You are here

Korg N364

Workstation Synth By Gordon Reid
Published October 1996

Though Korg's new N‑series keyboards aren't completely new on the inside, having much in common with the recent X‑series synths, they still have plenty to offer as powerful workstations, with some neat and contemporary embellishments for '96. Gordon Reid sketches out the Korg family tree...

In 1988 Korg launched the M1, which, if we set aside unaffordable monsters such as the Fairlight CMI and Synclavier, introduced on‑board sequencing, 8‑part multitimbrality, and the workstation concept to the world. It also introduced Korg's AI (Advanced Integrated) method of synthesis. This, although it lacked resonant filters, was bright and flexible, and was a great success, so much so that Korg produced a rackmount version (the M1R), a cut‑down module (the M3R), and three souped‑up keyboards (the T3, T2, and T1). There was even a rackmount version of the T3, although, with somewhat arcane logic, Korg named this the M1R EX.

Still with me? OK... The next chapter in the AI story came in 1991 and, predictably enough, was called AI2. This came in the form of the second‑generation O1/W and O1/W FD keyboards, and the rackmount O1R/W. As before, there was a baby rackmount, the 03R/W, but this time there were no equivalents to the T‑series. The big difference between the 'M's and the 'O's was a facility called 'wave‑shaping'. In principle, this allowed you to distort the waveform, adding many high harmonics and, therefore, brightening the sound. And there was no doubt about it: the 'O's were noticeably brighter than their forebears, although (and few people seemed to notice this at the time) you could count on your fingers and toes alone the number of factory programs that used wave‑shaping. Clearly, the second‑generation machines were definite improvements upon their predecessors.

The third generation, introduced in 1993, used another new method of synthesis. This lacked wave‑shaping but, for reasons that Korg have never explained, was still called AI2. Nevertheless, the new AI2 retained much of the brightness of the O‑series, and the basic model of this generation (the 32‑voice X3 keyboard) leapt straight to the top of the best‑selling lists, where it resided for the next couple of years. Other third‑generation products included the inevitable X3R, the X2, and the 05R/W. Uh‑oh, this is getting confusing...

Despite its 'O' prefix, the O5R/W was an X3 module, the only difference between this and the X3R being the omission of one bank of sounds and the lack of a sequencer. But now things get really crazy... The X5 was the next X‑series keyboard, but this was an O5R/W with a keyboard — or, to put it another way, an X3 without a disk drive or sequencer. The X5D was a 64‑voice version of the X5, and the X5DR was the rackmount of that.

Now, eight years down the line, we have the fourth generation, the N‑series, comprising the N364 and the N264. Snappy names, aren't they? But here's the joke: while the '64' refers to the number of voices (no problem there) and the '3' and the '2' refer to the keyboard size, in time‑honoured T3/T2 and X3/X2 fashion, rumour has it that the 'N' refers to 'Next Generation'. Why's that funny? Because, while the 'T', 'O' and 'X' prefixes referred to derivatives of Korg's original synthesis, the sound generation in the N‑series is identical to that of the latter X‑series synths and modules. The N364 is the keyboard version of the X5DR, with the same amount of PCM ROM, and the same chip‑set, but with the disk drive and sequencer reinstated from the X3. Indeed, the N364 sounds identical to an X5DR. 'N' stands for 'No changes', maybe?

Well, of course there are changes, but they are more subtle than the introduction of a new method of synthesis. So let's delve below the surface and discover whether Korg have struck gold again, or whether the enhancements embodied in the N364 are more along the lines of the Emperor's New Clothes...


Physically, the 61‑key N364 is very reminiscent of the X3, with an almost identical body and end cheeks. Indeed, the top panel is the same aluminium extrusion, the only difference being the width of the screen cut‑out. The colouring is different, however, and I rather like the gun‑metal blue, which reminds me of the OASYS prototype I first saw 13 months ago, but which has yet to appear in public.

