I think the most important thing really is just getting the routings right,
if the signal flows through all the right things and you have all the
right modulations available, you're going to be about as close as you
can get and you ought to have a usable and fun synth in any case. Mixers
are easily the most useful module, since they can gather all the modulation
sources and set their levels prior to modulating pitch or whatever.
A few things that might not be totally obvious. For example, the
Octave Cat is a 2 VCO synth with 1 LFO and 2 envelope generators (ADSR
and AD). However, to emulate it, I had to build a 7 oscillator synth
with 2 LFO's and 2 ADSRs, just so all the suitable signals were always
present in the same way. But a synth like this is one heck of an
instrument, especially if you break from some of the limitations of an
accurate emulation (tune the oscs separately, add modulations etc).
It is useful to use master and slave oscs, and you should note a
couple things that might not be obvious. A master will propagate its
pitch modulations to its slaves, but not its FM modulations, so you
have to FM modulate the slaves the same way with similar signal
routings as the master. Also, when a master osc is sync'd it doesn't
sync the slaves, so you have to do that routing too, though not all
slaves have a sync input.
Some of the old monosynths would have made superior
polysynths, which makes them fun to emulate. You can do "what-if"
exercises, such as "what-if" my ARP Odyssey had a vocal tract filter.
Finally, the one thing that you can't easily copy is the instability and
signal overshoots of VCO's, it's helpful to remember that the nord oscs
are perfect in shape and stability, so you might want to take steps to
compensate, like mixing in noise, or modulating with a bit of noise
or adding FM from an unpredictable source. AM can be useful here too.
15.2 Yamaha DX7
by Wout Blommers
Wout Blommers has undertaken an extensive project to convert a
vast library of DX7 sounds to the Nord Modular. In so doing, he
wrote a nice description of FM synthesis and an explanation of how
each of the DX7 "algorithms can be mapped to the Nord Modular. His writeups
are available in the chapter
DX7 to NordModular.
The results of the conversion project, some 9400 different Yamaha DX7 patches, can be found at http://nordsynth.zevv.nl/010_NordModular/011_Patches/FM_for_Xmas/.
An example of one of these converted patches is shown in the figure below:
Editor Conventions
Transient Generator section
Lower Left section
VCO1 section
VCO2 section
Filter section
VCA section
Knob organization
Interesting modifications
I sold my original white Odyssey to finance the purchase of bigger synths,
and then bought another early black face Odyssey several years later during
a sudden attack of gear nostalgia. I've used it from time to time, but
maintenance is an issue, as it is with all vintage analog synths. Thus, it
seemed like a good idea to replicate the Odyssey in the Nord Modular.
The Odyssey used an ingenious system of patch switches to quickly select a
variety of patching configurations. This was a good compromise between
versatility and speed, without the hassles of patch cords or the limitations
of a completely hard-wired synth. My goal was to replicate the Odyssey as
completely and accurately as possible, including the method used to set up a
patch. This meant devising a basic patch architecture that included patch
switches instead of requiring the user to draw patch cords.
On the Virtual Odyssey, the patch switches are represented by the switches
in the 4-to-1 switch modules, and the names for these modules show the
available switch selections. For example, the module labeled "NG/RING MOD"
selects between the noise generator and the
ring modulator as the input to the first knob in the audio mixer.
Thus, by simply selecting switch positions and turning virtual knobs, you
can replicate any sound that was possible on the real Arp Odyssey, without
drawing any patch cords.
In fact, you can make sounds that were not possible on the real Odyssey,
because some parameters on the NM have a greater range than the equivalent
parameter in the real Odyssey. For example, the LFO can run slower and the
envelope release times can be longer than the same parameters on the real
Odyssey.
There are some modules which are used to
process the Nord's signals to more closely resemble the signals on the
Odyssey. The waveforms in the Odyssey's audio oscillators and LFO were
positive-going only, unlike the bipolar waveforms in the Nord Modular. Thus,
a collection of Level Shifter/Inverter modules were used to alter the Nord's
waveforms. Although an audio oscillator sounds the same whether it is
bipolar or unipolar, it makes a difference when you use these waveforms as
modulation sources.
Since the Odyssey had more sliders than the Nord has knobs (I think the
Odyssey has 32 sliders), I opted not to assign any parameters to the Nord's
front panel. I'll leave it up to the user to assign them as needed when
setting up a patch.
The real Odyssey had one major feature that I was not able to replicate on
the Nord. The Odyssey had a unique "duophonic" keyboard. When you pressed
two keys, OSC1 played the low note and OSC2 played the high note. This was
useful for producing distortion effects when using the ring modulator, among
other things. Note that this is not the same as simply setting the Nord to
have two voices. If you get ambitious, please feel free to try to figure out
how to achieve the duophonic effect.
