kurzweil k2700 In the Lab: Kurzweil K2700

[Stephen Fortner]

Bear with me a few paragraphs while I act like a recipe blogger and tell a rambling personal story before we get to the real topic.

 

In 1995 I was in grad school and the last new synth I had acquired — an Ensoniq VFX — was languishing unused in Vermont. Opportunities to play in bands in Santa Barbara were coming up, and I needed a keyboard. After a fair amount of shopping and bugging Sweetwater guys on the phone, I settled on a used Kurzweil K2000 from the Guitar Center in Sherman Oaks — the one on Ventura Blvd before they moved to the shinier, newer digs. I took out a supplemental student loan to cover the $1,500 price, and for that it was fairly pumped up: sampling option with extra RAM, the PRAM option for more program storage, the Orchestral ROM, and a few more things.

 

The diversity and quality of sounds I got out of that single synth blew me away at the time. I found myself playing in a startup cover band which became very popular within a year or so — if you‘re in or near Santa Barbara you probably know Area 51, which is still gigging today. Gig money let me eventually augment the rig with a second used K2000, then a Roland VK7 organ.

 

I got tons of compliments on my sounds from those who were hip to listening. When the K2661 came out I had a fair amount of writing for Keyboard magazine under my belt, and got to review it. I bought that unit and still have it. That even had the KB3 mode, letting me ditch the separate organ for simpler (read: lower paying) gigs.

 

How much I could cover on just one keyboard really drew me into the Kurzweil ecosystem. Around 2004 I even got the gig writing the instruction manual for the VA-1 virtual analog synth, a still-amazing instrument that was killed off by Samick’s abortive buyout of Kurzweil parent Young Chang. (Bonus: I got to keep the unit, I still have it, it works, and it’s one of a dozen or fewer in existence.)

 

Reviewing and gigging on Kurzweil gear also led to my knowing Dave Weiser, then a sound designer and engineer at YCRDI near Boston. He’s a regular on these forums and the guy behind a whole lot of the best vintage keyboard sounds across Kurzweil’s products. He’s done a truckload of pit orchestra programming for Broadway and similar shows, all on Kurzweil, including the 2018 TV production of “Jesus Christ Superstar” with John Legend.

 

A couple months ago, Dave rings me up and says, “Hey man, me and the people at AM&S [American Music and Sound, Kurzweil’s U.S. distributor] really want to get a K2700 into your hands.” This was sort of a full-circle moment because I had seen partial prototypes of a synth workstation meant to be the K2600’s successor during visits to YCRDI, oh, ten or more years prior. That too was benched by Kurzweil’s financial woes in the oughties, but finally, here we are.

 

So is my review unit. Let the fun begin.

[Stephen Fortner]

My goal for this review is that it (eventually) answers these questions:

• Is the K2700 a worthy heir to the K-series legacy? Does it have the same caché?
• What is there to recommend the K2700 over other recent do-it-all keyboard workstations, such as the Roland Fantom or Yamaha Montage?
• What recommends it as a main studio keyboard?
• What recommends it as a main live performance keyboard?

What I don’t want to talk about, at least not in this review:

• Do fast computers running great software make hardware workstations obsolete?

-----

Right off the bat, the value grabs attention. Many recent workstations have multiple synthesis engines (sample-based, virtual analog, FM, organ modeling, etc.) and the K2700 is no exception. But searching online retailers shows the street price of the K2700 — which at this time comes only with an 88-key weighted action — is $2,999. That’s right: three large. Even with post-pandemic inflation and supply chain issues.

 

By comparison, the new Roland Fantom-6 (61 keys, semi-weighted synth action) is $3,299. The 88-key Korg Kronos, which runs on a platform now over a decade old? $3,999. Nord Stage 3? Between $3,499 and over $4k, depending on model.

 

So, one initial thought is that before we get any deeper in this review, it would be a coup indeed if Kurzweil, who’s brand image used to be “ridiculously nice if you can afford it,” now came in as the bang-for-buck leader for the one-keyboard-to-rule-them-all musicians among us. It would be even more of a thing if it did so while delivering the sound quality and user experience that historically accounted for the “ridiculously nice” part.

 

To the end of finding out if that’s the case, I’m going to write about whatever I’m using on it or whatever is inspiring me (or pissing me off) about it at the moment. (Brother Bryce has suggested I be less formal and print-magazine-like here in the GearLab, and I’m trying.) Meanwhile, ask me anything about it, at any time, whether it has to do with what I’ve written about or not, and I’ll move answering your question to the front of the line.

 

I have to say I’m excited to be playing a Kurzweil whose model number begins with K once again.

[Al Coda]

Can´t wait for your in-depth review !

Since the K2700 was announced I´ve been interested in all it´s functionality in detail.

 

But as a long time PC3 user, I don´t trust Kurzweils hardware anymore,- almost.

I experienced issues from ribbon cables pin and socket connectors, usage of unsealed pots in pitchbend and modwheel electronics, mechanical play in pitchbend mechanism unable to eliminate by the possibly also too slow acting scanner electronics  and unexpected thermal properties too.

The latter means,- some issues only occured at (room-) temperature above 23°C and w/ my PC3.

To be precise,- deviations from normal conditions in regards of pitch-bend center value (ideally zero) in combination w/ pitchbend up and down max. values. Modwheel controller min/max valuedeviations increase(d) too when room temperature increased.

It became more obvious when using wider pitchbend ranges,- typically 7 semitone steps up/down in my case.

It´s not so dramatic when using the "standard" +/- 2 semitone steps because deviations are smaller then.

 

I use pitchbending and modwheel a lot w/ very fast movements and my left hand is always busy moving between keys and wheels,- at least for leads and basslines.

 

And shortly I received a user report of the PC4 introducing similar issues too and I´ve seen a photo of the, compared to PC3, now changed PC4 wheel mechanisms.

That looked not much better and made the report plausible.

 

So,- my 1st questions are:

 

1.)

Does Kurzweil now use sealed and better overall quality pots for pitch- and modwheel ?

2.)

Is play in wheelbox mechanics eliminated or is it now the same as in the PC4 ?

