Kurzweil K2600XS
SYNTHESIZER/SAMPLER WORKSTATION
by Jim Aikin
Synthesizer workstation with drawbar organ mode and optional sampling.
The K2600 has a long list of mouthwatering features. It even looks impressive, thanks to the two ribbon controllers, eight sliders, and two left-hand buttons, all of them programmable. The rear-panel jacks are thoughtfully labelled along the rear edge of the front panel. Yes, that's a floppy disk drive tucked away by the wheels -- and no, your eyes aren't deceiving you: The K2600 is a subdued shade of purple. Inside, it sports a few thousand (!) new synthesis algorithms.
The rear panel of the K2600XS. In the upper layer are (L to R) a stereo digital audio optical input, low-impedance left and right audio ins (XLR), high-impedance line-level stereo audio in (1/4" TRS), multipin in and out connectors for 8-channel digital audio interfacing with Kurzweil's DMTi audio interface, a stereo digital audio optical output, AES/EBU stereo digital audio in and out (XLR), the line-level A, B, C, D, and mix outs (all 1/4" balanced), and the stereo headphone out. In the lower layer are 25-pin SCSI and SCSI thru connectors, MIDI in, out/thru, and out connectors, an LCD contrast knob, two 1/4" controller pedal inputs, a minijack input for a breath controller, and four 1/4" switch pedal inputs.
Designing a new keyboard instrument by building on your existing technology is a longstanding tradition: Think of Roland's series of JV modules, or Yamaha's line of VL wind modeling synths and add-on boards. So when Kurzweil unveiled the K2600 series, the fact that the design is directly based on their popular and powerful K2500 series (reviewed May '96) shouldn't surprise anybody. If it ain't broke, why fix it?
Building on an already extensive feature set, the K2600 sports some new features sure to grab the interest of anybody who is shopping for a pro-oriented instrument. Maybe the biggest news is that the KDFX multi-effects board (reviewed April '00) is now standard rather than an optional add-on, as it was in the K2500. This powerful four-bus board adds not only versatility but a stunning level of sound sophistication. Thanks to KDFX, you can program complex mixes in a K2600 without having to resort to any outboard gear except the mixdown DAT or CD-R.
Also new in the K2600, and available to K2500 owners, is live mode. This uses the sampling input option to process incoming signals in real time -- not only through KDFX but through the entire synthesizer architecture, with its filters and waveshapers. Assign a few filter parameters to the front-panel sliders, and live mode becomes a terrific resource for sculpting mixes of songs recorded on another platform. With the digital I/O option, which adds eight channels of I/O in Kurzweil's KDS signal format suitable for interfacing with their DMTi audio breakout box, you can process an entire eight-track master tape through the K2600 on its way to your mixdown deck -- pretty impressive.
In the synthesizer area itself, the K2600 sports a new voicing architecture called Triple Modular Processing, or "triple mode" for short. We'll have more to say about triple mode below. Essentially, it lets you design much more complex voice DSP chains than before. Also new, and available to K2500 owners as a software upgrade, is KB3 mode, which turns the Kurzweil into a realistic and surprisingly programmable tonewheel organ emulator.
Other new features are less obvious, but perhaps just as important. The K2600 has been redesigned internally so as to clean up the audio signal path. All 16 MIDI channels are now "drum channels," meaning that there's no restriction on how many channels can play programs with up to 32 layers. Eight channels of digital audio input are available, compared to two on the K2500. You can now program up to eight-way velocity cross-switching within a keymap, which is a handy shortcut for sound programmers developing realistic Rhodes patches and the like. And if you're running short of RAM ID numbers for your samples, you can group multisamples under a single sample ID.
Both K2500 owners and K2600 owners can take advantage of up to 8GB of hard disk space, in the form of four 2GB partitions. Twenty-bit digital output is supported. And the sequencer has a new "RAM tracks" feature that makes it easy to fly samples directly into a song without having to assign them by hand to keymaps and programs.
