Arturia Solina V

Arturia have revived a classic string synth in software, and the result is the Solina V.

Until Ken Freeman invented string synthesis by detuning and modulating three oscillators per note to create a chorus effect, the only ways to obtain an ensemble sound had been to play a Mellotron or to book an orchestral ensemble. Then, in 1972, a similar sound was heard to emanate from Eminent’s 310 Unique home organ, which employed a single oscillator per note, but thickened up the sound using a bunch of modulated BBD chips (analogue delay lines) to create the ensemble effect. The 310U had little impact when it was released, but Jean Michel Jarre’s ‘Oxygene’ (1976) ensured that it would later be recognised as one of the most important keyboards of the 20th century.

In 1974, Eminent extracted and extended the ensemble technology from the 310U and repackaged it in a four-octave keyboard. Small and light by the standards of the time, the Solina allowed bands to add string sounds to their arsenal without lugging around heavy and often unreliable instruments (or viola players). So they did. Pink Floyd were amongst the first, but Solinas were soon gracing the music of artists and bands as diverse as Elton John, Fleetwood Mac, the Eagles, Gentle Giant, Patrick Moraz, the Enid and Renaissance. Even Herbie Hancock and Stevie Wonder were to be seen playing them.

Numerous imitations soon appeared, and many of these improved upon the Solina. My favourite was the Logan String Melody 2, but there were many others worthy of acclaim. Nevertheless, despite its limitations, the sound of the Solina remained desirable. Even today, nearly every sample-based keyboard offers patches named ‘Solina this’, or ‘Solina that’, although none of these sound exactly like the original. So, can a physically modelled recreation do better?

A Direct Comparison

I liberated my trusty second-generation Solina (CV and Gate interfaces, but monophonic audio outputs) from storage, and set it up with an Arturia Analogue Experience keyboard sitting on top so that I could switch between the soft synth and the original. My initial tests were encouraging. I compared every Upper (paraphonic) voice and every Lower (monophonic bass) voice on the Solina with its equivalent on the soft synth and, in my view, the differences between them were unimportant. Without an A/B test I would never have known which was which, except that the Trumpet sound on my Solina has always been a bit hissy, and the soft synth’s isn’t. I then tested the various combinations of waveforms. Old analogue mixers (including those built into 40-year-old keyboards) don’t necessarily add sounds linearly but, within a narrow margin, everything was as it should be.

The amplitude envelope of the Solina’s paraphonic section is defined by just two parameters: Crescendo (attack) and Sustain (release). I found the shape of the soft synth’s crescendo to be a tad different from the original’s, giving a slightly different character to sounds that speak slowly. It’s also calibrated as a loudness (dB) rather than a time (ms), so that’s a bug, albeit one that doesn’t affect the sound in any way. At the other end of the note, the amount of ‘suck’ generated when re-initialising the envelope during the release phase was less pronounced on Solina V, and that’s a good thing. The contour of the bass voices was even simpler on the original; it responded to the crescendo control, but the length of the release was short and fixed. I found that the Lower amplitude envelope on Solina V is somewhat faster, making it possible to articulate notes more quickly, but losing something of the smoothness when playing legato. However, this can be adjusted by the (as yet unrevealed) Bass Section, so please read on...

Solina V emulates the design of the original Eminent instrument very closely with (if you look carefully) just three visible enhancements. Now it was time to engage the all-important ensemble effect, and this was where Solina V and I had a bit of a falling out. Before explaining why, let me preface the next few comments by stating that the ensemble effects on any two vintage Solinas can sound somewhat different from one another; quite apart from the inherent variations between units, you can open up a Solina and adjust the rates of the tremolo and chorus oscillators as well as the contribution levels of each of the three signal paths that create the ensemble effect. So now you’ll want to know how accurate I found the ensemble on the soft synth to be. It pains me to write this, but it isn’t right; at around 1.3Hz, the lower frequency modulator is running much too quickly and its effect is way too prominent. Whereas the genuine Solina, with a slow modulation speed of a little under 1Hz, creates a luscious wash of sound, Solina V warbles. Don’t take my word for it: Eminent’s original 1972 patent for its ensemble effect states that, “The effect of the present device is most apparent when the frequency of the signal generated by the first generator means is at most 1Hz”. Fortunately, this shouldn’t be too difficult to correct should Arturia choose to do so. If I have to find a positive (and I do) it’s that the halo of noise that accompanies the Solina’s ensemble effect (which was inevitable given the technology in 1974) isn’t present. Before anyone claims that this destroys the character of the original in some arcane way, I would remind you that sound engineers spent a great deal of time, effort and money trying to make the Solina sound as clean and quiet as possible, so this is a good thing.

Resonators, Vox Humana & Effects

With its lid closed, Solina V offers just three additional controls when compared with the original: a Volume Upper in addition to the Volume Bass and Volume Master of the original, plus pitch-bend and modulation wheels. But click on the lid, and the GUI opens to display a few dozen extra controls, while yet more appear on either side of the keyboard itself and, intriguingly, a new Upper timbre — Humana — appears.

