Browse reviews

@ a b c d e f g h i j k l m n o p q r s t u v w x y z

gear fetish

drum module (44 total)
:: random entry :: add entry

effects (61 total)
:: random entry :: add entry

manufacturer (118 total)
:: random entry :: add entry

microphone (5 total)
:: random entry :: add entry

MIDI tools (5 total)
:: random entry :: add entry

recording gear (25 total)
:: random entry :: add entry

review (459 total)
:: random entry

sampler (44 total)
:: random entry :: add entry

sequencer (19 total)
:: random entry :: add entry

soft-synth (27 total)
:: random entry :: add entry

synthesizer (174 total)
:: random entry :: add entry

limit to
drum module effects manufacturer
microphone MIDI tools recording gear
review sampler sequencer
soft-synth synthesizer
Location :: review :: n :: noisewreck
review :: noisewreck

moron basically compared it to a K2500 with "a few subtle improvements" and said it's not as up-to-date as it should've been.

On paper at least, one would agree. But if you're going by the spec sheet alone, you're completely missing the point of this instrument. Also, the K2600 adds Triple Modular Processing (TMP). Later on you'll see that this is far more than just a "subtle improvement", specially when one takes into account that the K2600 was released in 1999. I've had this instrument since 2000 and it still amazes me!

There are three basic versions of the K2600: the 76 key semiweighted K2600, the 88 key fully weighted K2600X, and the rack module K2600R. All three also come as "S" versions which adds the sampling option. The sampling option is not necessary if you want to simply load samples into the instrument from disks. However you need it for direct sampling, as well as external input (Live Mode in Kurzweil-speak). Also, the latest versions of these instruments come with the Contemporary and Orchestral ROM cards installed.

As I haven't seen any in depth reviews of any of the K2xxx series instruments here, I'd like to cover some of VAST that's common to all instruments in the K2xxx range, and then concentrate on what's new on the K2600 (specifically TMP as this turns the K2600 into a completely different monster). Hopefully, at the end the readers will have a better understanding why there is nothing else out there that can come close to these beasts, and why in many cases the amount of ROM samples is irrelevant.

Variable Architecture Synthesis Technology (VAST). The guts of the system.

A VAST program (similar to a "patch" in other synths) is made up of 1 or more layers. The best way of conceptualising a VAST layer is to think of it as a complete semimodular synth topology (I'll get to the "semimodular" aspect shortly). You can have 1 VAST algorithm per layer in a program, and you can have up to 32 layers in one program (in a sense, you can have up to 32 complete and different synths in a program, each with its own LFOs, envelopes, and FUN(ction)s. Put differently, each layer in a K2xxx program is a full fledged synth in its own right.

"Conventional" and "TMP" Layers:
"Conventional layer" is what I like to call the layers that are common to all K2xxx models. The K2600 (and K2661) also support what Kurzweil calls "Triple Modular Processing" or TMP. I'll talk about each below:

All in all, there are 31 VAST algorithms available to a "conventional" layer. Any of these algorithms can be used in any given layer in a program. These algorithms are common to all K2xxx models as stated above. With the OS v.2 of K2600, Kurzweil introduced TMP. Basically TMP turns 3 VAST layers into one huge modular environment where the signal flows from the algorithm in layer 1 to the one in layer 2, and finally to layer 3 before being output. Each TMP layer has it's own set of available algorithms. There are 29 TMP algorithms for layer 1, 37 for layer 2 and 25 for layer 3. As you'll see shortly the sound shaping/mangling/programming possibilities with TMPs are staggering, and although reviewers tended to brush over these, to me this turns the K2600 (and K2661) into a completely different monster.

Algorithms in detail:
Each algorithm basically incorporates a fixed signal path that connects various DSP blocks to one another. The power, and "semimodularness" if you will, comes from the fact that those DSPs are not fixed, but can be selected by the user from an available list. So, although you have a fixed signal path, and fixed number (1-3 depending on the algorithm) of DSP blocks per algorithm, you're pretty much free to chose any DSP for any of the blocks. The available DSPs range from 1, 2 and 4 pole filters (LP, HP, Notch, BP and Allpass) and EQs to DSP oscillators (which can be used in conjunction with the ROM or user samples) to various waveshapers and distortions and PWM. PWM is implemented differently in VAST. It's not really "true" PWM, but because of the way it's implemented, it allows you to perform PWM-like stuff on samples and external audio.

The last class of DSPs are the different kinds of mixers. Mixers come into play when dealing with split-wire algorithms. As the name suggests, in a split wire algorithm an audio signal gets split in two, where each stream then can be put through different DSP blocks and then be recombined using those various mixers. Other than your regular level mixer, there are some others that allow things such as ring mod, AM, FM/waveshaping and other kinds of crossmodulation.

All these DSP blocks can be modulated from various sources, such as the ones I listed above, but the actual list of available mod sources fills over 8 pages in the user manual. There are several clock based global LFOs, several keyboard based mod sources (velocity, key number, aftertouch), and some more esoteric ones such as the number of keys being held at a given time, the speed of sample playback, etc. Incidentally, you can also modulate the LFOs and the envelopes from all the available mod sources, including themselves.

Back to TMP: The maximum number of DSP blocks in TMP layers 1 and 2 is 4 although layer 3 still has 3. Now, imagine three layers worth of all these DSP blocks (11 alltogether). Now imagine that each layer is still a full fledged synth layer, with it's own sample source and it's own LFOs, Envelopes, etc... Obviously at the basic level it would allow a user to come up with some interesting DSP chains, such as having 3-4 filters with embedded distortion and/or waveshaping blocks all of which can be modulated (yes, including distortion and waveshaping amounts)... but... now imagine algorithms that allow you to split and combine signals on each layer, sometimes at more than 1 point on a layer... which leads you to those crossmodulating mixers... which allow you to crossmodulate one sample with another... or external audio!!! Now imagine the fact that each of these sound sources can be filtered, distorted, bent and twisted and pretty much mangled beyond recognition, before AND after crossmodulating each other, while you're modulating all those DSP parameters!!! The sheer number of sound mangling possibilities are truly staggering!!!