Sound enticing? You bet. One thing that hasn't been upgraded between the K2500 and K2600 is the ROM waveform set, which is identical. If you're using the instrument mainly as a sampler, this is not a big issue, but if you buy it without the sampling option, you may wind up feeling that the waveforms are not quite competitive with the wide variety of sparkling sounds you'd get in a new Korg, Roland, or Yamaha synth. Even without the sampling option, though, you can load sounds from CD-ROM in many popular formats, so the ROM waveform palette is less of an issue.
The factory programs have been seriously updated. Most are entirely new and modern-sounding. Be sure to try using the first few sliders when auditioning them. The new synth basses are wicked! "Industrial Kit II" takes advantage of KDFX to add some fresh aggression and weirdness in the percussion department, and the breathy bell tones of "Highlandistic" will be perfect for my next vaguely ethnic new age piece. I could fill pages describing the sounds that ship with the K2600, but we need to move on.
The voicing parameters, sample editing commands, and other features are mostly identical with the K2500 as well. This is not a bad thing, necessarily. The K2500 had such a wonderful graphic display for waveform editing, for example, that it's hard to see how Kurzweil could have improved it. Since many samplers these days max out at 64 or even 128 voices, though, the 48-voice polyphonic limit (24 with stereo waves) feels a bit skimpier than it used to.
5-Cent Tour
We don't have nearly enough space to cover every aspect of the K2600 in detail. This is a professional-level piece of gear, with the kind of operating system you'd expect. So let's hit a few high spots quickly, and then zoom in on the new features.
The K2600's voice architecture uses a variable algorithm design. There are 31 algorithms with various signal routings -- and that figure ignores triple mode (see below), which adds a host of new algorithms. Within an algorithm you can choose DSP functions for up to four different "blocks." The available DSPs include not only various kinds of filters but also waveshaping and other types of controllable distortion. Each K2600 program can have up to 32 layers, each layer being assigned its own keymap (a set of samples laid out on the keyboard), algorithm, modulation routings, and so on.
Modulation sources include not only all of the standard MIDI controllers but unique things like square waves derived directly from MIDI clock in various rhythmic subdivisions. Complex modulation shapes can be created using four function generators (Kurzweil owners call them FUNs) per layer. The envelope generators have three attack stages, a decay stage, and three release stages. Suffice it to say that the K2600 is a programmer's paradise.
Programs can be combined in Setups of up to eight key zones each. Setup mode contains a programmable arpeggiator, and while this lacks the funky rhythm patterns of some newer arpeggiators, it compensates with some very slick latching and transposition options. You can also assign sequence tracks to be triggered by individual keys in Setup mode so as to do a little MIDI remixing on the fly. Only one sequence at a time can be played in this mode, so it's not as flexible as the phrase sequencing on some instruments, but it's useful nonetheless.
The K2600 sequencer is competitive with the best hardware-based sequencers in other workstations. Start with 16 tracks, a high clock resolution, groove quantizing, event editing, controller thinning and remapping, and the ability to import data from another song into the current one. Songs can be created by chaining short sections together, and you can run a chain (called an Arrangement) concurrently with a 16-track song, which gives you a total of 32 playback tracks. And if 32 tracks aren't enough, you can resample the entire output of the K2600 -- assuming you have the sampling option installed -- to a stereo audio RAM track (see below) and start a fresh layer of MIDI overdubs.
One small curiosity: While the sequencer's internal clock resolution, according to Kurzweil, is 768 ppq, the sequencer editing environment has only 480 ppq resolution. If your chops are good enough to take advantage of that extra resolution in unquantized tracks, you know who you are.
A fairly comprehensive suite of sample-editing functions is provided. You'll find both basic items like cut-and-paste and more advanced DSP of the pitch-shift and time-stretch variety. In the K2600, compression/limiting has been added to the DSP menu. Some unusual beat chopping/transforming operations are also possible. Best of all, you can view the waveform graphically while you work, zooming in both vertically and horizontally to see exactly where the desired sound starts and ends. Offhand I can't think of much they could add that would make my life easier.
