned the XT already. The AP helps a lot in some certain areas, but, Roland has hardly exploited the idea. My suggestion to you is that SOLARIS offers a wealth of sound-generating options and controls for *synthesis* but what about thinking of sculpting *responses* and *behaviors*. This means, for example, having a table that couples "impact-transients" with "velocity" values. So, for example, while you hit the key harder, you actually generate *two* control values: (1) The velocity as usual; and, (2) A simulated "impact" transient. The impact transient then becomes the modulator for things that can be percussive and non-pitched or even pitched noise types as in "Drum" synthesis or synthesis of percussive sounds. This is what I mean by keyboard intelligence --- there are a whole host of ways of making the SOLARIS feel like one is playing a Stradivarius and not a stock keyboard --- the Stradivarius is an analogy for expressibility and tonal nuancing, which is where things are given life in my view.

Also, another function for the keyboard can be "sample-and-hold" of chord velocities in tight groups --- for example, if someone plays chords with the left hand, but, plays key pressure solos with the right hand, then split the keyboard so that pressure is is only acted upon by the what the right hand does and not what the left hand is doing (that way the chords don't get modulated needlessly by the soloing lead lines). You can do this by keyboard splits or zoning where the channel pressure is accorded, or, by using a sample-and-hold algorithm for when notes are clustered (which they are when you do chords) and ignoring pressure from clustered notes (being held).

Another aspect can be that you select inter-note delays: for example, when hitting a chord, instead of the notes sound together, they "strum" and by having the delays ripple left-to-right and right-to-left, you get alternative strums.

I could go on and on about keyboard-intelligence for SOLARIS but I think I will stop here

In any case, I am willing to pay and if you need help with the feature list of keyboard intelligence functionality, I am here to help. Perhaps once the requirements are done, then a better idea of cost to the customer can be assessed. Most of the coding for keyboard intelligence has to do with note-transition tables and zoning as well as derived tables for impact simulation as well as some basic musical-knowledge (i.e. that closely held notes = a chord). I oversimplify but am very keen on being as supportive as possible to the SOLARIS and for my needs, and also, to the needs I perceive my colleagues that are *realtime* players (as opposed to pattern-based sequencer jockeys --- though nothing wrong with that because you program all this keyboard intelligence in off-line, and that is what I have done). One keyboard that permits this flexibility in programming is really a keyboard synth combo which is the Haken Continnuum fingerboard and the Kyma system. However, that requires personal engineering skills and custom programming.

Can you imagine the WOW factor if SOLARIS were to have, not only what every Synth manufacturer that puts computers in boxes has, but, something that none of them cater too? You do it already with the user-interface and workflows, but how about it for the "performance" and behavior flows? This means that we are no longer in the realm of pushing around data (which is what happens with sampling and menu-driven or any soft synth or hard synth interface) but we enter the realm of "feeling" and "touch", of "character" to the player and of the notion of what it means for an instrument to be an instrument (which clearly, the SOLARIS is fulfilling with its multi-screen soft-dedicated knobs) and aliasing free behavior.

I am excited about the SOLARIS, having had the time to hear the Arturia Origin, I was again so very disappointed --- they overstate facts and the system has all kinds of aliasing (they claim it is alias free, not true).

I am continuing work on the new Album and my studio space is ready for SOLARIS when it arrives, and, I am quite happy to pay deposits for next iteration features.

Any reactions are really welcome

[lotus-eater]

It's really interesting to read different members feature requests. What you describe above as 'Keyboard Intelligence' isn't high up on my own feature request list (perhaps a few audio examples could change my mind), but I really like the idea of modular upgrades for the Solaris. I can't think of any other hardware synths that allow you to choose which new software features you'd like to add; and I think if the upgrades are going to cost extra, you need to give owners the option of choosing which ones are important to them.

I pre-ordered mainly because of my experience with the Minimax and the promise of more stunning emulations of classic synths (can't wait to try the SEM and SSM filter types), so I'm all for adding new filter models to the machine. A nice mix of classic (Jupiter 8, ARP 25/600 as suggested above), rare (Polivoks, OSCar, Synthacon) and all-new filter designs will always draw interest -- particularly from the types of people who are likely to shell out for the instrument in the first place. My wallet is cringing already.

