Following on from last month’s praise-filled preview. Paul Ireson takes a longer, harder look at Ensoniq's new VFX synthesizer, and finds that the beauty of this machine is far more than skin-deep. Is this the synth of 1989? Read the review to find out...
Following on from last month's praise-filled preview, Paul Ireson takes a longer, harder look at Ensoniq's new VFX synthesizer, and finds that the beauty of the machine is far more than skin-deep. Is this the synth of 1989?
From the player's point of view, there are two principle levels of sounds on the VFX: Programs and Presets. Programs are the basic sounds/patches, or whatever else you might choose to call them (I do wish manufacturers could agree on a standard here), and Presets are performance arrangements of Programs. Up to three Programs can be combined and their key zones, transposition, volumes and other parameters altered to produce the performance Preset. Programs and Presets can both make use of the onboard digital effects processor. 60 Programs and 20 Presets can be stored in each of the VFX's internal RAM (INT bank), internal ROM (ROM bank) and a plug-in cartridge (CRT bank). There are two playing modes for the VFX: Preset and Multi. Multi mode is the means by which the VFX can be used as a slave for an external sequencer, providing 12 independent polyphonic parts, and Preset mode is more performance oriented.
As an alternative to playing the Presets, you can also select and play single Programs from within the Preset mode. Doing so, in fact, takes you into a temporary Preset buffer, where you can build up a Preset by selecting a single sound (which you've just done), and then layer up to two more Programs on top of the first. This combination of three Programs, together with whatever Preset parameters you edit, can be saved to a spare Preset location. However, I found myself using this edit buffer simply as a way of selecting single Programs to play, often prior to editing them.
Running through the VFX's factory sounds, my impressions were of a great range of sonic timbres, with 'movement' being perhaps the only identifiable link between them all, indicating the modulation depth of the synthesizer. There are excellent acoustic instrument samples (sax, trumpet, etc) alongside heavily effected synth sounds, resonance sweeps, bright digital zings... All the sounds are beautifully clean, as are the digital effects, but above all I found them inspiring. Synthesized sounds form the basis of the music I make, and a good sound, to me, is one that makes me create (fairly) original music: the VFX has a good number of those, and to me that is a good indication of its worth.
Programs provide the best starting point for examining the VFX sound structure, as they are the instrument's basic 'sounds'. Each Program consists of six Voices, processed by the effects unit, and contains all of the parameters relating to how aspects of the Voices and effects can be modulated. Each Voice consists of a digital oscillator playing one of 109 possible waveforms (or more, depending on how you count them), two filters, an LFO, three envelope generators, and a wealth of modulation possibilities. The two Patch Select buttons located immediately above the Pitch Bend and Modulation wheels are used for live control over which of the six Voices in the Program are actually played, as on the EPS.
The 109 waves that can be used by the oscillators are arranged into 10 groups: String Sounds: Brass Sounds; Bass Sounds; Breath Sounds; Tuned Percussion; Percussion; Transwave; Waveform; Inharmonics; Multiwave. The String, Brass, Bass, Breath and Tuned Percussion sound groups are similar in that all but one or two of their 'waves' consist of what are really complete sampled sounds, in some cases multi-sampled for maximum realism. Percussion is the only group of sounds that does not feature looped sustain portions.
The Transwaves are the most interesting group of waves: each of the 17 Transwaves is actually a wavetable, consisting of around 100 variations of a basic waveform. The variations sound similar, but have a different harmonic content which changes progressively through the wavetable. As the oscillator plays, the point in the wavetable at which it is reading the waveform can be modulated (by any of 15 sources), thus producing considerable timbral movement. The Transwaves have been carefully chosen to enable you to create some very exciting sounds: the waveforms are on the whole bright, aggressive digital tones, but also included are four resonance and two pulse width modulation waves. Although nothing quite as dramatic can be done with the other waves before they are processed further, a Start Point parameter is available for all wave groups up to Percussion: this allows you to set the point at which the playback of a wave starts, so as to lose some or all of the attack. The start point can be modulated, and velocity is particularly suited to this application. These waves can also be set to play backwards. All the waves also have a delay parameter, which simply delays the start of playback of the wave by between 0 and 250 milliseconds, or alternatively until the key is released. This allows the creation of some interesting echo effects. In order to make best use of this delay facility, the onset of the envelopes can be likewise delayed until the wave actually starts to play.
