Nearly four years since the DX7 gave the world "FM synthesis", and changed the face of modern music, Yamaha have introduced two new instruments to replace it.

IT SOUNDS LIKE a DX7. In fact, it sounds like two DX7s. The back panel reads "DX7". But take a look at the front panel, and what you see is a new instrument.

There is little doubt that the DX7 is the most popular synthesiser ever made. Not, perhaps, in sheer number of units sold, since the DX7 didn't sell for less than £1000 in any territory of the world, and not every musician can afford to spend that kind of money on a synthesiser, no matter how good it is. But in terms of endearing itself to musicians and music consumers, in terms of the number of pop records, live concerts, cabaret events, academic recordings, jazz jam sessions and ethnic fusion experiments that have benefitted from a single instrument's existence, the DX7 reigns unchallenged.

And the fact that it has consistently sold over the past four years indicates that Yamaha would have to make some serious mistakes in its redesign to change the trend. As it turns out, however, the new DX7s really are worth a close look, even if the old DX holds few secrets from you.

There are two new models: the DX7IID and the DX7IIFD. These instruments are basically the same as the original DX7 - both are 16-note polyphonic, six-operator, 32-algorithm FM synthesisers with five-octave velocity- and pressure-sensitive keyboards. But the new DX7s feature improved sound quality (through new envelope generators and quieter DACs), a revitalised MIDI implementation (a department in which the original DX7s suffered), improved controls and displays, and most importantly, the ability to play two distinct voices simultaneously. Accompanying this last feature are doubled internal and cartridge memory capacity for access to twice as many voices as before.

Both models operate identically, except that the IIFD features a 3.5" floppy disk drive for storage of its own voices as well as those of other instruments connected to it by MIDI.

The new DX7s cost a little more than the original model, but this is due more to the current strength of the Yen than to the added features, many of which have been implemented in software and are therefore relatively cheap to install.

Format

THE ORIGINAL DX7 was, for all its complexity, very simple to operate: you plugged it in, switched it on, adjusted volume levels (probably the hardest part), then pressed one of 32 voice switches to get the sound you wanted.

In this sense, very little has changed. The main difference is that now, when you press one of those 32 switches, you select not one voice, but two, along with an assortment of programmed performance control settings. But you needn't think about all that while playing the DX7.

The orignal DX7 was also a challenge to program. Now, this is neither the time nor the place to go into detail about FM synthesis, Yamaha's implementation of it in the DX series of instruments, and why it is inherently difficult to use intuitively the way some other synthesis systems are. Suffice it to say that out of all the people who have bought and now use a DX7, only a small percentage ever get beyond using their instrument's preset sounds (or add-on ROM equivalents), even if their intention is to do otherwise.

The new DX7 makes it easier to create customised sounds by simply combining distinct voices, and adjusting a handful of controls in Performance Mode, the highest level of operation. But creating new voices and editing existing ones is just as demanding as ever, even though some attractive new features make programming in Voice mode (the equivalent of the old DX7's Edit Mode) much more interesting.

Then there is the appearance of the DX7IIs. Gone are the mylar-covered switches (vague-feeling and unreliable), the smooth-finish chassis, and the 32-character LCD (uninformative and frustrating). The DX7IIs feature "proper" switches, dual two two-character LED readouts, and an 80-character back-lit LCD, all housed in a textured metal chassis.

Most of the controls are laid out as on the original model, though (so existing users shouldn't suffer too much culture shock), with the addition of a few new switches here and there to accommodate some of the new features. A second slider has been added next to the data entry slider, not for programming purposes, but for real-time control of any programmable parameter in performance. In fact, the data entry slider also has control over any programmable parameter when the DX7ll is not in Edit mode, so now you can easily adjust any sound parameters while playing the keyboard. (Remember when synthesisers had knobs?)

The master volume control remains where it has always been - to the left of the data entry slider - while the breath controller and headphone jacks remain below the pitch and mod wheels.

And nothing has changed on the back panel, with the exception of the addition of a second audio output which allows the DX7IIs two voices to appear at separate outputs, or for them to be panned across the stereo image with the new Pan feature (more on that later). The number of footswitch and footpedal inputs is the same, though there have been some changes in how the DX7 deals with these controllers.

The sound quality of the DX7IIs is much clearer than it used to be, and, though this may well be due to the new voice layering feature, warmer than that of the old DX7. There's no mistaking the clang of many of the DX7's characteristic FM voices, which would be sorely missed if omitted, but many of the factory-supplied performance presets avoid overemphasis of the bell tones, and instead blend the complex timbres with simpler ones for more usable sounds. And if the 32 performance presets aren't exactly what you're looking for, there are still 128 distinct voices which you can easily combine to suit your needs.