The K2600 sports a comprehensive set of disk utilities. You can create macros to load whatever combinations of objects you might need, load them into memory in various ways (such as renumbering objects to create continuous banks), partition hard drives, audition individual objects from the disk page prior to saving, and so on.
Live Mode
Maybe you've already recorded a great drum or bass track in your digital audio multitrack, and you'd like to see what it sounds like when processed by the K2600's filters and effects. The old-fashioned way is to sample it, assign the sample to a keymap, assign the keymap to a program, and so on. At this point, you'll have to figure out where to position the note-on message so as to get the sampled audio in sync with the rest of the song. The K2600 gives you a far easier method: With live mode, you can send audio to the sampling input, process it through a live mode program, and bounce it back into your recorder as a new track.
In fact, that's just the start of the fun. If you have the DMTi eight-channel digital I/O option, you can process an entire eight-track mix through KDFX in real time. This is a quick, easy way to add automated effects to an ADAT master, for example, on the way to mixdown. Even without the sampling option installed, you can use the digital I/O option to process eight tracks of audio through the voice filters and KDFX. (Sorry -- we didn't receive a DMTi in time to try this.) Live mode would also be ideal for onstage processing. My only criticism is of Kurzweil's choice of their own proprietary KDS format for the multichannel digital I/O. Surely an instrument of this caliber ought to have ADAT lightpipe I/O. The optical connectors are already in place, after all.
Kurzweil provides a handful of live mode programs as templates to get you started. The only caveat is, you have to press a key on the keyboard to hear the live mode audio (or choose a Setup in which the key-down event is included in the Setup definition), because the K2600 is interpreting the audio input as just another waveform in its voice architecture.
The subversive among you are doubtless thinking, "Just another waveform, eh? What happens if I grab the pitchbend wheel?" Kurzweil thought of that. The K2600 will do its best to change the pitch of the input audio in response to both pitchbend and MIDI key performances. It will even play the contents of the input buffer backwards! Inevitably, trying to fool Mother Nature this way will result in some hiccuping as the audio input data buffer overflows (with downward pitch-shifts) or empties out and repeats (with upward shifts). If you don't insist that your music be in 4/4, though, this hiccuping can be a great way to add odd phrasing to an unsuspecting drum beat. I had great fun with it, that's for sure -- just don't ask me how I'm going to line up the rest of my tracks with this spastic new "beat."
Triple Mode
If you hanker after synthesis resources available nowhere else, you've just found them. The K2600's triple mode voicing (not available to K2500 owners -- sorry) serves up some of the deepest voice programming I've ever seen. With a couple of hours of experimenting, I was able to come up with some seriously disturbed (in a good way) patches.
I also found myself baffled by the sheer number of possibilities. Getting decent sounds out of a triple mode patch seems to me -- and I've been programming synthesizers for a few years now -- more a matter of luck than of planning. If the manual provided a tutorial with some concepts to pave the way, I might have a different impression. The manual, which overall is quite thorough and readable, explains clearly how triple mode programming works in a technical sense, but when it comes to the musical applications, you're on your own. Kurzweil provides ten triple mode programs in ROM to get you started, and another 100 on disk, but to my ears these don't produce many sounds that are clearly different from the rest of the K2600's analog-type sounds.
The idea behind triple mode is simple: Three layers within a program can be linked to form a single signal path. The entire triple then produces a single voice, reducing the K2600's total polyphony correspondingly: If you're using only triple mode voices, you'll be limited to 16 notes, just as if you were using standard a three-layer program. As with normal Kurzweil layers, each layer in a triple can have its own algorithm. The block diagrams for the algorithms, conveniently displayed in the K2600's large LCD and also included in the manual, explain the signal flow.