The Waveform group contains simple, single-cycle loops of organ, synth and some more acoustic sounds (excellent raw material for creating straightforward emulations of simple analogue synthesis), and the Inharmonic waves are longer loops containing more dissonant harmonics. These are excellent for adding some serious 'grit' to a sound. The final wave group is Multiwave, which strings together all the waves in the preceding nine groups - multi-samples and all. Start Point and Length parameters determine which of the waves is a starting point, and how many waves are played before playback repeats from that first wave again. The main use of this facility is simply to create rhythmic effects by repeating one or more waves over and over.
The range of waveforms provided by the VFX is broad, covering a lot of both 'acoustic' and 'electronic' ground. Many of the acoustic sounds can be used with no further processing, and sound damn good on their own with just a single Voice playing. Likewise, the Transwaves don't need any further processing to sound good, but it helps, and it's the processing and modulation side of the VFX, along with its ability to combine up to six Voices into a single Program, that enables each of the waveforms to be the starting point for a wide spectrum of glorious sounds.
The most important and most direct form of modification for the basic wave sound is filtering, and the VFX actually provides two filters per Voice, placed in series. Four poles are split between the two filters (a 'pole' is a measure of the roll-off of a filter, equivalent to 6dB per octave). Filter 1 can be configured as either a 2-pole or 3-pole low-pass filter, whilst Filter 2 can be a 1-pole or 2-pole low or high-pass filter. By varying the type of filters and the cutoff frequency for each, a high degree of tonal shaping can be introduced. Band-pass filtering, 'classic analogue' 24dB per octave low-pass, and many more possibilities present themselves. Each filter's cutoff frequency can be modulated by Envelope 2, and also by any one other modulation source. Therefore this shaping could also be controlled by any of the mod sources. Positive or negative modulation is possible.
On playback, the pitch of each Voice can be altered in a number of ways. Voices can be detuned by octaves, semitones and cents. Both the LFO and Envelope 1 are permanently assigned to Voice pitch as modulators, though their mod amounts can be varied so as to turn them off. In addition, any one other mod source can also be used to modulate pitch, and its depth varied.
I've mentioned two envelopes so far - Envelopes 1 and 2 - which are hard-wired to modulate pitch and filter cutoff frequencies (though they can also be used as mod sources for other destinations). A third envelope controls the amplitude of the Voice. All envelopes are identical in terms of their parameters: each has five level parameters, and five time parameters that determine how long it takes for the envelope output to change from one level to the next. The five levels are: Initial; Peak (though it is not necessarily the highest level); Break 1; Break 2; Sustain. The five times are: Attack; Decay 1; Decay 2; Decay 3; Release. Each envelope's output level can be modified either positively or negatively by keyboard scaling and note velocity (the latter with a choice of 10 response curves), and its attack time can also be compressed with increasing note velocity.
Any of the VFX's 15 mod sources can be assigned to modulate Volume, and a Pan position (with mod source and mod amount, of course) can also be set. These parameters are accessed in the Output page. This Pan position refers to a place in the stereo image of whichever of the three possible output busses the Voice is assigned to. These busses are Dry, FX1 and FX2, the uses of which will become apparent when we look at the VFX's effects. Also from within the Output page, a keyboard range for the Voice can be set, and a velocity threshold (above or below which the voice will not play). An extra 12dB of boost can be applied at this point, if necessary, to compensate for heavy filtering.
Another Output page parameter is Voice Priority, which can be set to High, Medium or Low. This determines which voices are 'stolen' first if you try and play too many notes at once (the VFX can play up to 21 notes simultaneously): careful use of this function ensures that less important parts of a sound are lost first. In addition, the VFX has an advanced form of dynamic voice allocation that stretches the oscillators over still more notes. Whenever an oscillator's output can no longer be heard - for example, when its amplitude envelope has dropped to zero, or when it has reached the end of a wave in the case of a one-shot sample - it is returned to the 'pool' of oscillators waiting to be played, even if the key that originally triggered it is still held down.