3.)

Is the scanner faster w/ significant higher update rates now ?

4.)

Became VAST envelopes and LFOs faster and is LFO aliasing eliminated ?

5.)

Received KB3 mode improvents other than "double leslie" in the FX domain,- p.ex. percussion NOT running thru C/V anymore ?

6.)

Are more "non-aliasing" VA OSCs available in VAST now ?

7.)

What about aliasing in upper keyboard ranges when using sample layers or VA hardsync OSCs ?

 

 

O.k., that´s for the time being.

Looking forward your reply,-

 

thx in advance

 

☺️

 

A.C.

 

 

 

[Stephen Fortner]

Hi Al!

 

Great questions. Some of the under-the-hood stuff will take a little time to research, but I’ve already forwarded your questions to some trusted contacts. Right now I can tell you that I’ve been beating the living daylights out of the pitch-bend and modulation wheels and have experienced no problems. The pitch-bend always has a crisp return-to-center, and I’ve been testing it with an up/down range of an octave. (For those who don’t know, yes, up and down ranges can be set separately. I personally like setting an up range of a whole-note and a down range of an octave for dive-bombs.)

 

Regarding number 5 about the KB3 harmonic percussion, I just cranked up the decay time and listened to just the percussion with no drawbars active. It sounds like it’s still running through the vibrato/chorus. That said, this is the best-sounding iteration of KB3 mode I’ve heard yet. Even the Leslie simulation sounds better to my ears. For what my $0.02 is worth, the routing of the harmonic percussion is one of those purist/realism things that’s invisible in 99 percent of real-world musical applications. Especially with fast decay and going through a Leslie anyway, one is just not going to hear the difference.

 

Number 6, about KVA oscillators. I haven’t had much time on the PC4, so I can’t speak to more or less, but I just did a quick test. I created a program with 32 layers, each layer using a KVA sawtooth oscillator as the sound source in its algorithm. The K2700 has 256 voices of polyphony, and you and I know that means audio streams, not notes. Assuming a KVA oscillator uses the same DSP resources as a multi-sample (keymap in Kurzweil parlance), 256 / 32 should have meant I could play 8 notes before hearing voice-stealing. I got more in the range of 12 to 16. I punched up the voice allocation display in Global > Tools, and watching its little Ms and Ds dance confirmed I should have only been getting 8 notes, but me ears told me things were a little better than that. (By contrast, we do know that an FM voice uses 4 voices of polyphony here.)

 

Number 7, aliasing. Once you get away from the newest sample-based sounds (mainly grand pianos and EPs), I’m sorry to say there’s a bit of it present in the top octave or octave-and-a-half. That’s because there’s a lot of legacy content in the K2700, such as many of the orchestral sounds. For gigging, I think it’s a non-issue if you’re playing actual melodies and chords in the range of the emulated instrument, but yeah, it’s there.

 

This is far less so with the KVA oscillators, but if I listen critically to that top octave, I can hear the tiniest amount. As is the nature of the aliasing beast, playing certain intervals can make this more pronounced (but again, not musically intrusive in a real-world situation) by generating sidebands. “Aliasing-free” oscillators was a bragging point of the VA-1, and KVA oscillators are presumably running the same code. Interestingly, the VA-1 itself really does sound free of aliasing, however much I “squint with my ears.” You’d think with newer DSP processors and D/A converters in the K2700, the opposite would be the case. To be fair I should clarify that nothing I am talking about here is anywhere close to the oscillations at near-dog-whistle level sort of aliasing you routinely heard in ROMplers from 20 years ago. The K2700 has a gorgeous hi-fi sound overall.

 

Number 4 — In this version of the firmware, the 6-operator FM engine has been given faster envelopes. I am told that these may migrate to the rest of the VAST engine in future updates.

 

I’ll get back to you on other under-the-hood questions. Hope what I’ve posted here so far is a help.

 

 

[jeffincltnc]

Is it also a worthwhile successor to Forte?   Forte = 13GB onboard sounds, K2700 = 4.5GB onboard sounds.  Both have user FlashPlay available, but where I think they cut back on the PC4 was in the size of the Forte pianos.   I definitely noticed that the 9 foot Steinway and 7 foot Yamaha big samples of the Forte were definitely compromised in subsequent products like the PC4.   It seems like the new pianos from the Kurzweil flagship are not up to what they had on Forte.  More polyphony, I guess.

[Stephen Fortner]

jeffinpghpa, it doesn’t sound to my ears like they conserved memory by compromising piano or EP sounds. Not sure about the PC4, but I do own a Forte7, so sometime (probably over this weekend), I can set it up and A-B them.

 

A lot of people have scratched their heads over why the K2700, the latest flagship instrument, has less sound memory than the Forte. The reason is polyphony. The Forte had 128 voices, and on the K2700, they couldn’t figure out how to get 256 voices to address the same amount of memory. So they made a trade-off and chose polyphony. Honestly, I might have gone the other way, but I can see the reasoning.

 

The Forte7 was a great size and weight for gigging — the K2700 is 11 pounds heavier (41 vs 52). Sadly, the Forte line appears to have been discontinued. On a retail site that will remain nameless, I just saw a demo 88-key unit advertised for $4,500, and used Fortes on Ebay and Reverb are hitting at about $3,000. That speaks to the demand for them.  

[ElmerJFudd]

11 hours ago, Stephen Fortner said:

jeffinpghpa, it doesn’t sound to my ears like they conserved memory by compromising piano or EP sounds. Not sure about the PC4, but I do own a Forte7, so sometime (probably over this weekend), I can set it up and A-B them.

 

A lot of people have scratched their heads over why the K2700, the latest flagship instrument, has less sound memory than the Forte. The reason is polyphony. The Forte had 128 voices, and on the K2700, they couldn’t figure out how to get 256 voices to address the same amount of memory. So they made a trade-off and chose polyphony. Honestly, I might have gone the other way, but I can see the reasoning.