Each of the layers in the triple has its own set of algorithms: There are 30 for layer 1, 38 for layer 2, and 26 for layer 3. My trusty pocket calculater tells me that makes almost 30,000 possible configurations for a triple -- and we haven't even started plugging DSP into the block diagram yet. Each block in a layer's algorithm can have one of between three and 18 possible DSP functions assigned to it, ranging from simple lowpass or highpass filtering to waveshaping, added "analog" waveforms (sine, saw, pulse, noise), crossfading, and parametric EQ. Multiply it out, and you'll find that the total number of configurations for a triple runs into the trillions. And that's before you start editing the parameters.
I found the triples best suited to creating highly synthetic (in other words, "digital"-sounding) analog-style sounds. By assigning five or six of the K2600's eight front-panel sliders to control various aspects of the sound, I was able to generate some very expressive sweeps -- not just filter sweeps, but much more far-reaching types of waveform modulation. This is a real strength of the K2600, and of triple mode programs.
I also found, however, that in order to make effective use of the sliders, I had to assign my new triple programs to Setups and play the instrument in setup mode. Here's why: Individual K2600 programs can't be given entry values for the sliders. As a result, when you call up a new program, any parameter being modulated by a slider will be offset by the current position of the slider. This is quite different from how a knob-laden analog modeling synth works. In an instrument of the latter type, the current position of the knobs doesn't matter when you call up a new patch, because the associated parameters will use their stored values. K2600 Setups do have entry values for the sliders, so when you select a new Setup containing a triple program, you don't have to manually grab the sliders and jam them down to the bottom (or, worse, try to find a position halfway up) before playing the keyboard.
KB3 Mode
Realizing that many players yearn for the playability of a drawbar organ but can't afford both a B-3 clone and a multi-purpose instrument, Kurzweil developed KB3 mode. Available as an upgrade to K2500 series instruments, KB3 is also part of the Kurzweil PC2 and PC2X, and we wrote about it in our review of the latter instrument (June '00). The implementation in the K2600 and K2500 goes further, however, allowing you to create both authentic Hammond sounds and wildly inauthentic -- but interesting -- timbres.
All of the players who tried playing KB3 organ while the K2600 was in my office were impressed with the realism and playability of the sound. A couple of B-3 snobs said, "Well, it's not the real thing, but it would work fine at a gig." I'd go further: In a rock band recording, not even an expert will be able to tell whether the track is real or a KB3 organ.
The main differences come down to playability issues, such as the fact that KB3 mode doesn't provide dual manuals, pedals, or panel switches for controlling a real Leslie. (There's a panel switch for controlling the KDFX Leslie simulation, but that's not quite the same thing.) On the other hand, you can't edit a performance on a real Hammond in a MIDI sequencer. All of the KB3 controls transmit and receive MIDI.
In KB3 mode, the K2600's eight front panel sliders function as the drawbars, and the buttons above the sliders are dedicated to switching between rotary speaker speeds, turning the chorus and vibrato on and off, and switching the percussion. When percussion is off, the mod wheel works as the ninth drawbar; in other words, switching percussion on "steals" the top drawbar, just as it would on a real organ. (You can choose a different drawbar to be stolen if you want to.)
The sliders operate just as if they were drawbars, creating less or no sound when pushed away from you and a louder sound when pulled toward you. In one of many enhancements over a real B-3, you can choose whether the sliders will have a 0-8 stepped response or a 0-127 smooth response. You should avoid touching the mod wheel "drawbar" if you've switched percussion on, because its movement will cause the percussion envelope to smear out in a very unrealistic way.
Select a KB3 program and press the edit button, and you'll be whisked into a world of sound customization that most B-3 emulators can't touch. For starters, the lower (1-4) and upper (5-9) tonewheels can be assigned various waveforms. Want your B-3 drawbars to play sawtooth waves instead of sine waves? The K2600 is happy to oblige. One set of drawbars (though not both at once) will play a sampled waveform, so you could even assign drum loops to the drawbars if you dare. For more off-the-wall possibilities, each drawbar can be tuned up or down in half-steps.
