By Jim Aikin
THIRTY YEARS AGO, ANALOG SYNTHESIZERS (AND THERE WASN’T ANY OTHER
kind) were accused of sounding cold and sterile. Today they’re sought after for their
warmth. Go figure. If I had plenty of money and a spare room, I’d buy a few monster
analog synths. As it is, I have several excellent virtual analog plug-ins to choose from.
But nothing I’ve heard yet nails the analog sound as solidly as U-he Diva. This soft
synth is so hot, it’ll raise blisters.
Diva’s lineup of modules is flawlessly classic.
Even so, there’s a lot more here than meets the
eye. Across the top (the black panel) are four
mega-modules. The first houses the oscillators,
the second a highpass filter, the third the main
resonant filter, and the fourth a pair of attackdecay-
sustain-release envelopes. The lower area
(the red panel) houses two LFOs, two effect
processors, extra modulation routings, and voice
Each of the mega-modules offers three or four
different choices. For example, in the oscillator
panel you can choose the Triple VCO, Dual
VCO, DCO, or Dual VCO Eco (less CPU usage)
mega-module. The Triple VCO is reminiscent
of, though certainly not identical to, the oscillator
layout on the Minimoog. Other modules
seem to be based on early Roland synths or the
Korg MS-20. I’m told the developers listened
to hardware instruments in the studio while
Diva has no arpeggiator—an odd omission
for an analog-emulation synth that does such
a fine job of sounding authentic. The effects
choices (chorus, phaser, plate reverb, stereo
delay, and rotary speaker) aren’t luxurious, but
they’re the right choices for a vintage sound. On
the plus side, Diva has a rich set of controls for
adding subtle analog instability to the sound,
such as individual voice detuning and multiple
“slop” controls—more on these later.
The user interface is clean and professional.
Double-clicking any knob zeroes it, and rightclicking
arms the MIDI Learn. By shift-clicking
a knob, you can make fine adjustments, which
are displayed numerically. In use, Diva was rocksolid:
I didn’t encounter a single glitch.
Th e patch browser occupies the whole panel
by itself, and includes hundreds of mouth-watering
presets. After saving your own patches, you can
drag them into whatever category folder you like.
Patches can be marked as favorites or junk, but
I’ll be using the favorites mark a lot more than
the junk mark. Whether you’re doing dance,
hip-hop, new age, or fusion, you’ll find plenty
of leads, basses, and pads to get the creative
By now, you’re probably realizing as I did
that there are too many features to list, much
less talk about how they sound, so we’ll hit the
The Triple VCO has a Minimoog vibe, right
down to the rocker switches, but the waveform
selector, unlike the one on the Minimoog, is continuously
variable from sawtooth and triangle
through pulse. Oscillators 2 and/or 3 can be
synced to oscillator 1, which itself can be an FM
source to modulate 2 and 3. You can switch modulation
on or off individually for the pitch and
waveform, but all three oscillators will share the
same source signals and modulation amounts,
which is a limitation.
Modulating the waveform of the Triple VCO
from triangle to saw produces a slight and very
analog-sounding pitch instability. As the modulating
signal speeds up, the pitch changes get
larger. Though I’m a bit skeptical, I’m going to
call this a feature.
The Dual VCO doesn’t offer continuously
variable waveforms, but it does have pulse
width modulation and sync. Oscillator 1 can
cross-modulate oscillator 2, with an input for
the modulation amount.
The DCO module has a single oscillator with
two waveform selectors (the two are mixed), a
sub-oscillator with its own waveform, and a noise
source. All of the oscillator modules have dual
inputs for pitch modulation; in the Triple VCO
these inputs can be switched on or off individually
for the three oscillators.
According to the manual, some fancy signal processing
is being employed to produce low-latency
resonant filters. The difference is subtle, but I feel
the filter tone is more “present” than in other
modeled analog synths I’ve played. The filter
modules are called Ladder, Cascade, Multimode,
|An alternate view of Diva, showing the Dual VCO, Bite filter, envelope sliders, and the voice trim panel (lower center).
The Ladder filter is Moog-like, right down to
calling the resonance knob “Emphasis.” This filter
will self-oscillate, and can be modulated at audio
rate from oscillator 1 for rich, swirling sidebands.
Th e Cascade fi lter has a clean/rough switch and a
12/24dB switch. The Multimode filter is similar,
but replaces these two switches with a fourposition
switch for four-pole lowpass, two-pole
lowpass, highpass, or bandpass.