The parameters described so far affect only single Voices within Programs - the Program Control page deals with a set of parameters that affect all six Voices globally: Pitch Bend Range, Delay Multiplier, Glide and Pitch Table. Delay Multiplier (x1, x2, x4 or x8) acts on all the waveform delay times set within that Program, to raise the maximum delay to two seconds. Although the Glide (portamento) time affects all Voices in a Program equally, within the Pitch Mod page you can specify whether Voices will glide or sound normally, and if they glide, whether they do so monophonically or polyphonically. The Program Control page also gives access to the Pitch Table function: each Program can have a custom Pitch Table programmed into it as an alternative to the standard equal temperament tuning. The System Pitch Table (which defaults to equal temperament) can also be reprogrammed if necessary; though in this case, in order to get a Program to play with equal temperament, you would have to programme equal temperament as a custom table for that Program.
The VFX includes a high quality digital effects processor through which its sounds can be processed. It is at this point, just prior to the final output stage, that the previously separate audio paths for the six Voices of a Program come together. I refer you to the box on this page for a list of the algorithms, but briefly, the 15 effects include reverb, delay, flange, chorus, rotary speaker simulation, and several combinations thereof. Within each algorithm, you can adjust reverb decay, chorus depth parameters and so on, to tweak the effect to your personal requirements - the algorithm and parameter values are stored with each Program.
As mentioned above, each Voice can be allocated to one of three busses: Dry, FX1 and FX2. How these busses (all of which are stereo) are routed to effects depends on whether the effects algorithm consists of only one effect (Reverb or 8-Voice Chorus), or of two or more effects. In the case of algorithms with only a single effect, mix levels for the FX1 and FX2 busses act like an auxiliary send and allow different amounts of processing to be applied to each signal on these two busses. The various Voices in a Program can therefore be treated with different amounts of the same effect, rather than a blanket effect. One further refinement: in single effect algorithms, the FX2 bus can be used in a special mode where, rather than operating as a conventional stereo bus, its Pan position determines the dry/wet effects mix. This then allows real-time control over the dry/wet balance.
In multiple effects mode FX1 feeds Effect 1, and the mix level for Effects 1 controls how much of this output is sent to Effect 2 rather than straight to the stereo outputs. FX2 feeds Effect 2 (usually the reverb part of a combined algorithm) by way of the Effect 2 mix control. The result is that the effects can be arranged in series or in parallel, or a bit of both, for maximum processing flexibility. After all this talk of stereo effects busses, you won't be surprised to hear that the VFX has stereo audio outputs (but then what synth doesn't these days?).
The depth of modulation allowed within VFX Programs is the key to adding a further dimension to what are already basically very strong sounds, simply because so many aspects of the sound can be varied simultaneously, and by different mod sources. The 15 modulation sources that can be used wherever a destination has been listed above are: LFO; Envelope 1; Envelope 2; Mod Mixer/Shaper; Noise Generator; Note Velocity; Keyboard Scaling; Timbre Level; Voltage Control Foot Pedal; Pitch Bend Wheel; any external MIDI Controller; Pressure and Velocity; Mod Wheel and Pressure; Mod Wheel; Pressure. Most of these are self-explanatory, though one or two call for a little elucidation.
Timbre refers to the Timbre parameter, which is explained in the description of Presets below. The LFO Rate and Level can in turn be modulated by any of these sources (including the LFO itself!), and there are seven LFO waveshapes to cater for all tastes: triangle, sine, sine/triangle, sawtooth, square, positive sine and positive triangle. These last two are conventional sine and triangle waves, but the modulation level is adjusted so that zero is its lowest not its average value, which can be useful when trying to emulate a guitarist's finger vibrato technique (as strings can only be bent up in pitch). The Noise Generator simply produces a signal whose level changes randomly at a variable rate.
The most interesting of the modulation sources is the Mod Mixer/Shaper, which takes any two of those 15 mod sources (including itself) and mixes and transforms them. Mod Source 1 remains unaltered, but Mod Source 2 is adjusted by a scale factor and then further transformed by a shaper. The scale factors range from 0.1 to 8, so a fair range of compression or expansion can be applied to a source. The 16 curves which can further transform the scaled modulator level include the familiar variations on convex and concave, and also quantise shapes: these quantise the mod levels to one of 2, 4, 8, 16 or 32 discrete values. Why use such a shape? To ensure that a pitch bend effect jumped in semitone steps, for instance, rather than happening smoothly. The final shape is Smoother, which takes the rough edges off a rapidly changing input. This proved most useful for making the aftertouch response of the keyboard a little 'lazier', to compensate for slightly unsteady fingerwork.
