You are here

Cherry Audio Elka-X

Cherry Audio Elka-X

Cherry Audio recreate that rarest of beasts, the Elka Synthex.

Several reasons have been advanced to explain the spectacular commercial failure of the Elka Synthex. Some have blamed the name — Elka were primarily associated with low‑cost organs and accordions. Others have blamed the company’s failure to appoint an appropriate distribution network. Yet others blame the DX7, although that’s daft because the Yamaha appeared nearly two years later. But one thing seems clear; no‑one claims that it was a failure because it sounded bad.

Despite its sophisticated looks, the Synthex was a basic, bi‑timbral polysynth. In each of its eight voices, the outputs from two DCOs plus noise were presented to a multi‑mode filter and then an audio amplifier before the mixed audio was passed through a chorus effect and onward to the outside world. Ring modulation of one oscillator by the other plus an unusual cross‑PWM feature created complex tones for sounds such as bells and chimes, two ADSR contour generators provided shaping, and low‑frequency modulation was generated by a global LFO per layer plus a dedicated vibrato oscillator. The eight voices could be split between its Upper and Lower parts to provide splits and layers, but neither velocity nor aftertouch were supported. If this sounds much like an Oberheim OB‑Xa or OB‑8, it’s because it is. The only unusual extra was the provision of a four‑track sequencer.

Introducing The Elka‑X

Although it looks like a Synthex, Cherry Audio have followed a well‑trodden path with the Elka‑X by taking the original architecture and extending it in numerous ways. So, for example, whereas you could only detune DCO2 on the original, you now have independent fine‑tuning of both oscillators. More significant changes can be found in the filter (where a 12dB/octave low‑pass mode has now appeared) and amplifier sections, as well as in the contour generators, where you can now determine the desired amount of velocity sensitivity. What you won’t find, however, are any ‘analogue feel’ controls to add small inconsistencies between voices. Perhaps because the Synthex used DCOs, Cherry felt that these would be inappropriate.

There are now three onboard effects units; the original chorus has been emulated, but the soft synth also boasts a sync’able, multi‑mode delay unit and a reverb. When you use the Elka‑X bi‑timbrally, a complete set is available in each Layer. The Master output pane has also been extended, with controls for layer volume and pan as well as master tuning and volume, and a simple limiter has been added to minimise any nasties.

The primary modulation source is again an LFO per layer rather than per voice. This generates your choice from six waveforms (a uni‑polar square wave and a random mode have been added) with independent depths for the Group A bus (DCO1’s and DCO2’s pitches and pulse widths) and the Group B bus (the filter cutoff frequency, amplifier gain and the sequencer rate). The LFO can be synchronised to the master/MIDI Clock and has also gained a Reset function, reinitialising the waveform when a new note is played after all previous have been released. The performance pane containing the triangle‑wave LFO2 and pitch‑bend controls has also been retained, but with three important enhancements: master/MIDI Clock synchronisation, an initial amount control for the modulation depth, and an Acceleration slider that has nothing to do with acceleration but instead allows you to control the LFO frequency using the mod wheel. When using the Elka‑X bi‑timbrally, you can apply all of the parameters in this pane to the Upper, Lower or both layers.

For some reason, the buttons selecting the Single, Double and Split modes and those selecting which Part you’re editing are found at opposite ends of the GUI. Those selecting the mode share a pane with the switch for the number of voices (up to a maximum of 16) and, beneath this, you can select polyphonic, monophonic and mono unison modes as well as unison detune, chord memory and multi‑triggering. Only ‘last note’ key priority is available when using the Elka‑X or one of its Layers as a monosynth, but that’s fine — last note priority with multi‑triggering is many players’ preferred combination. Finally, a bunch of utilities perform useful housekeeping tasks, allowing you to copy, paste and otherwise manage the layers and their effects.

In Use

Integrating the Elka‑X into my MIDI/DAW setup proved to be simple thanks to an excellent MIDI Learn window that allows you to determine a CC for each control, with minimum and maximum values as well as definable response curves. Happily, you can direct a given CC to multiple parameters simultaneously, which means that you can do things such as use aftertouch to control (for example) loudness, brightness, growl and effect depths at the same time. However, there’s no support for MPE, which, given its appearance on other Cherry products, surprised me.

But what of the sound? Given all of the differences described above, is the Elka‑X really an emulation of the original or just a vanilla soft synth dropped into a GUI designed to evoke your gear lust for a piece of revered unobtainium? Before answering this, I need to address a handful of issues...