 

The Forte7 was a great size and weight for gigging — the K2700 is 11 pounds heavier (41 vs 52). Sadly, the Forte line appears to have been discontinued. On a retail site that will remain nameless, I just saw a demo 88-key unit advertised for $4,500, and used Fortes on Ebay and Reverb are hitting at about $3,000. That speaks to the demand for them.  

If the Forte was a slow seller, it’s not easy to determine what feature choices, UI choices, expandability, sounds, programs, look/design, etc. etc. are the driving force.  It may be related to the Forte being seen as more a gigger’s board vs their past success with the K line that had a production/studio vibe to them.
 

However, price is most likely a factor. Korg clearly feels in the current market the Nautilus needed to come in well under $3k (the 88k is advertised at $2799), and they don’t dilly dally around about offering it in 76 and 61k variants (though I’m told Kurzweil historically sells a lot less 76 and 61k instruments than 88k).  
 

With nostalgia being a big chunk of the MI market lately, a return to the K series was probably an excellent idea. The question is did they hit the right price point ($2999 street) and does it sound great in the areas that current buyers are hot for. 

[Doerfler]

liking this thread so far. I was a salesman at Washington Music Center 30 years ago when the K2000 was released. Great memories

[Delaware Dave]

i recently saw this video and was surprised by how little difference there was between the sound quality of the K2700 vs. the K2600.  

 

 

[Stephen Fortner]

Great share Delaware! Even with compressed YouTube audio, I believe I can hear that the D/A converters on the K2700 are smoother, but the two are certainly very close.

[Al Coda]

On 2/24/2022 at 6:34 PM, Stephen Fortner said:

Hi Al!

 

Great questions.

 

Hello Stephen ! Thank you !

 

On 2/24/2022 at 6:34 PM, Stephen Fortner said:

Some of the under-the-hood stuff will take a little time to research, but I’ve already forwarded your questions to some trusted contacts.

 

Great idea !

 

On 2/24/2022 at 6:34 PM, Stephen Fortner said:

Right now I can tell you that I’ve been beating the living daylights out of the pitch-bend and modulation wheels and have experienced no problems. The pitch-bend always has a crisp return-to-center, and I’ve been testing it with an up/down range of an octave.

 

That´s indeed good to know !

Since when did you use it ?

With my PC3, the issues started about half year after I bought it new.

1st the pots crapped out, had been replaced by the original replacement parts,- same story after about half a year.

Then we replaced pots by better and way more (EUR 10,-) expensive sealed ALPS pots which fixed the issue of early dirty worn out pots almost permanent.

The trimmer for PB dead-zone was also replaced by a very high precision one, but this didn´t fix the PW - center issue because of the small play between wheel and spring, which in addition to pot quality caused deviations.

 

So, I´m pretty sure, when someone buys a K2700 new,- it will work perfect for a period of time, but I also fear it MIGHT show issues possibly early again.

 

I really want a K2700, last but not least as a improved replacement for my PC361 and because I love Kurzweil keys for it´s ultra-flexible MIDI controller abilities.

 

On 2/24/2022 at 6:34 PM, Stephen Fortner said:

 

Regarding number 5 about the KB3 harmonic percussion, I just cranked up the decay time and listened to just the percussion with no drawbars active. It sounds like it’s still running through the vibrato/chorus. That said, this is the best-sounding iteration of KB3 mode I’ve heard yet. Even the Leslie simulation sounds better to my ears. For what my $0.02 is worth, the routing of the harmonic percussion is one of those purist/realism things that’s invisible in 99 percent of real-world musical applications. Especially with fast decay and going through a Leslie anyway, one is just not going to hear the difference.

 

Well, yeah,- it is somewhat nitpicking, but when it comes to organ, I also like to play 8884 2nd perc / long decay, C3 and NO leslie at all, just only an ampsim.

And in that case, it can be recognizable, not for the band or audience, but for the performer.

To be true,- it cann be a nice effect and I´m not a purist at all,- so I´d wish, "Perc thru CV" were user selctable !

But IIRC, i the PC3 series using MARA Dsp, it wasn´t possible by chip architecture itself.

 

☺️

 

A.C.

 

 

 

[Al Coda]

On 2/24/2022 at 6:34 PM, Stephen Fortner said:

Number 7, aliasing. Once you get away from the newest sample-based sounds (mainly grand pianos and EPs), I’m sorry to say there’s a bit of it present in the top octave or octave-and-a-half. That’s because there’s a lot of legacy content in the K2700, such as many of the orchestral sounds. For gigging, I think it’s a non-issue if you’re playing actual melodies and chords in the range of the emulated instrument, but yeah, it’s there.

 

In my 1st post you reply to, I didn´t mean ALL the samples, it was more about the sampled synth waveforms which were often used for synth patches in the PC3 as also KORE64 expansion.

By nature, it´s imposible playing a sampled waveform over wide keyboard ranges.

The other and much more important source was the KVA sync algo/waveform consumating as much Dsp blocks as the MOOG filter emulation.

Most of the factory hard-sync sounds alias like hell in PC3.

I have to mention I was able programming some which didn´t alias that much by using custom algorithms and inserted some carefully adjusted 1 Dsp-block LP filters, sometimes also in combination w/ "RollOff" Dsp at the right places within VAST architecture.

 

So, It might be matter of patch programming and improvement of VAST itself to eliminate aliasing.

 

On 2/24/2022 at 6:34 PM, Stephen Fortner said:

“Aliasing-free” oscillators was a bragging point of the VA-1, and KVA oscillators are presumably running the same code.

 

In PC3 VAST KVAmode, I differenciate between the "aliasing free" OSCs and these using less Dsp blocks and alias.

Sawtooth is the most aliasing free, w/ PWM you already enter aliasing territory and "sync",- even consumating the max Dsp blocks,- aliases the most.

Now, for sawtooth, square etc. you also find less Dsp consuming OSCs in VAST, and they ALL alias.

 

And as soon as you used WARP Dsp or any LFO speed higher than 6.5Hz, the LFOs alias too.

You hear w/ pitch modulation in high ranges, even vibrato (w/ Modwheel).

Aliasing also appeared when using distortion on DSP level or as FX.