Other similarities to the X3 abound. The disk drive is mounted on the side (unlike the more user‑friendly O1/W FD and Trinity, which have it facing forwards) and, most unfortunately, the N364 shares the X3's major failing by offering a truly horrible keyboard. I hated the X3 and X2 keyboards, which I've previously described as black‑striped marshmallows glued to the front of the synths. While I found the N364's keyboard much less spongy, I also found it to be over‑light and extremely shallow. But maybe Korg are expecting users of the N‑series to come from a less 'pianistic' mould than me? It's not an unreasonable perspective.

...I should pause to give the GM bank special mention because it's one of the best I've yet encountered.

On a more positive note (oops, sorry), the N364 has one overwhelming advantage over the X3: it has four outputs. This is not trivial, because while the N‑series share the same chip‑set as the X3 and X5, they will be much more flexible in the studio. Funnily enough, the X2 also had four outputs, as did the X3R, and the more expensive Trinity, so Korg clearly understand that outputs are important. What's more, the inclusion of four on other X‑series instruments shows that the hardware is capable of supporting them. What I still don't understand is why Roland should be able to offer us six or even eight outputs, while Korg never seem to exceed four, even on their flagship products.


Continuing in this mode of comparing the N364 to the X3, I can summarise the voicing of the N‑series very easily. The N364 is simply two X3s bundled together, so it's got twice as much of everything. For example, the X3 offered a maximum polyphony of 32 voices, so the N364 has 64. Likewise, where the X3 offered a single bank of GM Programs plus two non‑GM banks of Programs and Combis, the N364 has four non‑GM banks, though it retains the one GM bank. Furthermore, the N364's 8Mb of sample ROM is 1.33 times that of the X3 (OK, so most generalisations don't bear close examination) and, as already mentioned, there are twice the number of outputs as on the X3. If there's a subjective difference in sound (for example, I always felt that the X5DR sounded better than the X3) it's probably a consequence of better factory programming. Once you've worked with a synthesis engine for a year or two, you learn a few tricks which enable you to create better voices. This happened with the T‑series (Korg's programmers were considerably better at voicing the T3 than they were the M1) and I'm sure that, out of the box, the N364 sounds a whole lot better than its ostensibly identical predecessors.

On that basis, there's not much point in lingering over the N364's programming capabilities, nor the quality of the results, in great detail. These have been covered in our reviews of the X3R (February '94) and the X5DR (May '95). Still, I should pause to give the GM bank special mention because it's one of the best I've yet encountered. What's more, it's editable, so full marks to Korg here. It's also worth pointing out that, while the N364 has lost the O1/W's wave‑shaping facility, it's brighter, cleaner, and offers extended parameters regarding the panning of the oscillators, and the way in which you can apply effects to individual parts. Finally, I should mention that the N364 retains Korg's traditional quick editing mode: just press one of the appropriate buttons on the top panel, then use the data slider or increment/decrement buttons to change primary values. It's hardly new — the M1 had it — but it's still a super feature.

With regard to its Combis, the N364 takes a step backward when compared to the M‑series and T‑series. While the screen will tell you how many Programs are inserted within a Combi, it can only give you the numbers of Programs 1‑4 or 5‑8 at one time. This is not a consequence of the size of the screen, but of the way in which the space has been (mis)used. And it's a pain. Before Korg jump on me, I'll admit that I'm aware that the higher cost of a larger screen would have to be passed on to the consumer. But better programming should have cost no more, and besides, if the M1 could do it...


We now diverge from the N364's predecessors, because we come to the all‑new, real‑time Pattern‑play function, part of the N364's onboard sequencer (see 'Sequencer Strengths' box for more sequencer details).