At first I didn’t hear anything special from the
Bite filter, but it turns out this filter is sensitive
to input level. Turning the oscillators’ output
knobs to a low level can give you some grungy,
unstable tones that are very satisfying. I got
some truly nasty tones by combining the Bite
filter with the Triple VCO’s feedback routing.
All the filters have three modulation inputs
for cutoff frequency (two assignable and one
hardwired to key number), plus modulation
inputs for resonance and FM, the latter being in
the lower panel.
The two envelope generators are individually selectable
from three types: an analog-type ADSR,
a similar digital ADSR, and a very Minimoogstyle
ADS (attack, decay, and sustain only) with
an on/off release switch. All three have controls
for velocity sensing and keyboard scaling. Velocity
controls the envelope amount, not the attack
and decay times, which are controlled by keyboard
The digital envelope has mysterious buttons
labeled Q and C. According to the manual, the Q
button quantizes the envelope output, producing
a rougher sound. I just about couldn’t hear the
difference. The C button changes the envelope
segments’ shape, but again in such a subtle manner
that I had trouble hearing it.
By default, many of the modulation inputs
on the upper panel come from envelope 2
and LFO 2. From a pop-up menu, however,
you can select numerous other sources, such
as MIDI pitch-bend, mod wheel, aftertouch,
breath controller, or velocity; either of the
LFOs or envelopes; or any of the processors in
the Modifications panel.
These processors add a lot of versatility.
With the lag processor, you can smooth out the
edges of an LFO’s square wave. The add processor
lets you mix two modulation signals, albeit
with no control over their individual depth. The
quantizer produces a stepped output from a
smooth input, with precise control over how far
apart the steps will be.
The LFOs aren’t exactly vintage in design,
though they can be pushed up into the audio
range for weird clangorous effects. In addition
to the standard waveforms, they’ll do random
stepped and random glide. You can sync them
to various metric values, and modulate both the
rate and the depth from various sources.
Also in the lower panel are eight supplemental
modulation routings: inputs with trim
pots for two oscillator parameters (FM and
cross mod depth, noise and dual VCO mix), two
filter parameters (resonance and FM amount),
feedback depth, main volume (normally envelope
1, but can be set to gate), and volume and
***Click here if the player below doesn't load correctly.
Voice Trims and Global
In addition to the usual master parameters,
Diva gives you some features that are less standard,
such as separate glide times for oscillators
1 and 2. You can load a microtuning scale,
stack and detune multiple voices, and choose
low-, high-, or last-note priority. In duo mode,
oscillator 2 tracks the highest keyboard note
and oscillator 1 the lowest, a nice performance
option borrowed, if memory serves, from the
Diva is up to 16-note polyphonic (if you have
a fast enough computer—I don’t). Its voice trim
section has eight columns of knobs for detuning
the voices. (Voices 9–16 duplicate the settings for
voices 1–8.) You can detune each oscillator in microscopic
amounts. This very effectively simulates
the sound of synths like the Prophet-5, which had
individual analog circuits for each voice.
Adding similar off sets to cutoff , envelope,
pulse width, and glide time would involve way
too many knobs. Instead, you get four knobs—
one for controlling the overall amount of “slop”
for each of these things, together with a randomize
button that invisibly juggles the off set
amount for each voice.
The Drift knob has a very subtle but audible
effect. Unlike detuning from voice to voice (and
thus from one note to the next), this knob introduces
a slow, tiny, random pitch drift while a
note is sounding.
Along the bottom of the voice trim section
is a row of knobs called Voice Map Modulators.
These are a modulation source you can
use anywhere. If you want rotating pitch
offsets, for instance, you can route Voice Map
to oscillator pitch. This setup effectively emulates
some of the possibilities of early Oberheim
I was a bit ho-hum about a soft synth with no
wild-eyed digital features—until I started playing
Diva. After that, I just about couldn’t stop. An
arpeggiator and a step sequencer would be welcome
additions, and wouldn’t break the vintage
analog concept. Other than that, Diva has it all.
If you’re jonesing for analog, this synth will blow
your doors wide open.
PROS Extremely authentic
analog sound and features.
Modeled on vintage hardware
of great presets.
CONS High CPU usage. No
Up to 16 voices, CPU-dependent.
SYNTHESIS TYPE Analog
FORMATS VST or AU plug-in.
(Pro Tools version planned.)
SYSTEM REQUIREMENTS Mac:
OS 10.5 or later. PC: Windows XP
or later. Both: Dual-core Intel or
AMD (but not PowerPC) processor,
2GHz or faster.
I can’t believe it’s not analog.