 

Much to test w/ the K2700, which by nature needs time,- so many users, just only being preset jockeys, might not recognize early enough or at all.

 

As a consequence, the question to the R&D should be "is the VAST technology now improved at all or is it identical to the PC3 series?"

And I don´t mean new samples or not.

 

I won´t buy a K2700 as a piano, EP and organ.

It´s good to have these, the orchestra´s samples and KB3, but for me it is a MIDI controller and sophisticated flexible synth engine too,- which unfortunately imported some of the disadvantages from K2000 times already.

 

On 2/24/2022 at 6:34 PM, Stephen Fortner said:

You’d think with newer DSP processors and D/A converters in the K2700, the opposite would be the case. ... The K2700 has a gorgeous hi-fi sound overall.

 

Sounds promising and I´d highly appreciate when new (LENA) DSPs and D/As eliminated the for me in PC3 very obvious aliasing when playing factory synth programs using sync.

 

On 2/24/2022 at 6:34 PM, Stephen Fortner said:

 

Number 4 — In this version of the firmware, the 6-operator FM engine has been given faster envelopes. I am told that these may migrate to the rest of the VAST engine in future updates.

 

Well, that´s really good news !

 

 

On 2/24/2022 at 6:34 PM, Stephen Fortner said:

 

I’ll get back to you on other under-the-hood questions. Hope what I’ve posted here so far is a help.

 

Yes it is,- thank you !

 

☺️

 

A.C.

 

On 2/24/2022 at 6:34 PM, Stephen Fortner said:

 

[Stephen Fortner]

Controls

 

Still working on the penetrating questions (I see you Al) and the A-B setup with the Forte, but I wanted to write a little interlude about controls. The K2700 has every control I ever wished for on previous K-series instruments and, what’s more, they’re all mapped to useful parameters. This was less the case on the Forte and certainly the K2600s I’ve had. Adding knobs above the nine faders was certainly a necessary move.

 

Left to right, the first five faders control filter cutoff, resonance, volume envelope attack, volume envelope release, and something called “motion amount,” which is program-dependent and has to do with modulations. The final four are dry/wet mixes for effects — often distortion, chorus, delay, and reverb but changing depending on the program (phaser instead of chorus on a Rhodes, for example). 

 

The first four knobs control a three-band EQ with a sweepable mid band. Each of the next five controls a key parameter for the effect on the fader just below it: rate for modulation effects like chorus, distortion warmth, delay and reverb time, etc.

 

Of course you can re-assign all these to any parameter you want at the program level, but the point is, you don’t have to. Filter cutoff and res, and envelope attack and release are the first things I reach for in live performance when I want to change the character of the sound but not to the extent that I need to change programs. These all seem to have the faders scaled to what the sound designers think is a useful range for each program. I can see their thinking here, but you can’t necessarily take, say, a pad sound and pull the attack all the way down to where it’s Minimoog-crisp — at least not without editing that range in the program. I’d just as soon have each fader sweep the full 0-127 range. I prefer cars with manual transmissions, too.

 

The nine buttons below the faders are also all mapped. Some bypass/enable the EQ or the effect in each one’s control column. On most programs, buttons 2 and 3 do things called Presence and Impact, respectively. I haven’t yet traced what these do on most Programs, but on a piano sound, for example, Presence boosts the mids and high-mids and Impact adds an edge to the attack, as if the sound were going through a compressor. Both could be very effective when it’s your turn to solo. Button 4 is Freeze, a feet-free sustain pedal.

 

Wiggle or press anything, and the display momentarily tells you what it does. The Variation button above the master volume fader often adds a pad or other secondary layer.

 

Of course, this is all different with KB3 organ Programs, with the buttons controlling the B-3’s signature harmonic percussion, vibrato/chorus, and key click; the Variation button defaults to your Leslie slow/fast switch.

 

The sound selection buttons are in category mode by default. Pressing the View soft button below the display cycles between single view, list view, and Quick Access banks of eight programs, which appear as a grid on the bottom of the screen. (Quick Access banks are user-definable “set lists” that can contain both single Programs and Multis. You simply press the corresponding number button to bring up the Program or Multi.) Or, press the Keypad button and access your Programs and Multis by punching in their numbers because you’ve memorized them, as longtime Kurzweil users tend to do.)

 

Arpeggiator buttons, a tempo knob and tap button, and sequencer transport are on the right, as are 16 velocity-sensitive trigger pads. (I will devote a separate post to their various functions, but they trigger programmable chords by default.)

 

Big props for Edit and Exit buttons I don’t have to squint to find, and for nearly every button being backlit.

 

On the left we have gain knobs for the two XLR combo audio inputs, a button to switch input 1 to Hi-Z instrument impedance, and a button to turn on 48V phantom power for both. This is very handy given that the K2700 functions as a USB audio interface at sample rates up to 192kHz, and given that you can process external audio through the K2700’s effects. That said, that there’s no separate volume control for audio streaming in over USB is a significant omission IMHO. The parameter does exist, in the Audio I/O menu of the Global settings mode, but if anything should have a physical widget, it’s this. Both the Yamaha MODX and Roland Fantom have USB volume controls.

 

The other nit I’d pick is the display. On the plus side, you get a knob on the back for brightness, not a menu item. It’s crisp and high-contrast. But the 7" touchscreen panel common to the Montage, MODX, and Fantom is becoming an industry standard; its resolution is WVGA at 800x480 pixels. The current Kronos has an 8" touchscreen at 800x600 (SVGA). The K2700? Not a touchscreen, 4.3", and 480x272. To be fair, the K2700 modes and menus don’t lend themselves to touchscreen operation as you can see the family resemblance with even a K2000, and I get choosing consistency over reinventing the user interface wheel.  

 

Ok, half-past midnight and time to relax with however much of an obscure horror movie on Shudder it takes to get me to sleep. More in the next day or so, getting into sounds!

[Al Coda]

 

On 2/25/2022 at 5:10 PM, Delaware Dave said:

i recently saw this video and was surprised by how little difference there was between the sound quality of the K2700 vs. the K2600. 