Now that we understand Programs fully, it's time to look again at what happens when three are combined into a Preset. The most fundamental Preset parameters are the choice of the three Programs, and which of these are active (ie. sounding). The Programs are played from within Tracks (nothing to do with sequencer tracks!), and a Track contains all of the parameters that tell the VFX how to modify a Program when it is played from within a Preset. These include: Volume; Pan (which enables you to override individual Voice Pan settings); Timbre (editing this with the data entry slider gives real-time control over whatever aspects of a Program have Timbre selected as a mod source); Key Zone; Transpose; Release (which shortens or extends the amplitude envelope release time of Programs in a Track). You can also specify which of the Tracks will respond to the sustain pedal.
Besides modifying some aspects of VFX Programs on playback, Tracks are also the means for determining how the VFX acts as a controlling keyboard: each Track has a MIDI channel, and up to three channels of MIDI data can therefore be generated simultaneously from the VFX's keyboard. Each Track generates MIDI data only from the region of the keyboard defined by the Key Zone. Each of the Programs in the Preset can be set to respond only via MIDI (local off) or only from the VFX keyboard (local on), or both. A Program Change number can be set for each Track, which the VFX will send when the Preset is selected: the number does not have to bear any relation to the Program that the VFX selects internally for that Track.
The final important aspect of Presets is how effects are dealt with. The default effect algorithm for any Preset is the same as that used in whatever Program the Preset is based on. However, if it is not suitable, this selection can be overridden from within the Preset. The Voice routings can also be altered if necessary, to re-direct Voices to a different bus to suit the new algorithm and effects parameters. When changing from one Preset to another (or one Program to another), the VFX does its best to ensure a continuity of sound: any held notes are not cut off, and continue to play the same old sound until you release them. If a new effect algorithm has to be loaded for the new Program or Preset, a brief silence will be heard to cover up any glitches that might be caused - unless the old and new effects are very similar, in which case the changeover will remain smooth.
Multi mode is, in a way, just like a giant Preset, but with 12 Tracks instead of just the three. Although it is primarily intended for allowing the VFX to be played under external control (from a sequencer, say), all 12 Programs on the Tracks could be layered on top of one another for a really monster sound, and played from the keyboard if 11 Tracks were set to play under Local (or Both) control (as opposed to MIDI or Off). There is no storage space for Multi patches as such: effectively, Multi mode configures the VFX as 12 polyphonic synths, and if you want to change the Programs on each Track you can either do so from the front panel with the Replace Program function, or with an externally-generated Program Change message. Personally, I find this easier to deal with than a whole set of multi-patches of any sort. The effect selected for Multi mode can be switched under external MIDI control.
It's in relation to Multi mode that my only significant criticism of the VFX raises its head - the lack of multiple voice outputs. VFX offers stereo audio outputs only. If you are using 12 different sounds from the VFX, it's a fair bet that you'll want to process some of them separately, so the absence of individual polyphonic outputs is a real disappointment. Even two extra assignable outputs would have helped - Ensoniq have not even provided a means of connecting an external multi audio output box, as on their EPS sampler. An unfortunate oversight, and one that may put a few potential studio customers off.
I'm a complete convert to the VFX. It has a wealth of lively, exciting sounds that span a huge range of the aural spectrum - from traditional analogue, to acoustic, to entirely original digital sounds. The clarity and movement in the sounds is outstanding, as is that of the effects. But best of all, the sounds are not entirely the result of choosing some good basic sampled waves: they represent merely a few of the range of sounds that the eager programmer will be able to coax from the VFX.
The absence of a sequencer is a bit of a surprise, and means that the VFX is not really competing for your money if you're looking for a keyboard workstation. The consequence of this is that the VFX is cheaper than it would have been as a workstation. As it is, the expected price of the VFX looks a real bargain given the depths of this machine's synthesis, its polyphonic aftertouch and 12-part multitimbrality. The lack of separate outputs is a sad omission, but I think I can live with that. The bottom line is that the VFX is the most exciting and inspiring synthesizer of the moment. And at £1350, it's an absolute steal.
£1350 inc VAT.
Ensoniq GB, (Contact Details).
Gear in this article:
Review by Paul Ireson
Previous article in this issue:
Next article in this issue:
> Korg M1R
mu:zines is the result of thousands of hours of effort, and will require many thousands more going forward to reach our goals of getting all this content online.
If you value this resource, you can support this project - it really helps!