Patterns are phrases that you can assign to keys. They may contain drums, or basses, or organ, or guitar — or whatever you like — and different keys access different Patterns. So, for example, bottom C could trigger your basic bass and snare, while D and E access variations, and F and G access fills. Then, C# and D# could offer hi‑hats, ride and crashes, with other percussion under F# and G#. Above these notes, A, B and C could have bass riffs appropriate to the basic rhythm, with suitable fills on A# and C#... and so on. These phrases can then be played in real time from within the sequencer. The Roland XP50 has a similar feature, but the bonus of the N364 is its immediacy. The Roland demands that you decide which Patterns you wish to use and where, and then place them in the appropriate tracks. Then Korg simply invite you to put the sequencer into multitrack mode, initiate recording, then press the keys that relate to the desired Patterns and listen to the sequence develop as it's recorded. If you press a key a little early or a little late, there's no problem. Provided that you're deriving MIDI clock internally, the N364 doesn't allow you to record anything out of time, and quantises the Pattern 'on the fly'.

There are important constraints, however. The most important of these is the need to match the Patterns to the song in which they're being placed. Failure to do this will result in incorrect instrumentation and, in all likelihood, the wrong tempo. Serendipity isn't my chosen method for writing music, and I've discovered that playing funk patterns with a light jazz kit and double bass doesn't work. You have been warned.

Despite the fact that the N364 can store up to 100 Patterns, the nature of Pattern mode precludes a great deal of variation, and the tracks generated display a great deal of repetition. In other words, nobody is going to write a violin concerto in this fashion. On the other hand, the dance fraternity are going to love it. Programming bass and drums can be a mind‑numbing operation for some people, and this is a very quick way of getting the basics recorded. Then, once the rhythm track is in, you can use the sequencer to add transposition, variation, and changes in instrumentation. Mind you, it's hardly going to make software such as Band in a Box, or even Korg's i‑series keyboards, obsolete. It might if it, for example, recognised chords played in the upper half of the keyboard (or something like that) and tracked them in real time. But it doesn't. You can play over the top of the Patterns, but that's all.

In addition to the Patterns pre‑loaded in the N364, and the others that Korg supply on the factory disk, you can, of course, create your own. These can be edits of the factory data, or new Patterns programmed from scratch. You can even load sequences from elsewhere, select bits and pieces that you like, and allocate these to Pattern memories.


The final major facility, and another one 'new' for '96, is the N364's arpeggiator. This offers five modes: Up, Down, Alt1, Alt2, and Random. Of these, three are self‑explanatory, but the alternate modes bear explanation. On most arpeggiators, the upper and lower notes are played once per cycle — for example, a C major triad would be played over a single octave as C E G E on the first cycle, followed by C E G E again on the next. This is the behaviour exhibited by Alt1. Alt2, on the other hand, plays the highest and lowest notes at the start and end of each half‑cycle, so that our triad now becomes C E G G E C, followed by C E G G E C again. You may think it a fine distinction, but this allows you to keep arpeggios within time signatures, shifting between common and complex times at the touch of a button.

The parameters that you can then apply to the arpeggio are range (1‑4 octaves); note sorting; gate time (to make arpeggios more 'choppy'); MIDI velocity (1 to 127, or the played velocity); sync; latch; and speed. Oddly, if you continue stepping sideways within the arpeggio parameters, you find yourself in the edit map of the program playing the arpeggio. Whether this is intentional or not, I don't know, but it could be very useful, allowing you to modify the sound as the arpeggio is playing.

Because at the time of the review there was no manual for the N364, I looked at a Prophecy manual (as the Prophecy is also a Korg synth with an arpeggiator) to see whether there were any hidden arpeggio functions. I found that most parameters were identical, although the N364 has lost a couple of features, such as gate modulation and velocity modulation. But I suspect that few players will feel the lack of these.


It's time to sum up and, as is usual for a magazine review, space constraints mean that there are features we've ignored, and classic voices we've overlooked. Indeed, that last point is no throw‑away line. The sheer range of sounds in the N364 is beyond review. With 936 Programs and Combis derived from a mind‑numbing 430 multisamples and 215 percussion samples, there's no way I can give you a flavour of what this synth can do. The best I can say is that, after a Trinity Pro or fully‑populated JV1080, the X5DR was next on my list of sonic lusts. Since the N364 sounds no different, I see no reason to change this view. Notwithstanding a few minor niggles regarding less than perfectly mapped multisamples, it sounds great!