 

 

 

 

I´m shocked the Rhodes, beginning at 03:25 and the following phased Rhodes sound better in the K2600 !

 

🤔

 

A.C.

[Mighty Motif Max]

On 2/25/2022 at 10:10 AM, Delaware Dave said:

i recently saw this video and was surprised by how little difference there was between the sound quality of the K2700 vs. the K2600.  

 

 

Something to note - for the pianos and piano layered patches, that demonstrator is using the old Triple Strike samples. They're still in the K2700 but there are much better German 9ft and Japanese 7ft grand pianos - he is just using the carried over Triple Strike patches.

[Stephen Fortner]

Piano and EP Sounds — First Impressions

 

Hi friends. I got word from Kurzweil that the K2700 contains 406 new Programs and 721 new Multis. I am still not quite clear on how much multisample content — i.e. stuff based on new recording sessions — is involved. But as soon as I know, I will report. The ears say that compared to my Forte7 and certainly to my most recent Kurzweil acquisition, an SP6, there is a lot going on here. Keymaps, velocity breakpoints, sample loops, and the like have all been tweaked and optimized for the K2700 and its Fatar TP-40L keybed.

 

What I can say is that the range and quality of sounds leaves me wanting for just about nothing so far. I’m going to take them in tapas-sized chunks that are easier to read — and to write.

 

As Max mentioned, the triple-strike pianos that were Kurzweil’s signature for so long have been surpassed by German D and Japanese C7 sample sets that have many more velocity layers. We all know that means Steinway D and Yamaha C7, but of course another manufacturer can’t use those terms.

 

The “location 1” piano is called Bristol Piano, based on the German D sample set and enhanced with sympathetic resonance and sustain resonance. (Refresher: Sympathetic resonance is the vibration strings from held notes on the keyboard in response to other new notes played. Sustain resonance is what happens when the damper pedal is down and all strings are somewhat affected by the ones actually hit by the hammers.)

 

“Bristol Grand” is perhaps one of the best-balanced workstation piano sounds I’ve ever played. It has just enough high-end sparkle to be useful in solo passages, thunderous bass, and a sweet, singing midrange. Onstage with rock instrumentation, however, I’d reach for the aptly named “Sizzle Grand,” which also uses the German samples. From the first note I struck on “Bright Jazz 9ft” I immediately wanted to play Vince Guaraldi tunes, and the odd thing is that it too uses the German Samples and Guaraldi recorded A Charlie Brown Christmas on a Yamaha C7 at Fantasy Studios in Berkeley, CA. (Fun brag: I got to play that actual piano once, and you’re damned right I tore into “Linus and Lucy.”) Nothing that used the C7 sample was closer.

 

As far as I can tell, Programs 1-23 are where piano patches new as of the K2700 live. After that I began to see some familiar names, beginning with a couple of K2600 triple-strike grands.

 

Kurzweil EPs have always been some of my favorites to play since the early 2000s when the ROM4 expansion for the K2600 came out. There was not only body and realism, but just the right amount of randomness about the balance of tine and tonebar you might hear on any given note while playing a Rhodes patch. A lot of that was the Weiser factor, and the Programs only gotten better since. So have everyone else’s beginning with the original Yamaha Motif, but there’s still something about the Kurzweil sound when it comes to vintage electro-mechanical keyboards.

 

“Steely Dyno” crushes my benchmark: the phaser Rhodes characteristic of Steely Dan and the Doobies’ “Minute by Minute.” Beyond that, any Suitcase or Stage character you could want is on hand, effected and not, bright and dark, more tine or less. Likewise for the Wurly, whether you’re looking for Ray or Supertramp or anything else. 

 

For EPs, 49-62 are seemingly the Program locations for the new stuff; after that things start to look and sound familiar again. But there are plenty of gems among those legacy sounds, including some CP electric grands that are still more vibey and evocative of Peter Gabriel, the Buggles, and Simple Minds than those in any other keyboard — I’m looking at you, Yamaha.

 

Clavinets mostly match those in my Forte but are still excellent. They get the key release noise right, and variations are available to cover the pickup rocker switch settings on the real deal. On the Clavinet C and D6 models, the A-B rocker chose between two pickups and the C-D rocker chose whether one or both were heard. They were inter-dependent, with the four combinations creating very different tones, like so:

• AC: Harp pickup only.
• BC: Bridge pickup only.
• AD: Both pickups, in phase.
• BD: Both pickups, out of phase.

Point being, the K2700 has Programs to cover these, as well as a bunch designed to sound like the Clavs on famous songs and named appropriately.

 

Even entry-level keyboards have EPs and Clavs that don’t leave you wanting for realism, and then there’s the whole world of plug-ins. So it’s impossible to say whether those in the K2700 are “best” by any objective standard. But to me at least, they are special. What’s special about them is that dialing up any Program and immediately makes me want to play this or that song. Practically speaking, that’s a cover band warrior’s playground, and I’d say the same thing about the synth sounds in the K2700, which we’ll get to. Stay tuned!

 

 

[Stephen Fortner]

Intermezzo — Programs and Objects Lists

 

For the curious and the obsessed like me, here are a coupe of Excel spreadsheets you can also download from the Kurzweil website, but I figured it would be nice to have them at arm’s reach here. One is a list of sound Programs with a different sheet tab for each category; the other is a multi-sheet workbook of every type of memory object in the K2700: Programs, Multis, keymaps, etc. K2700-Complete_Object_Lists_1.04.1.xlsK2700-Programs-Categorized-Objects-1.04.1.xls

[Stephen Fortner]

New Programs List

 

Ok, it appears I was a bit off the mark about what Programs are spanking new in the K2700. They sounded new, and are IMO some of the star sounds in the instrument. Attached is a definitive list from the software manager at YCRDI. If it’s on here, it's new.

 

My source there also gives this summary of the samples (keymaps), comparing recent products to the K2700:

 

Forte SE, SP6, PC4, and PC4SE vs. K2700

 

• No German 9-foot from original Forte (not SE).
• Carries over most Forte samples but uses fewer samples per velocity layer (depending on instruments) to accommodate Flash memory size.
• Fewer orchestral percussion samples.