For me, the most significant of the new facilities is the Upper Resonator. Often called a formant filter, this acts on the Upper voices and is based on the equivalent panel on the Polymoog Synthesizer. It comprises three filters with predetermined ranges of 60 to 300 Hz, 300Hz to 1.5kHz, and 1.5 to 7.5 kHz, and each can be bypassed or act as a low-pass, band-pass or high-pass filter, although all three have to act in the same way at any given time. Each offers an independent cut-off/centre frequency, variable resonance, and gain of -72dB to +6dB. You can get a coarse idea of what this does by raising the resonance and gain of all three filters to maximum and stepping through each of the filter profiles in turn. Adjusting the cut-off/centre frequencies in each case will then allow you to sculpt all manner of exaggerated timbres. Inevitably, the nicest sounds are achieved with less extreme settings, and these can yield excellent results.

Given the presence of the Resonator, you would expect the Humana voice to be an approximation of the Vox Humana preset on the Polymoog Keyboard. It occupies the same sonic space, and you can use it in the same way — Gary Numan’s ‘Cars’ is never more than a tweak or two away — but purists will claim (with some justification) that it’s unlike the unstable warble of the old Moog. Whatever your stance, it’s a doubled and slightly chorused sound that offers some movement even without the application of the onboard effects, and very useful it is too.

Opening the top panel reveals a wealth of additional facilities that take Solina V through ARP Polyphonic Ensemble territory, past the Moog Polymoog, and into areas that the original string synth could barely have imagined. Ah yes, the effects... At first sight, Solina V seems to be endowed with three additional effects — chorus, delay and reverb — whose contributions are controlled by the three knobs to the right of the keyboard. But this is misleading. FX1 offers four effects — a dual-stage phaser and three complexities of chorus — although only one can be selected at any given moment. Following this, FX2 offers both single-channel analogue (solid state, not tape) and stereo digital delay models with maximum delays of up to one second. Finally, FX3 offers 24 parameterless (is there such a word?) reverbs, 18 of which are derived from Roland’s DEP5 and RSP550 rackmounts. Why these? I haven’t a clue. They are neither contemporaneous with the Solina, nor do they have the caché of something like a Lexicon or Eventide. Perhaps someone in Arturia’s development team had a couple knocking around the studio.

The Bass Section

The area in which Solina V most exceeds the capabilities of its inspiration has to be the bass section. Where the original offered just two sounds — generally used to emphasise the lowest note played — the soft synth adds a resonant 24dB/octave low-pass filter with an independent AR contour generator offering maximum attack and release times of 3s. For this to be meaningful, the bass section also needs a dedicated amplitude release, so this is also added, although, in keeping with the nomenclature elsewhere, it’s called Bass Sustain. You have to be careful with this filter; at one point during this review, I was monitoring on bookshelf speakers, and I nearly popped the cones out of their enclosures as its self-oscillation swept down toward 20Hz.

I was even more surprised to find an arpeggiator in the bass section. Happily, there’s nothing complex here — just up, down, up/down and random modes with On and Hold options, and tempos in the range 0.01 to 50 Hz or sync’ed to MIDI with a sensible range of clock options. Used together with the bass filter, you can obtain a number of unexpected sounds from this. You can also transpose the output from Lower up a couple of octaves and the output from Upper down a couple of octaves to place the pitch of the arpeggio above whatever you play with your right hand, which suggests further possibilities. Mind you, it would be even better if the effects could be applied individually to the Upper and Lower voices so that a precise bass arpeggio could be married to a whooshy chord sequence.

Modulation

The mod wheel (which, like the pitch-bend wheel, is constrained by an Amt knob when the lid is open) controls the amount of LFO modulation directed to the global vibrato and tremolo, as well as to the filter cutoff frequency of the bass section. The LFO itself offers five waveforms, key-sync, and controls for the rate (0.01 to 13 Hz or sync’ed to MIDI), for the delay before the onset of modulation, and for the length of fade-in. MIDI sync is particularly useful when other aspects of the synth are also synchronised, because you can then create sounds in which the modulation and other effects are in some way linked together. However, there’s a bug here. When you use delay and fade-in, but release the note before these have completed their durations, the LFO depth jumps instantly to maximum when you release the note, and its release phase is subject to the full modulation effect. You might find ways to use this, but it’s wrong.

In Use

Solina V offers a great deal more in the area of performance capabilities than you might expect. For example, it’s both velocity- and aftertouch-sensitive, and the Master section contains the four knobs that affect how the overall loudness and brightness of the instrument responds to these. While the way that these work may seem a little odd at first, it’s clever; in all cases, turning up the control reduces the initial level (therefore quietening or dulling the sound) to an appropriate degree so that a high MIDI Velocity or MIDI aftertouch value (or a suitable combination) brings you back to the original sound. Well, that’s the theory, but you can overdrive your sounds considerably if you’re not careful. As long as you’ve left some headroom, you’ll be fine. If not, distortion may ensue. Oh yes... and while you can generate different MIDI velocities without a controller keyboard by pressing the keys in the GUI at different positions — from low values at the back to higher ones at the front — you can never generate a velocity as high as 127 in this fashion, which would seem to be another bug to me. On the other hand, there’s a nice trick to be discovered here. If you increase the aftertouch brightness sensitivity, but don’t send aftertouch to Solina V, you can filter the output of the whole instrument to create a much darker range of sounds, which is great.