Let’s start with the good news. Several teething problems were eliminated or at least ameliorated during the course of the review as the soft synth updated itself. Now the less good news. Some remain. For example, you obtain the common and undesirable do‑do‑da‑de effect when you step through voices whose oscillators are tuned to the same frequency. This suggests small, random offsets, but the situation is more complex than that. If you set up a patch with the oscillators in tune with one another and then turn the volume control of one of them to zero and back again, you’ll obtain a different tone. Because there’s no drift, the resulting inconsistencies between voices can be quite severe, although detuning the oscillators a touch helps to minimise this. Another issue lies with the Elka‑X’s implementation of oscillator sync. Programming this is unusual, and the downloadable manual currently refers to controls that don’t exist. Then, once you work things out (you obtain sync sweeps using the downward glide option in the Glide/Portamento pane) the sweep is linear and comes to an abrupt halt when it reaches the root pitch, which doesn’t sound great. I also felt that the whole synth was just a tad too well behaved; sure, the Synthex used DCOs, but there could still be minor variations between voices. But, on the bright side, all of this can be fixed if Cherry feel it appropriate to do so.

So let’s now return to the question... does it sound like a Synthex? In my view, if you ignore the additions to the architecture and sidestep the issues discussed above, it emulates the sound and flavour of the Synthex rather well and can sometimes sound remarkably similar. Is it identical? No, you’ll always find differences between an original synth and its digital recreation if you go searching for them. But as I’ve asked before, do you really want to spend your time hunting for minutiae or would you prefer to be playing music? Consequently, a much better question is, ‘does it sound good?’

Stepping through the factory patches shows that it can sound excellent and, when I programmed my own patches, I quickly and easily obtained some first‑class results, most of which were far beyond the dreams of an original Synthex, especially in terms of performance capabilities. I particularly liked the clear and transparent top end, courtesy of the 4x oversampling that all but eliminates aliasing. Inevitably, this places significant demands on the CPU, especially at high sampling rates with short buffers so, when recording, I would be tempted to render its output one Elka‑X at a time rather than sequencing multiple instances simultaneously. I would also test it extensively before using oversampling on stage. But I’ll rarely complain about nice sounds being reproduced with high fidelity.

The Elka‑X is an affordable and musical alternative to a synth that very few will ever have the chance to own.


The Elka‑X is an affordable and musical alternative to a synth that very few will ever have the chance to own. There are still some teething problems to be addressed but, given the frequency with which I received updates during the course of this review, it’s reasonable to hope that Cherry Audio are tackling these (as well as any others that I may have missed). In the meantime, it really can sound rather nice, and the additions of velocity, aftertouch and the excellent MIDI implementation make it far more useable than a real Synthex. It also costs less than one percent of what you might pay for the original. I like it today. I hope to like it even more tomorrow.

The Arpeggiator & Sequencer

The Elka‑X includes an arpeggiator with up, down, up/down and random modes, key hold, up to four octaves’ range, a frequency range of 0.25Hz to 30Hz, and the ability to sync to clock with ratios ranging from eight beats to 1/64th triplets. But more significant than this is the updating of the Synthex’s four‑track step sequencer because, by simplifying its operation and adding some useful utilities, Cherry have made it relevant again.

To program it, you first select what it affects — the whole synth in Single mode, or the Upper or Lower layer. (Although it can play four notes simultaneously, it is strictly mono‑timbral.) You then choose the track that you wish to record and enter up to 128 notes and rests. If wanted, you can record the other three tracks in the same way and, since none of these have to be the same length as the others, you can create lines that, when looped, rotate against one another. You can synchronise playback with the LFOs, delay line and arpeggiator, and this can be magical because you can arpeggiate the output from the sequencer to generate almost stochastic compositions. You can also transpose the output in real time, or play along with it as you choose. I didn’t expect to get much use from the sequencer but I found that it’s one of the most creative and satisfying elements of the Elka‑X. However...

The arpeggiator and the sequencer always output notes with a velocity of 127. Not only does this make everything loud, it means that the filter of any sound programmed to respond to dynamics is always maximally open, which is very different from what you would hear when playing the same sound from your keyboard. Investigating this also revealed that the limiter distorts when pushed hard, so it’s particularly important to keep the combination of the oscillator, layer and master volumes at suitable levels when using the arpeggiator and sequencer lest things start to sound nasty.


The Elka‑X has great potential. I view it as a work in progress but, even before any remaining issues are addressed, it’s a very pleasing and affordable addition to your arsenal of soft synths.