K2700 vs. previous workstations and stage pianos

• Reintroduces 1977 EP keymaps.
• Reintroduces Forte German 9-foot keymaps.
• Reintroduces reed EP (Wurly) keymaps from Forte.
• Reintroduces Forte pitched orchestral percussion samples such as chimes, bells, and glockenspiel.

I think it’s safe to interpret this to mean that no sounds in the K2700 are based on new sample recording sessions. That in itself is not a bad thing, as the Forte samples were damned good to begin with and, to my ears, a lot has been done in terms of keymap editing (layers, making loops more seamless, etc.) and VAST programming to make these sounds really shine. At the same time I can see the desire for some all-new content in a keyboard this long-awaited.

 

 

 

 

K2700 NEW programs_multis - Multis.pdf

[Stephen Fortner]

KVA Synth Sounds

 

KVA Programs use the modeled analog waveforms originally developed for the unicorn VA-1 synth, with the benefit of 2020s processing power and D/A conversion. Before we get there, I want to drop a brief comment on all the synth Programs that are based on sampled keymaps. The variety is too wide to cover in that old Keyboard magazine let’s-go-through-’em-all sort of way. The quality is impeccable. Programs that are supposed to sound analog are on par with the best virtual analog synths from Nord, Modal, and other high-end brands.  There is something here for every taste and every musical application: leads, pads, comping. In fact, I prefer these overall to what I can get out of most virtual analog (read: modeled) synths, sample-based though they may be. On the more sparkly side, tons of pads, soundscapes, and electronic film score fare feature internal motion and evolution courtesy of the arpeggiator, riff generator, or just clever VAST programming. There’s also a six-operator FM engine that can load original DX7 patches as Sys-Ex, but we’ll get to that in its own post.

 

The KVA waves take it up a notch and are incredibly smooth. Programs that use them don’t live in any one place, though there are several sections of contiguous Programs with “VA1” in the name. The sure way to know a Program’s sound source is to hit the Edit button and then choose the Keymap page (via a soft button below the display).  If you see “999 Silence” in the Keymap field at the top left, then that Program — more correctly, that layer of that Program — is not using a sample-based Keymap. It’s using a KVA oscillator or an FM layer.

 

This actually brings me to something we need to cover before I get any deeper into the VA stuff: the whole architecture as regards VAST vs other stuff the K2700 can do.

 

 

[Stephen Fortner]

To VAST or Not To VAST: K2700 Sound Architecture

 

Dating back to the K2000, VAST (Variable Architecture Synthesis Technology) can really be thought of as the first virtual modular synth to exist in hardware. Back then, the raw materials for making patches were just sample-based Keymaps, and VAST offered a deep and powerful way to get maximum musicality, realism, and sonic variation from the samples the ROM and RAM amounts of the day could hold. You had a bunch of Algorithms, which were basically signal chains with different numbers and arrangements of modules. You could dial various processors into each module such as filters, modulators, math functions that turned two inputs into a different output, and so on.

 

The K2700 still uses VAST as its main sound engine, but it does other stuff, too: KVA waveforms, KB3 drawbar organ mode, and FM.

 

Here’s where things get a bit interesting and possibly confusing for the keyboardist who didn’t come up on Kurzweil. When in a company’s marketing materials, you see something like “It has all these multiple synth engines,” you expect that which synth engine a sound patch/program is using is going to be some kind of top-level choice. That is, you start there. The Korg Kronos handles this in a straightforward way. When programming a new sound from scratch, the first question you answer is, “Do I want the SGX piano engine, the CX-3 organ engine, the HD-1 sample engine, the STR-1 string modeling thingy, the Polysix emulation, the AL-1 analog modeler, or something else out of the nine in total?” If you want to combine two or more different engines, you split and layer Programs in a Multi. It’s all intuitive and sensible.

 

Kurzweil does things a bit differently in the service of flexibility, and it can be a head-scratcher at first. Before we even get to Multis, a single Program can have up to 32 layers. By default, these are VAST layers with all those algorithms and modules in them, and they draw from the sample ROM as their main source ... except when they don’t.

 

The KVA oscillators live inside VAST as an alternate sound source. Keymaps refer only to sample-based stuff, so to construct a KVA layer, you punch up “999 Silence” in Edit mode on the Keymap page, then go to the Algorithm page. There, you find the block just to the right of Pitch (always the first block in the chain), cursor to it, and dial in one of the KVA waves, which include the usual analog synths, a Roland-style supersaw, and more goodies.

 

The FM engine lives next door  to VAST as its own sort of layer. The top-level Edit mode page has a “New FM Layer” option, which is the quickest way to see what one looks like as opposed to trolling through Programs to find those with FM layers — though the DX-style EPs are a good place to start!  Do that, and you’ll then have access to the FM Main page via the display’s soft buttons. Punch that up, and look — it’s a DX7 with the 32 original FM algorithms! What you won’t see on this layer are the usual VAST parameters like the ALG and DSPCTL pages. The idea is, you can freely combine VAST and FM layers within a program.

 

KB3 organ Programs live across the street from VAST. They’re their own thing, more of a top-level sound engine choice like in the Kronos. That’s because they’re re-tasking a bunch of the K2700’s DSP resources to emulate tonewheels and other aspects of the Hammond B-3 organ. So, you can’t have KB3 layer and a VAST layer in the same Program — but you can combine KB3 and non-KB3 Programs in a Multi, subject to polyphony constraints because KB3 eats more voices than VAST.

 

Confused yet? Here’s a visual of the hierarchy:

 

 

Other than proving why I’ll never work as a graphic designer, this shows in a binary way how the objects nest inside each other. The point being, you start in a very different place depending on the kind of sound you have in mind at the outset. Of course, you could play the excellent factory presets and just tweak when you need to, but among other things the K2700 is supposed to be a sound designer’s machine.