Voice Mode

ALTHOUGH IT IS not the highest level of operation on the new DX7s, Voice Mode is perhaps the best place to start since it resembles the old DX7's Preset Mode, and a basic understanding of Voice Mode is useful when it comes to customising performance setups. In Voice Mode, you have direct access to each of the 64 internal and 64 cartridge sounds, but can stil hear the effect of combining two voices in split or layered keyboard configurations.

When the DX7II is first switched on, the Play LED lights to indicate that the DX is in Performance Mode. Both the LED readout and the LCD convey information about the current performance program, while to the left of these, three Voice Mode switch LEDs show the current program's keyboard configuration - Single, Split, or Double. Pressing any of these three switches puts the DX into Voice Mode with the selected keyboard configuration. So, if you want to work with just one voice, press Single; to try various voice layers, press Double; or to try split combinations, press Split.

Once in Voice Mode, the 32 number switches select individual voices, as before. Since it is possible to play two voices simultaneously in Voice Mode, an A/B switch (to the right of the displays) toggles between the two voices, unless Single Voice Mode is selected, in which case the A/B switch is inactive. Since there are 64 internal voice memories and only 32 switches, another switch (labelled "1-32/33-64") selects one of two banks.

The only peculiarity about Voice Mode that I found was that when Split Voice Mode is selected, the keyboard split point is fixed at C3. Since Voice Mode is not intended as a performance mode, this detail might have been easily dismissed, were it not for the fact that voice combinations cannot be selected in Performance Mode, and it would seem a simple matter to have a split point transferrable from one mode to the other.

Once in Voice Mode, and having used the A/B switch to select the voice you want to edit, editing that voice's parameters is accomplished in much the same way as on the original DX7. But the larger LCD displays more parameters, so less button-pressing is required in Voice Edit mode, which is invoked by pressing the Edit/Compare switch. Once a parameter, or group of associated parameters, is selected using the numbered switches (which do not select voices once in Edit Mode), two cursor switches position a cursor in the LCD over the parameter to be adjusted. There is not much in the DX7IIFD's manual regarding editing in general, though a helpful Quick Reference Guide shows all of the DX7II's menus and parameters.

As on the original DX7, many of the parameters affect only one operator at a time, but to make life a bit easier, the DX7II's feature individual Operator Select switches rather than one switch stepping through the operators. These switches come in handy when it comes to copying envelopes from one operator to another (something which, if you've ever spent a lot of time with a DX7, you'll know to be a necessary and incredibly frustrating plan of action).

As far as actual voice parameters go, there are several welcome additions such as multiple LFOs (one per voice) which lend a more realistic "motion" to voices when played polyphonically. Of course, the LFOs can still operate in unison by setting the LFO Mode parameter to "Single", rather than "Multi".

Each voice can have its own key mode, which allows a variety of voice-stacking options for thicker sounds without giving up polyphony altogether. One of the factory preset bass voices works well in "Unison/Poly" key mode, which renders two-note polyphony with split keyboard configurations. Switching on the Oscillator Sync then gives a bit more "oomph" by doing away with some phase cancellation.

Improvements in the pitch-bend department include three new modes. Two of these modes allow only the lowest or highest note to be bent, while the third new mode bends all notes held without affecting those being held with the sustain footswitch.

Yamaha have also added a Random Pitch Depth control so that voices can appear to drift ever so slightly out of tune just like real instruments, giving voices that approximate to the "warm" analogue sound of earlier synthesisers. Not always necessary unless you're emulating an acoustic instrument notorious for possessing tuning inaccuracies (harpsichords, for example), but effective all the same.

But my favourite keyboard tweak is Fractional Operator Level Scaling (there, said it), which could make each person's IIFD slightly different from another's. Fractional Scaling is an alternative to the standard keyboard scaling of individual operator output levels, which allows independent scaling of groups of three adjacent notes across the entire keyboard. Just testing out the feature, with no specific effect in mind, I found it remarkably easy to make certain groups of three notes sound slightly (or incredibly) different from the notes around them. On the electric piano voices, this made it easy to have certain "tynes" brighter than others, just like a real electric piano.

Ironically, the sophistication of this instrument allows it to sound as imperfect, and therefore as endearing, in many cases, as the old instruments so many of us have locked away in the attic in order to make room for the new.

The real meat of the matter is in the application of these fractional scalings to create multiple timbres across the keyboard, using only one algorithm. Thus each voice can contain multiple splits, limited only by the imagination as applied to the DX's algorithms, and by your patience; it still takes a long time to generate useful voices.