In edit mode, the meanings of the percussion switches can be redefined. Some of the edits may increase the authenticity of the sound to your ears; others will be more visionary. You can control the amount of key velocity applied to percussion volume. The pitch of any drawbar can be used for either the low or high percussion pitch. The level and decay time can be set for all four switch combinations (loud/fast, loud/slow, soft/fast, soft/slow).
Key click is nearly as programmable. You can set the basic loudness, the velocity response, the length of the click (from realistic to an absurdly long noise burst), and its pitch. The amount of randomness in the key click loudness can be programmed.
The choice of three settings for vibrato and chorus is very authentic, as is the fact that you switch between one and the other using a front-panel button. John Krogh and Ken Hughes, both of whom have played a lot of Hammond, praised the realism of the vibrato and chorus.
Each KB3 program has four stages of parametric EQ, plus the standard KDFX processing. You also get the standard complement of LFOs, ASRs, and function generators per program. There's not much they can be applied to compared to the modulation destinations in a regular K2600 program -- mainly pitch and amplitude. But if you need to add vibrato from the mod wheel or a footpedal, you can do it.
In the realism department, you can program the amount of "leakage" between tonewheels. This is possible because KB3 really does model a Hammond's tonewheel setup, by using DSP voice channels for the various tonewheels. One consequence of this is that KB3 sounds use up a lot of polyphony -- 40 voices for an organ model with 79 wheels. This 79-wheel organ, like a Hammond, is fully polyphonic, because all of the tonewheels are running "in the background" all the time. When you press a key, you're gating one or more wheels, depending on which drawbars are out. If you need more polyphony for other sounds, you can reduce the number of tonewheels.
At first I thought Kurzweil had missed a golden opportunity. Why aren't there fine-tune controls for the drawbars? In the middle of a sleepless night I figured it out: As on a real tonewheel organ, the sliders and keys are gating voices that sound continuously in the background. Therefore, a detuned "tonewheel" would correspond to different drawbars depending on what key you play. Given the nature of the KB3 organ module, it would be flat-out impossible to detune, say, the fifth drawbar on every key unless the K2600 had about nine times as many voices in its synthesis section.
RAM Tracks
Not too many composers and arrangers these days work in a MIDI-only environment, so it makes sense that Kurzweil would add a feature that lets you overdub audio directly into your sequences. You'll need to have the sampling option installed, and enough free memory to hold your audio, as the K2600 uses RAM for the audio tracks.
Recording to RAM is a bit more limiting than being able to record audio direct to hard disk, as on the Korg Trinity, for two reasons. First, your total track length for all tracks can't exceed 128MB (about 25 minutes of 44.1kHz recording). Second, you have to save the tracks to disk at the end of your session and then reload them each time you want to work on the song, which takes a couple of minutes. The advantage of RAM tracks is that even if you're storing the RAM tracks to an old, slow hard drive, you can easily add a couple of dozen tracks, and they'll all play perfectly. You're not limited by the rotation speed and access time of a hard drive.
The idea is this: You arm the K2600 for sampling -- I found that forcing sampling to start by using the automatic threshold detection worked well -- and then put the song into record mode with an empty record track selected. Record the audio, and stop the song. At this point, the K2600 will ask you to strike the root key for the sample. It will create a "vanilla" sound program with this sample assigned to that key, and also insert a MIDI note-on message at the proper point in the sequence. When you hit the play button, the new MIDI note will trigger your new sample, and you'll hear the RAM track. What could be simpler?
In practice, it took me several tries to get my first RAM track recorded, as the directions in the manual are vague on a couple of important details. I found that the K2600 preferred to insert my new RAM track sample into whatever sound program was selected for the track I had armed for recording, rather than creating a new sound program as the manual indicated it would. I later learned that it will create a new program only if the track is empty. This is a thoughtful bit of design, as it allows you to record as many as 32 RAM "tracks" to a single sequencer track.