 

To experienced Kurzweil users, this may be all “No duh,” but to a newbie saying “I want to make a sound out of those virtual analog waves I keep hearing about because that’s, like, cooler than samples,“ there’s an undeniable learning curve. I say this not to bash Kurzweil at all — the architecture of the K2700 is incredibly powerful and this UX makes sense if you’re thinking from the guts of the instrument out. You can also draw a clear through-line from the K2000’s menus and pages to where we are right here, and there’s something to be said for consistency.

 

But if Kurzweil were to put me in charge of UX design for future products, I’d shoot for one that can access this same architecture while giving the musician a clearer and more “flowcharted” ride. Having done things as seemingly (but not really) straightforward as writing synth instruction manuals, I’m quite aware of what a big, hairy deal that would be, so I’m not saying the K2700 is bad for having the organizational scheme it does.  Moreover, the performance aspect of the UX, covered in the post above about controls, is mostly brilliant.

 

Rather, my main hope for this post is that it offers a bridge between the multiple sound engines described on the K2700 product webpage and where you actually find, instantiate, and edit this stuff once you have the beast out of the box.

 

Now that that’s out of my system, we can get back into how this synth sounds and plays.

[Al Coda]

On 3/4/2022 at 5:05 AM, Stephen Fortner said:

KVA Synth Sounds

 

 If you see “999 Silence” in the Keymap field at the top left, then that Program — more correctly, that layer of that Program — is not using a sample-based Keymap. It’s using a KVA oscillator or an FM layer.

 

On the PC3 that also ruled for layers w/ the keymap set to "none".

Did they ditch that ?

 

The programs done w/ sampled synth waveforms,- are they identical w/ the ones being introduced in PC3 and KORE64,- or are there lot of new ones ?

I own KORE 64,- and most of the synth programs are crap,- sorry.

Doesn´t mean the waveforms were crap as also not the DSP possibilities coming w/ VAST,- but the programming ... 😒

I remember,- in the other Kurz related forums, there were many thinking the same about.

 

☺️

 

A.C.

[Al Coda]

On 3/4/2022 at 4:24 AM, Stephen Fortner said:

• Reintroduces 1977 EP keymaps.
• Reintroduces Forte German 9-foot keymaps.
• Reintroduces reed EP (Wurly) keymaps from Forte.
• Reintroduces Forte pitched orchestral percussion samples such as chimes, bells, and glockenspiel.

 

is.pdf 90.99 kB · 2 downloads

 

This is really good news !

 

☺️

 

A.C.

[Al Coda]

On 3/4/2022 at 6:09 AM, Stephen Fortner said:

To VAST or Not To VAST: K2700 Sound Architecture

 

- but you can combine KB3 and non-KB3 Programs in a Multi, subject to polyphony constraints because KB3 eats more voices than VAST.

 

 

 

IMO, the most wanted feature would be 2 KB3 programs in a multi and together w/ a single FX chain for C/V, preamp- and poweramp- overdrive and rotary sim,- for a KB3 dual manual organ !

The rotary sim needs mic distance parameter instead the individual volume parameters for each virtual mic,- or even better,- both, for quick or in-depth editing.

"Double Leslie" only masks the weaknesses of single leslie sim and is also not the simulation of 2 independent Leslies receiving individual input signal and being set up in different positions in a room or on stage.

Single leslie needs optimization !

Are there improvements like more/less/ other parameters available in K2700 to get rid of the pitchiness of the rotor sim ?

For the leslie sim, there are so many parameters available which are unusable for the average user unless he´s willing to investigate the complexity of a mechanical leslie in depth.

I was, but never came to perfection.

It´s not really user friendly.

 

☺️

 

A.C.

[Stephen Fortner]

Hi Al and everyone else,

 

Al’s questions are timely because next up I’m digging into KB3 mode. Rather fried after a family-filled weekend but I’ll start fresh tomorrow. Kurzweil has also provided a way for me to take screenshots right off the K2700, so we’ll have more visuals. 

 

Meanwhile, brother Jim Alfredson posted this comparison video on FB and I wanted to share it here. From Anderton’s music in the U.K. (no relation to Craig) and rather unscientific but still shows that the K2700 sounds really damned good in the current field of workstations. Just from what’s here, I am really digging the sound of the Kurzweil, especially in the piano, EP, and synth categories.

 

 

 

I mentioned in a previous post that the K2700 has a lot of legacy content in terms of samples. To be fair, all current workstations do. In fact, the last entirely new recording sessions of any significance, at least that I’m aware of, were some of the orchestral sounds for the Yamaha Montage, done by players for the Seattle Symphony.

 

Vertically through a manufacturer’s product line, you have a master sample set that gets parsed out differently according to the product price point, available memory, processor, etc. That’s pretty much how things work across the board. That’s not necessarily bad, as the “gold master” of a well-recorded sample set can have a lot of longevity.

 

Ok, more on the way!

[Al Coda]

 

 another one ...

 

☺️

 

A.C.

[Stephen Fortner]

KB3 Organ Mode, Part 1 — First Impressions

 

KB3 has been around since the mid-to-latter life cycle of the K2600, and as far as I can hear, it reaches a new level of realism and optimization in the K2700. If you’re running a recent dedicated clone (XK5, SK series, Mojo, Legend, Numa, etc.) through a real Leslie or high-end sim like a Ventilator, then the  Hammond/tonewheel sound is central to your work and tastes, you’ve invested a lot of time and research in nailing it, and you may be less than blown away here.

 

 

That said, the fairer standard for evaluating KB3 is: drawbar organ modes that are part of general-purpose workstations. In the current field we have the Korg Kronos and Nautilus, the Roland Fantom, and KB3. Am I missing anything? The Yamaha Montage (and Motif XS/XF before it) had programs that could bring in individual drawbar tones on the sliders, but these were sample-based. IIRC, one or more of the sample layers (called Elements in Yamaha land) on a program might have nuances like key click or tonewheel leakage baked in. What they weren’t doing is using DSP to model tonewheels and all the quirks that come with them, not to mention the B-3 vibrato/chorus scanner, etc. 