And there's a catch: the fractional data can only be stored on RAM cartridges, so you have to have a RAM cartridge before it can even be stored on disk. This came as an unwelcome surprise, and though there is probably an explanation for it, it does limit fractional scalings to cartridge voices - an unusual inconsistency.

Performance Mode

HAVING CREATED, EDITED, or simply selected voices in Voice Mode, the next thing to do is return to Performance Mode to assemble these into final performance programs. Pressing the Performance switch is the first step, and the Play LED lights to confirm that the DX is in Performance Mode. Editing in Performance Mode works the same as in Voice Mode; only the parameters differ.

Pressing Edit to enter Performance Edit Mode does not always do the job, unfortunately, and this is where the DX7IIs can be confusing. For example, the voices used in a particular performance program cannot be changed in Performance Edit Mode; they can only be selected in Voice Mode. On the other hand, even though there are three switches labelled Single, Split, and Double, these select Voice Mode, so they are inactive even in Performance Edit Mode. Instead, there is a separate "Voice Mode" parameter which can be edited in Performance Edit Mode. Another instance in which your instinct must be suppressed until you get used to the controls.

On the subject of splits and layers, if either is selected in Performance Mode, corresponding split point and detune parameters become accessible as appropriate for the selected Voice mode. In either case, a Balance parameter determines the relative mix of the two voices, and a programmable Volume parameter sets the overall performance level.

Accompanying the three Voice modes are four Pan modes which determine how voices appear at the two audio outputs. Although there are several modulation sources for each voice's "pan" setting, the applications are a bit limited. Pan Mode 0 is useful for actually moving combined voices across the stereo image, but unfortunately, this affects all voices simultaneously, rather than individual voices. Thus a note which initially appears in the left output does not remain there if another note (in the same voice) is panned elsewhere, so the effect is not the same as on, say, the Prophet VS, which features polyphonic panning. When two distinct voices are layered in Dual Voice Mode, modulating the panning of one voice and not the other brought some interesting, more subtle effects.

Performance Mode is where the two Continuous Sliders (CS1 and CS2) come to life. These can be assigned to generally useful parameters such as the voice balance for basic performance control, or they can be assigned to parameters like algorithm number for outlandish effects. The assignment is programmable for each of the 32 performance programs, so one moment you could adjust the brightness of a sound by controlling a modulating operator's output level, and the next moment the same slider could determine the LFO speed. Easy to work with, and useful, to say the least.

The sliders are a big step forward in the real-time control of FM, even if they're still a compromise in view of the rest of the DX7II's complex programming facilities, and even if Korg's pair of envelope sliders on their new DS8 synth are even easier - and more rewarding - to use.

It's likely that many DX7IIs will spend most of their time layering two voices, even if this does reduce the instrument to eight-voice polyphony. For these situations, the Performance Mode Note Shift feature helps put voices into compatible registers.

In a move which should prove popular with academics and those who want to use their DX7II in - shall we say - specialist applications, Yamaha have provided micro-tuning facilities. The 11 preset temperaments available include Pythagorean (handy for providing incidental music to Ancient Greek plays), Werckmeister and Kirnberger (for a touch of the baroque), quarter-tones (full of Eastern promise), and eighth-tones (which give just an-octave-and-a-quarter across the entire length of the DX's normally five-octave keyboard). There are also a couple of memories for storing your own tunings, and for some reason cartridge memory has room for 64 user-defined tunings, and though this is not a bad thing, it makes me curious as to why the DX's internal memory has been short-changed again. And rather disappointingly, these alternative temperaments are available only in Single Mode.

These tunings could open up several new cans of worms, but they could also lead to the creation of some interesting new music. After all, there's no shortage of musicians and composers interested in microtonality, but there has been a shortage of instruments offering the right facilities. Until now.

The DX7II also deals with note retriggering with a "forced damp" parameter affecting all envelopes in a particular performance setup. When the forced damp is on, envelopes are reset each time a note is played, regardless of whether or not they have completed their final decay. This is a subtle effect, but noticeable enough to make it well worthwhile.

Storage

ANY OF THE parameters discussed so far can be stored in the DX7II's internal non-volatile memory, so power can be switched off without losing programs. Should there not be enough room internally, voices and performance setups may be stored on formatted Yamaha RAM cartridges. Since the new DX7 can hold more voices than older models, cartridges for the old DX7 can be loaded with half the new model's memory, and a special adaptor, the ADP1, is required for this. Otherwise, the DX7II relies on the new RAM4 cartridges. If you do not plan on creating your own voices, additional voices may be loaded from ROM cartridges.