Another plus point is the consistency of the editing system. If you can program an M1, you can find your way around an N364 without recourse to a manual. You might miss a few bits and pieces, but all the major facilities are where you would expect them to be. Unfortunately, many players no longer program their own patches. Maybe they would if workstations had the immediacy and accessibility of analogue synths, but nowadays most people just want a bucketful of sounds. Here, again, the N364 scores well, and with obvious reason. Where it fails to score is in its lack of expandability. The Roland XP50 hosts expansion boards that include new PCM sounds and additional patch memory. Even the humble M1 accepted PCM data cards. In comparison, the N364 is a dead end. Whether you would ever reach its limit is another matter, but it's a point worth noting nonetheless.

With regard to the N364's enhancements over the X‑series, the new Pattern mode is very easy to use, and I can see third‑party companies dashing to be the first to offer extra Patterns. The arpeggiator is an even more welcome addition, which lifts the instrument above the level of a simple workhorse to something a bit more immediate and 'synthy'.

To sum up: with the N364 and the 76‑note N264 costing just £1299 and £1499 respectively, Korg have possibly regained the top spot in price/performance terms, and it's very likely that we'll see no more of the X3. Indeed, I wouldn't be surprised to learn that Korg had already deleted it from their product lists.

Though N364 is, for all practical purposes, identical to an X5D, its extra bank of sounds, disk drive, sequencer, real‑time pattern mode, and arpeggiator add up to an attractive bundle of facilities at an attractive price. I suspect that the N364 could be the keyboard buyers' Christmas Number One.

Sequencer Strengths

There are no real differences between the X3's and the N364's sequencers. Indeed, since the review unit was a pre‑release model without documentation, Korg supplied an X3 manual with it, and this proved to be completely adequate. Whether this shows a lack of effort on Korg's part — it's been almost four years since they developed the X3 sequencer — or is a result of Korg deciding that if it ain't broke, there's no need to fix it, I'm not sure. What I mean is this...

One of the reasons that the X3 sold so well was its sequencer. The instrument's major competition was, and remains, the Roland XP50 (see review in SOS June 1995), and this has a derivative of the now‑revered MC500 sequencer built‑in. However, the MCs were never the most intuitive of units. If you took the time and trouble to get to grips with them they were extremely fast in operation, well structured, and flexible. However, their screens were small, and the messages displayed on them were often cryptic, so relatively few players took the time and trouble to learn them, nor their XP50 progeny. Since the X3 sequencer was easier to learn and use (although, of course, not up to the standard of a computer‑based sequencer) this became a selling point.

So, once again, there's little point in regurgitating a review of the N364's sequencer. It features all the normal facilities — cut, paste, delete, insert, event edit, and so on — and its capacity is also unchanged from the X3, at 32,000 events. However, there's one facility that warrants specific mention: the N364 holds its sequences in static RAM rather than dynamic RAM. This means that sequences are retained when you switch it off, and these become available again immediately upon power‑up. Mucho brownie points for Korg.

If I have to criticise, 32,000 events is not a huge amount in 1996. It's not exactly mean, but RAM — even S‑RAM — is relatively cheap, and if the Trinity can offer 80,000 events, surely Korg's latest workstation could show an increase over the 1993 model?


  • Huge range of classy Programs and Combis.
  • The re‑introduction of an arpeggiator.
  • Continuity with previous Korg workstations.
  • Good price/performance ratio.


  • Over‑light keyboard.
  • Badly‑utilised screen.
  • The name.


Great‑sounding, good value workstation. Though very similar to the X3, it has some valuable extras, while retaining a very competitive price.