Now you can see why the Motifs, Tritons, Fantoms etc, have the NEED to add more and more waveforms to their ROMs. In a sense this is a way to make up for their rather limited synthesis capabilities. The K2xxx series, and specially the TMP capable K2600 and K2661 don't really have the need for them, specially when using waveshaping and crossmodulation one can create new and constantly evolving timbres that are unheard anywhere else!

A few words about Live Mode. As mentioned at the beginning of this review, you need the sampling option for it. What makes Kurzweil's Live Mode unique, compared to other synths that also allow external audio input, is the fact that the incoming audio has to pass through the sampler's buffers on its way to VAST and KDFX. Also, in order to hear a sound, you must press a key to hear it. The fun begins when you start playing different keys, while the audio is streaming in, as this effectively causes realtime and polyphonic pitch-shifting! This is not, however glitch free. Obviously the pitchshifting comes from the fact that different keys play the playback buffers at different speeds (just like how it or any conventional sampler plays them). However, this causes asynchronicity between the input and playback buffers, causing underrun and overrun errors, which is were all the weirdness and the fun begins for the experimental and the industrial types. The Live Mode inputs are dual mono, which means that you can either use them to treat stereo signals, or completely unrelated mono sources, and then use Triple Mode Processing to have them mangle each other!

KDFX (Kurzweil Digital FX):

With the K2500 Kurzweil introduced KDFX as an option which is standard equipment on the K2600 and the K2661. It is still the most powerful FX section of any synth, and can hold its own against multiFX devices in the $1500-$2500 range. KDFX is basically a 4 bus mixer with an aux section. It has 4 stereo or 8 mono (or a combination of the above) inputs. Each input has a two band EQ section. You may think that this is not that flexible, but each EQ "block" can be switched between a high/lo-pass, hi/low-shelf or fully sweepable parametric EQ, in a similar manner to how one chooses DSP blocks in VAST algorithms. Each input can be sent any two of the four available FX busses. Each FX bus may or may not contain an FX algorithm, or you can chain up to 3 FX algorithms on a single bus. You can send any or all of the FX busses to the Aux bus, before sending to the physical outputs. The signal path of KDFX is highly configurable, giving the user full command as to what goes where and what gets sent to which physical output. Again, there is very little that is hardwired here. This brings a certain level of complexity with it, yet there is nothing else out there that can match the sheer flexibility of this arrangement.

The FX algorithms themselves are all 1st class. There are the usual suspects of various reverbs, delays, flangers and choruses. There are distortions, some of which model guitar amps and cabinets. There are several waveshapers, quantize/aliaser, multimode filters, and some esoteric ones such as "LaserVerb", Ring Modulator, a couple of frequency shifters, Pitcher, and EQMorpher. The EQMorpher is basically an infinitely programmable 4-band formant filter. The Pitcher is sort of like a tuned waveguide, it takes unpitched sources and imposes a certain pitch on them (so what you say? You can't really "play" melodies using it right? Well... You can in fact! I'll get to that shortly). There are a couple of FX that deliberately induce feedback such as Chaos and RegenDegen (a multitap delay that can be sent into feedback instability, yet be tamed by a built in compressor). There are 3 and 5 band parametric EQs, single and multiband compressor/expander (both hard and soft knee), enhancers, gates, and even stereo phase alignment tools).

Remember that earlier I said that everything and anything can be modulated on the Kurzweil? Well, this extends to KDFX. In fact this is implemented to such an extent that would make any other multiFX box green with envy. The only limitation is that KDFX must be modulated by Global sources (although KDFX also has it's own mod sources, the VAST Global mod sources can also be called into action). This includes things such as GKeyNum (Global Key Number), which is a monophonic control source that can be used to apply keytracking to KDFX parameters. Usual suspects would be things like filter cutoff, or when you want to play melodies using stuff like Pitcher. For more adventureous, you can for example switch a Filter from an LP mode to HP to BP to Notch and back (or forth) using an LFO or MIDI controller or anything else, WHILE MODULATING THE CUTOFF, without any hiccup! Or you can just go all nuts and decide to modulate the room type, size, diffusion amount, early and late reflection amounts and time and reverb time of a reverb algorithm with... say... an LFO (or a random value generator or...) to some truly bizarre effect!

Folks, this instrument inspiers the tweakheads. In fact it's the tweakheads' ultimate wet dream! It really opens up some synthesis and sound design possibilities that unless you've come across it in this instrument, you'd not even think of.

Currently the K2600XS sells for $3,899.00 if you shop around, while the 76 key K2600 (w/o the sampling option) goes for $2,999.00. Sure, these are not inexpensive instruments. However, considering that you get what amounts to limitless number of synths, a multiFX section that rivals the best of them, a decent mastering tool (with all those high headroom EQs, multiband compressors and phase alignment tools), throw in full MIDI controller capabilities, a flexible sequencer and a very live friendly Setup Mode (similar to a "Multi" on other instruments, where you can also assign keys to trigger different sequences!) you can see that the sum of all the parts of this ultimate workstation more than justifies the price. It is certainly money well spent!
posted by: noisewreck on 2005-12-07 00:23:35
view: other review entries posted by noisewreck