The DX7IIFD features a 3.5" disk drive which can be used for storing many memories' worth of voices, either from internal memory or directly from cartridge. Having onboard access to so many memories makes it especially easy to create custom voice collections for specific applications. The DX7IID, which has no disk drive, still has the option of data storage over MIDI to dedicated storage devices, or to computers equipped with appropriate software.

One topic which the manual does not cover in any detail whatsoever is using the DX7IIFD's disk drive to store MIDI data received from other instruments. After some trial and error, I managed to get the disk drive to store a few files containing System Exclusive dumps from a non-Yamaha instrument using the Disk MDR (MIDI Data Recorder) option. Since the DX7IIFD is set up to accept dumps of up to 20K of memory, it would probably have a hard time accepting samples, but could very well handle drum machine pattern or sequencer data dumps over MIDI. This alone could make the DX7IIFD a suitable master keyboard controller in small MIDI systems.

MIDI

ONE AREA IN which the old DX7 caused trouble was MIDI, but considering the fact that it was one of the first MIDI-equipped instruments ever built, it's hard to blame it for lasting all these years with a minimal implementation. And in any case, the new DX7s do more than make up for lost time.

First of all, the DX7IIs can receive on any two channels so that voices A and B can respond independently to incoming MIDI data. The transmit channel is also variable, and though this is an improvement over the original DX7 (which only transmitted on channel 1), it would make sense to have two transmit channels to follow the A and B voices (useful in Split Mode).

A neat trick I discovered is setting the A and B voices to different MIDI channels, then creating a performance program in Split Voice Mode with the split point at B0, so that the DX keyboard plays only the B voice, while an external keyboard can play the A voice. Well, I liked it...

Elsewhere, an unusual Note On/Off option determines whether the voices respond to all incoming note messages (assuming mode and channel settings allow), or only odd- or even-numbered notes. Not for everyone, perhaps, but...

Program changes are well looked after; when programs are selected on the DX7, it can transmit either the same program number, or a corresponding number which the user programmed into the DX7's memory. This is similar to the way the Oberheim Matrix 6R does program mapping, but applies to transmitting program changes. All in all, in fact, the DX7II makes program-number assignment remarkably easy to accomplish.

Similar flexibility is lent to assigning MIDI continuous controllers to the footpedal, footswitches, and continuous sliders (CS1 and CS2). This applies both to transmitted and received MIDI data.

A wide assortment of System Exclusive dumps are handled: voice, performance, micro-tuning, and other system information can be transmitted either from internal or cartridge memory, or directly from the edit buffer, in which case you needn't commit the data to cartridge or internal memory prior to transmitting it over MIDI. As mentioned earlier, the DXIIFD could really steal the show with the MDR feature; recording MIDI data from outside instruments is a simple matter of selecting the MDR "in" option, naming the input file, pressing the "Yes" switch, then initiating the data dump from an instrument connected to the IIFD's MIDI input. Sending the MIDI data back to instruments is simply a matter of selecting the desired MDR "Out" file, and following the same procedure, except that the receiving instrument needn't initiate the data transfer.

Verdict

AFTER PLAYING WITH the DX7IIFD for some while, getting used to its idiosyncracies, and thoroughly enjoying its sound, there is no doubt in my mind that the few compromises I encountered were not significant enough to matter. Programming the DX7 has not become any easier, except at the performance level, in which case you need only deal with the somewhat confusing interaction between Performance and Voice Modes to get satisfying results with a fraction of the effort demanded by the original DX7.

This is a wonderful-sounding instrument. I'm willing to bet that many of those who found the old DX7's sounds too stark will give FM another chance now, though that is not to say that the DX7 doesn't still sound like a DX7, or that it will replace analogue synthesisers.

The new packaging, the double-voice capability, the disk drive, and the versatile performance controls make the DX7IIFD a new instrument, rather than a mere revision, even if the technology inside it has changed little, if at all.

The DX7 is already well-known by many musicians, and its new incarnations are similar enough for those musicians to feel almost completely at home with them within a matter of minutes - something which, as so many new keyboard instruments appear with technology that seems entirely foreign, must count for quite a bit.

So the DX7IIs won't cause the stir the original DX7 did back in 1983, because they simply don't need to. Die-hard DX7 owners will probably upgrade to the IIs, and recording studios will probably feel obliged to get one of them for their improved sound quality alone.

Prices DX7IID, £1699; DX7IIFD, £1899; both including VAT

(Contact Details)