 

By that standard, I find KB3 to have a lot of, well, balls. The Roland and Korg sims are based on those developed for their VK/VR series and (new) CX-3, respectively, and I’ve always found KB3 to have a more depth. On the one hand, it’s a little less polished, more raw, and perhaps favors the midrange a bit, but at gigs I just get more of what I want out of it. (Side note: A shootout between all three workstation organ sims would be fun. I have two out of three here already. Hmm.) The Leslie simulation in particular retains a good sense of motion without layering the vibrato-chorus on top of it — I’ve always found that on the Roland and Korg stuff, this resulted in a high end that sounded more like it was panning between the speakers than moving around in a circle and bouncing off the walls, which is literally what a Leslie does to a signal. YMMV.

 

 

In fact, higher drawbar frequencies at fast Leslie speed is sort of my acid test for simulations, as those frequencies are more directional and it’s harder to fool the ear into thinking they’re being spun around. Whatever else KB3 gets wrong, it gets this right IMO.

 

Here are three quick clips I recorded using different drawbar registrations from the KB3 Program “Glow Tube Gospel.” I found that in KB3 mode, the Sustain pedal changed to a Leslie speed switch without my doing anything, by the way. These are all 16-bit, 48kHz WAV files.

 

Full drawbars out:

 

 

“Erroll Garner” (800008888):

 

 

Jazz organ (8684 with 3rd harmonic percussion):

 

 

Yes, the overdrive on this particular Program is a bit more “hairball”-y than I’d prefer, but I hope these little noodles convey the tone and the rotary sim.

 

Next posts, we’ll try out more KB3 sounds and dig deeper into the KB3 mode parameters. Al C, I will try and answer your questions about the double-rotary effects chain at that time.

 

 

 

 

[Delaware Dave]

On 3/8/2022 at 11:33 PM, Stephen Fortner said:

KB3 Organ Mode, Part 1 — First Impressions

 

KB3 has been around since the mid-to-latter life cycle of the K2600, and as far as I can hear, it reaches a new level of realism and optimization in the K2700. If you’re running a recent dedicated clone (XK5, SK series, Mojo, Legend, Numa, etc.) through a real Leslie or high-end sim like a Ventilator, then the  Hammond/tonewheel sound is central to your work and tastes, you’ve invested a lot of time and research in nailing it, and you may be less than blown away here.

 

 

That said, the fairer standard for evaluating KB3 is: drawbar organ modes that are part of general-purpose workstations. In the current field we have the Korg Kronos and Nautilus, the Roland Fantom, and KB3. Am I missing anything? The Yamaha Montage (and Motif XS/XF before it) had programs that could bring in individual drawbar tones on the sliders, but these were sample-based. IIRC, one or more of the sample layers (called Elements in Yamaha land) on a program might have nuances like key click or tonewheel leakage baked in. What they weren’t doing is using DSP to model tonewheels and all the quirks that come with them, not to mention the B-3 vibrato/chorus scanner, etc. 

 

By that standard, I find KB3 to have a lot of, well, balls. The Roland and Korg sims are based on those developed for their VK/VR series and (new) CX-3, respectively, and I’ve always found KB3 to have a more depth. On the one hand, it’s a little less polished, more raw, and perhaps favors the midrange a bit, but at gigs I just get more of what I want out of it. (Side note: A shootout between all three workstation organ sims would be fun. I have two out of three here already. Hmm.) The Leslie simulation in particular retains a good sense of motion without layering the vibrato-chorus on top of it — I’ve always found that on the Roland and Korg stuff, this resulted in a high end that sounded more like it was panning between the speakers than moving around in a circle and bouncing off the walls, which is literally what a Leslie does to a signal. YMMV.

 

 

In fact, higher drawbar frequencies at fast Leslie speed is sort of my acid test for simulations, as those frequencies are more directional and it’s harder to fool the ear into thinking they’re being spun around. Whatever else KB3 gets wrong, it gets this right IMO.

 

Here are three quick clips I recorded using different drawbar registrations from the KB3 Program “Glow Tube Gospel.” I found that in KB3 mode, the Sustain pedal changed to a Leslie speed switch without my doing anything, by the way. These are all 16-bit, 48kHz WAV files.

 

Full drawbars out:

 

 

“Erroll Garner” (800008888):

 

 

Jazz organ (8684 with 3rd harmonic percussion):

 

 

Yes, the overdrive on this particular Program is a bit more “hairball”-y than I’d prefer, but I hope these little noodles convey the tone and the rotary sim.

 

Next posts, we’ll try out more KB3 sounds and dig deeper into the KB3 mode parameters. Al C, I will try and answer your questions about the double-rotary effects chain at that time.

 

 

 

 

 

Steve, can you advise whether the effects being used on the full drawbars out leslie is a single leslie vs. a double leslie?  if there are two KB3a and two KB3ba effect boxes in the effects chain then it is a double leslie, if one of each then it is a single leslie.

[Stephen Fortner]

Hi Dave,

 

It appears to be a double. Here is the first screen shown when you select the FX edit tab:

 

 

Press edit again to get the FX chain, and you see this:

 

 

More soon!

[Al Coda]

30 minutes ago, Stephen Fortner said:

Hi Dave,

 

It appears to be a double. Here is the first screen shown when you select the FX edit tab:

 

 

Press edit again to get the FX chain, and you see this:

 

 

More soon!

 

Thx for confirmation !

 

And I see the 222 StChorus in addition to the "double leslie".

 

☺️

 

A.C.

[Al Coda]

Here´s a link according to pitchbend mechanics design flaw in PC4 ...

https://www.keyboardforums.com/threads/pc4-pitch-wheel.32785/#post-220891

 

It´s already a different design compared to the PC3 which also introduced pitchbend problems.

Now I want to know if the same unit is present in K2700 models.

 

By interest, I bought a used DX7 left-hand controller unit which is well known for it´s reliability (I owned DX7 and,- since decades,- still own a DX7mkII and KX76)

It´s soooooooo simple and never fails.

I don´t understand why a company like Kurzweil, offering great and complex technology, is unable using such super simple, but insanely effective design.

The left-hand controller unit is so important for expressivity when it comes to synth patches.

 

☺️

 

A.C.