Your RAM tracks will "free-run" after being triggered, rather than being locked to the sequencer's clock. But with a normal-length pop song, it's highly unlikely you'll hear any audible drift, unless you change the tempo of the song after recording the audio. And RAM tracks have other advantages: After recording, you can select an algorithm with a parametric EQ to do a bit of premixing, or assign the RAM track program to a KDFX bus to add reverb. Because the tracks are playing through the K2600's voice architecture, each track can be given its own EQ without using up any of your effect buses.
In Use
To explore the power of the K2600 as a workstation synth/sampler, I put together a sort of techno/fusion track (very electronic, but with chord changes) using its own sequencer. Since the factory programs include a number of interesting drum kits, I went with MIDI drums rather than a sampled loop. The "SkoolBass
Simple" program gave me a fat bottom end, "Fluid Grand" provided some washy, unobtrusive chords, and "TchRezoid
Hungry" was the source for a resonant, animated riff. (In case you're curious, programs with that funny symbol in the middle of the name switch between two distinct sounds when you move the first slider up or down.)
The first challenge I faced was deciding which KDFX studio preset to use -- and how to dial it up. Unlike some workstations, the K2600 doesn't assign effects directly in a song. Instead, you can use the studio associated with one of the programs in the sequence, or choose a studio at the Master level. I decided on the former method, as the "TchRezoid" studio sounded pretty good. Unfortunately, it also put some slapback echo on my drum sounds, so I had to enter program edit mode (which can be done with one button-press directly from the main sequencer page), reassign each of my drums to a different KDFX bus, and save the edited program. Editing the KDFX preset would have been slightly more difficult, though ultimately more useful, as I really had no use for that slapback.
After creating an eight-bar riff I liked, I saved it to a second song memory location and changed the chord progression, keeping the same drum beat. The next step was to call up an empty song, switch to the sequencer's Arrange page, and string my two eight-bar "songs" together as steps in a master song. After listening to the two riffs back to back, I decided the hi-hat pattern in one needed to be heard in both of them. I could have recorded it by hand into the second song, but instead I used the convenient "grab" command to copy the hi-hat track from song 1 into song 2.
While working in the event list, I noticed that my tracks were filling up with unneeded aftertouch data. The sequencer itself provides no way to filter this out prior to recording, but I was able to delete it from all of my tracks with a single command. I later learned how to filter the aftertouch by disabling it in the K2600's "Control Setup." The event list also has a view filter, so you can keep the aftertouch if you like, yet not have to stare at it while editing.
The mix wasn't quite right, so I needed to reduce the volume levels of a couple of tracks. The sequencer's mixer screen has no data readout of the levels or pan positions for the tracks, but that turned out not to be a problem. While playing the song, I lowered the level sliders to where I wanted them and then hit the Keep button, which wrote the volume data at the head of the track. To automate a smooth fadeout at the end, I assigned volume controller data to a slider and did a few overdubs.
In the end, I felt the K2600's sequencer wasn't as convenient as a computer-based sequencer (which probably will come as no surprise), but it definitely had all the tools I needed to produce a finished piece of music. And with the huge palette of high-quality sounds, I never felt I needed to MIDI up another synth -- though I could have done so, as the sequencer tracks can be set to transmit external MIDI only.
Conclusions
No two ways about it: The K2600 is a serious tool for professional musicians (and aspiring pros). It offers gorgeous sound, a deep and comprehensive array of features, and a ton of live performance control. The latter is accessed in the keyboard models by a good set of ribbons and sliders, but the rack unit has some special software tools to facilitate remote control.
The enhancements to the operating system may not transform the K2600 into a whole new instrument, but they're very welcome. The only real drawback I can see when comparing it to its predecessor is that the ROM waveforms haven't been updated. And even here there's a bright side: This ensures backward compatibility with the existing sound library. In any case, it's less than a crucial feature for a sampler, but Kurzweil tells us two new 8MB ROM waveform blocks are under development, so help is on the way.