[galaxiesmerge]

Thanks lotus-eater. The other thing about instruments that makes them special is the variances and the ability to never have the *exactly* and *precisely* the same sound. To do that, you would need keyboard-intelligence as follows: each note would have to generate random value that would function through a set of scaling tables to affect filter, pitch and envelope variances. In this way, you would replicate the naturalness of what analog circuits are: imperfect. No one wants perfect, repeated, precise and exact sound *all* the time, only *some* of the time --- you get ear fatigue hearing the same exact, precise, digital replication whereas with variances, even if these are not directly audible (indirectly audible through their interactions), that is the naturalness. At least, that's my viewpoint. I am all for filters too - don't get me wrong, but, having experience first hand what can work with note transitions and other functionality, I am now a believer that this element, which is also an element of synthesis, should not be downplayed and sometimes, like a new toy, it can be very inspiring. If I had to rank the number *one* function to add to the SOLARIS, I would say it is random variances and, then, impact models to add a second modulation parameter (for example, the impact model can affect the envelope, filter, and other things that deal with strike our percussive hit sounds).

My order of filters in order of preference would be:
(1) CS80 (Custom IC), GX1 (diode array, maybe should be part of diode models)
(2) Diode - models include the TB303, Roland 100M, Synthacon (Steiner Parker), EMS/VCS3
(3) Arp 2500 - 1047
(4) Arp 2600 - 4075
(5) Jupiter --- Custom-IC that also is the same as Jupiter-4

Cheers!

[Neotrope]

Howdy... to add in my four cents.
My main wish list for the future would be
1) CS-80 emulation, as that is my #1 "want" as far as being able to mix different tones with the rev1 Solaris; I agree, the Arturia is good but disappointing when you compare it to some of the real world CS-80 stuff... kind of a holy grail in some ways, and far less common than Prophet or Moog emulation, really.
2) duophonic aftertouch trickery as found in the Arturia (note priority aftertouch basically; high note; play chord with lower half, but playing pressure on higher note would not cause chord to modulate ... seems like something could be done in software for that ... )

Second item is not critical, since I do have a Haken Continuum, am waiting to see what the Rhodes controller will be like (if has poly-AT like one of their other models); perhaps a nextgen VAX; and of course the wild new "EVO" (http://endeavour.de/) which is pretty darn kewl, I must say. Since the Solaris responds to poly-AT, obviously a second keyboard controller for that is an obvious add-on (I sold my SCI MIDIboard, unfortunately, and then Fedex crunched it in transit... sigh). Anyway... hoping for update on the January deliveries now that NAMM is in the can for now.

Happy new year...

[John Bowen]

...and of course the wild new "EVO" (http://endeavour.de/) which is pretty darn kewl, I must say.

This is not a new idea, but it looks like a nice implementation of the Eaton-Moog Multi-Touch Sensitive keyboard, which they started working on in 1970! It was worked on for at least 10 years or more, but I did have a chance to try it at one point. REALLY difficult to learn how to play, I must say!

You can check this video by John Eaton:
http://www.youtube.com/watch?v=q0RLaNtshJ4

john b.

[John Bowen]

2) duophonic aftertouch trickery as found in the Arturia (note priority aftertouch basically; high note; play chord with lower half, but playing pressure on higher note would not cause chord to modulate ... seems like something could be done in software for that ... )

If all you wanted to do is have aftertouch affect an upper keyboard sound, but not the lower ‘chord’, then you can program the Solaris do do that.
The duophonic aspect means you can have 2 different aftertouch signals active...

John B.

[Neotrope]

Hi, John
thanks for feedback. The Evo is pretty cool, but considering it's almost as much as the Solaris, I can't see that being something on my buy list this year. For half the asking price, I'd look at it, but almost $4K with shipping to US (ouch!). Still, my custom Haken wasn't cheap either, frankly.

Duophonic aftertouch would be cool ... if possible with software, but that would just be a wishlist thing, like Artiuria have on their Origin. But will be happy to just have the beast at all when the next batch gets underway! I think I have an old Ensoniq keyboard around someplace with poly-AT, need to look for that in the closet of doom.

Really excited to see it in the studio.
C

[3rdConstruction]

Although I'm more interested in exploring new territory with the Solaris, I have to agree that emulating the character of the CS-80 is an appealing goal, and I'm sure it would get Solaris a lot of attention.

My own more humble future feature wish is for a whole lot more LFOs. Everywhere there's the opportunity to modulate with an LFO, I'd like the option to assign an individual & unique LFO to the job.

What about release velocity? Is that available in the current version? If not, that would be handy as well.

[francois]

I would also cast my vote for a CS80 osc/filter emulation, with ARP 2500/2600 coming second.