In common with other Arturia soft synths, Solina V offers remote control over almost every facility, with a simple system for assigning MIDI CCs to voicing parameters and another for saving and recalling controller configurations.In addition to all this new sensitivity, the menu bar also offers an unobtrusive button marked POLY that switches the Upper voices on Solina V between the authentic paraphonic response (many notes but just a single contoured audio path for all of them) and a true polyphonic response (many notes, each with its own contour). You’ll be amazed at the difference that this can make.

You’ll wring some of the best sounds from Solina V when controlling its parameters in real time using MIDI CCs. To do so, switch on Assign Mode, and every control (with the exception of the pitch-bend wheel) goes purple. If you then click on one, a window pops open and you can create an assignment by tweaking something on your MIDI controller. The control itself then goes red to indicate that it’s assigned, and you can then choose maximum and minimum values for the parameter (which can be set ‘upside down’ to invert the response if wanted) before closing the window. Unfortunately, it doesn’t seem to be possible to assign multiple parameters to a single control, which is a great shame. For dramatic effect, there’s nothing like (for example) increasing the Qs of the filters in the Resonator while adjusting the depth of the effects and increasing the speed of the arpeggiator. Perhaps Arturia could look into this for a future revision. Happily, you can save your MIDI configurations so that, if you want to play Solina V from different controllers, you need only recall a saved setup.

As for the sound, it can be superb, and I would particularly recommend spending some time getting to know the effects because some of my favourite patches were created by applying a wash of the onboard effects to individual voices (often without the ensemble), using the Master Section to darken the initial sound, and providing expression with velocity or aftertouch. For example, the dual phaser stages can process both sides of the stereo signal simultaneously, or Stage 1 can affect the left signal while Stage 2 affects the right. This makes it possible to create huge, whooshing panoramas by using slightly different rates, feedbacks and depths for each. Likewise, the three complexities of chorus each offer the usual rate, amount and delay controls, but also add stereo rate and width knobs that apply a second ‘out of phase’ two-channel modulation for a huge range of spatial, chorused and alternative ensemble effects. Sadly, there’s another bug here; if you turn the mix of any effect to zero, the current sound is held within its delay lines and then emerges when you next turn the knob away from zero. This can be rather unpleasant so, until it’s fixed, you should only turn the mix knobs to zero when no sound is being produced. Notwithstanding this, I could rant for hours about the range and depth of the pads, ensembles and quasi-vocal analogue choirs that seem to pour out of Solina V when you start to plumb its depths. You can even construct sounds that sound choral at low pitches and orchestral at higher ones. But if you’re not ready to jump in at this level, the factory presets will give you a flavour of what’s possible. These are accessed using the company’s standard Banks/Types menus, and you can save, recall, dump and load your own creations in the normal fashion, so we don’t need to discuss this further here.

Finally, I’ve noticed that some users are complaining about the CPU load demanded by Solina V when the reverb is used. It’s true that I had to extend the buffer to 512 samples when placing heavy loads on the synth, but even the higher of these figures represents just 11.6ms at 44.1kHz, which is less than the latency experienced when playing some early but revered digital hardware synths. I’ve never been able to decide whether people who claim that a latency of a small number of milliseconds renders an instrument useless are posturing, or whether they are genuinely affected by it. Either way, I view it as inconsequential for a string synth and, with a buffer of 512 samples, I couldn’t get Solina V to exceed 25 percent load, even with everything turned on and my arm laid across the keyboard.

Conclusions

With one caveat, I think that Solina V is a very nice instrument, quite different from other soft synths. However, I can’t give it a clean bill of health because of its ensemble effect. It’s not that it’s bad, but, whether in mono or stereo mode, it’s not what I consider to be the Solina sound. Of course, I can mask its warble using vibrato, chorus, delay and reverb (and create some lovely new sounds while doing so) but, despite what Arturia’s endorsees claim, the ensemble effect is wrong. But perhaps I can turn this criticism into an opportunity. I suspect that the underlying algorithm is correct so why not, as on the rather splendid Godwin 749 String Concert, place the modulators’ depths under the control of the player? Having done so, why not go further and allow the user to tweak the rates of the LFOs driving them too? Notwithstanding a few bugs to be swatted, this would eliminate my only significant criticism, and allow users to fine-tune the instrument in ways that are much more fundamental to its sound than things such as bass filters and arpeggiators. In the meantime, Solina V is capable of creating some luscious sounds of its own and is well worth investigation.