In the plus column are such enticing new features as the triple mode sound algorithms, RAM tracks, the KB3 organ emulation, and live mode processing through the voice filters and the new KDFX processors. Oh, and did I mention the vocoder? I've said it before, but it's never been more true: If you can't make great music with a K2600, don't blame the instrument! There are several excellent samplers on the market, and more than a few great synthesizers, but the K2600 is the only instrument I can think of that's both.
K2600 Vocoder
Though it may shake your confidence in your favorite magazine, we need to make a tiny confession: We're not perfect. In the August '00 issue we tested 11 hardware vocoders -- all the models we could get our hands on. We used the Kurzweil K2600XS to record a test tone (a sustained chord), and then processed this chord with a couple of different modulator sounds (a drum beat and a voice speaking a phrase) in all of the vocoders. What nobody on the staff realized was that the K2600 has a vocoder mode of its own. Oops. Click here to hear the K2600's results.
The K2600 vocoder can do both realistic vocoding and some very unusual vocoder-based effects, depending on how you set it up. The sampling input option is required for vocoding, and the special vocoder programs and setups are loaded from a floppy that comes with the instrument.
The vocoder uses the K2600's voice architecture rather than the KDFX processor, which means that most or all of the 48-note polyphony is soaked up by the vocoder. If you want to create a carrier tone within the K2600 itself, you can use the 20-band vocoder on the floppy, which leaves you eight free voices, or the 22-band vocoder, which leaves four voices. (You can reduce the number of bands further by reprogramming if you need more free voices.) Since we already had the carrier signal recorded on a computer hard drive, we opted to test the 24-band vocoder.
Three of the front-panel sliders are assigned to control envelope follower speed, bandpass filter width, and the frequency offset of the carrier bands. (See the August issue for more on these terms.) If you need finer control, you can enter edit mode for the vocoder programs and tweak individual bands to taste. The manual mentions several possibilities -- panning alternate bands to opposite sides of the stereo field, using an LFO to modulate the center frequencies of individual bands, scrambling the frequencies of the carrier bands relative to the modulators, or passing the carrier signal through assorted VAST DSP blocks. You could also create a straight high-frequency pass-through if you need more sibilance to make the spoken words easier to understand.
Even without any adjustments at the program level, the vocoder was fairly intelligible, though not extremely crisp. It sounded quite solid when processing the drum beat. (We've added these soundfiles to our online vocoder archive.) Kurzweil tells us they're planning to release some additional vocoder programs that will bring the frequency bands closer together, putting them in a more useful range for processing typical sounds such as those we were using.
In the signal routing department, a couple of oddities popped up. First, the K2600's volume slider doesn't work on the vocoder output, because the output is sent out via the A pair of stereo outputs, not the mix outputs. Second, in order to use a carrier signal generated by the K2600 itself, you have to patch the B output of the unit back into the sampling input. Not quite space age, but it works.
Given the cost of the K2600, not too many people will be buying it specifically to use as a vocoder, but even so, the vocoder is a valuable addition to the instrument. For sheer programming power, it's at the top of the pack.
Models & Options
The Kurzweil K2600 is available in six variations. Three hardware models are offered: the K2600 (76 semi-weighted keys, $6,256), K2600X (88 weighted keys, $6,820), and K2600R (3U rackmount, $5,175). Each of these can be purchased with or without audio inputs for sampling; the sampling units are designated the K2600S ($7,140), K2600XS ($7,700), and K2600RS ($5,950). Sampling can be added to a non-sampling unit as an option ($900).
Even without the sampling option, the K2600 comes standard in the U.S. with 64MB of waveform RAM (expandable to 128MB), which can be used for loading sounds from CD-ROMs, as the SCSI connector is standard, not an option. The prices above don't include an internal hard drive; if you want to add a hard drive, inquire about the hard drive mounting kit.
For Vital Stats, see the October 2000 issue of Keyboard.
KEY INFO #120
Senior editor JIM AIKIN is a songwriter, science fiction writer, computer programmer, synthesizer programmer, and classical cellist. Not all at the same time, of course.