One minor difference between the Nord and the Odyssey is the way the filter
responds when resonance is set to maximum. On the Nord, you must send some
amount of audio into the filter to trigger filter self-oscillation. This was
not necessary on the real Odyssey.
The one feature I added, which is not true to the original Odyssey voice
architecture, is the additional Vibrato LFO, which is controlled by the mod
wheel of a midi controller keyboard (via the red morph group). The Odyssey
had really lame performance controls, and I just had to do something about
it. I hope the purists are not too offended :-)
Knobs in 4-to-1 switch modules should stay at MAX. Switches in 4-to-1
modules act as the patch switches across the bottom of the panel on
the real Odyssey. Module names in the 4-to-1 modules indicate switch
selections. Example - the switch module below the VCA is labeled
"ADSR/AR". Button 1 selects ADSR, button 2 selects AR to control the
VCA. GREEN cables - audio signals routed to patch switches (HIDE).
PURPLE cables - control signals routed to patch switches (HIDE). RED
cables - hardwired audio signals BLUE cables - hardwired control
signals YELLOW cables - gate & trigger signals (HIDE).
Some example patches illustrating the use of this template are:
The first try-out we did was during a gig and the sound was so terrible,
the organization of the party threw us out!
After that failure we build the Hallsey Mark 2.
(The name Hallsey was chosen 'cause our bassman made MARSHALL on two
plates,
MARS did fell off the fake amplifier and we added 'sey')
The Mark 2 was a real modular synth, only its oscillators wouldn't tune
in line...
(It had sample&hold, mixers, LFO's and all that stuff)
The Mark 2 was stolen during a gig and some days later our leadsinger
found
the poor thing at the side of a road, 'cause the thieves couldn't make
any sense out of it, me think.
Mark 3, of course also from the famous 'Hallsey factory', was, so I
believe, one of the first Synths
which used AD and DA converters. Two of our gang worked at the
university, where such
electronic equipment was developed for Philips. (No, they didn't steal
them: it was just an
university, you know, where the professors also carry uranium 238 in
their pockets)
I think only the Mark 1 is interesting to emulate here ;-)
Played in the lower regions, it was a wonderful beatbox, but we didn't
knew then...
(Press the low C and hold it a while)
Those Shapers were really there. The LFO's in a way too.
And the distortion remains!
ARP 2600, by Roland Kuit (this is just a basic template -
Roland has created many more patches using this template).
15.3 Octave Cat
by Sam Streeper
This patch mimics the architecture of the old Octave
Cat analogue synth, my favorite monosynth. As much as possible,
I also tried to duplicate the sound and feel of that synth.
I named all the interesting modules in this synth. If there are no
symbols preceding the name, either I am indifferent about editing the
parameters in this module in the editor, or the values are important
enough that I exported their control to an external controller. If I
put a '+' symbol before the name, that indicates a module that has
values that I want to encourage myself to tweak in the editor, like
significant or interesting modulations or routings that I didn't
export. In the case of the switches a '+' usually indicates that I
intend to try different routings, but the mix levels in the switch may
be specifically set and I generally won't mess with those. If a
module name is preceded by a '!', that indicates a structural element
or logic setup that is carefully set (a necessary routing or a test)
that I generally don't intend to mess with. In other words, these
tweaking these modules is more likely to break the patch than to
provide interesting modulations. I don't think I have any such
discouraged modules in this Moctave Cat. In this document if I refer
to the Cat I mean the Octave Cat, not the Moctave Cat.
The ADSR amounts are controlled by the sliders on my QS8 keyboard (my
master controller), which are set up as controllers 12, 13, 91, and 93
respectively. The Attack/Release (AR) sliders are not exported. They
can be set in the "+AR" module, which looks suspiciously like an ADSR
envelope. The Cat's "ADSR repeat" function is not exported. It can
be set in the editor between off and gated in the "+ADSR repeat"
module. The additional non-gated auto-repeat functionality of the Cat
can be achieved by decoupling the envelopes' gates from the voice gate
and holding it high. The Cat's sample and hold switch is not
exported, but it can be set in the editor with the "+S&H source"
module. It samples either VCO1 or noise, and is triggered by the LFO.
The duties of the pitch slider are relegated to the pitch controller.
The Cat's octave shift button is not implemented as I have a big
keyboard. The Cat's glide (portamento) slider is not implemented in
the Moctave Cat; it should be there but I haven't figured out the
right way to do it yet. The LFO frequency is exported on knob 1. The
Cat's LFO supplies a square wave (this is also used to trigger the
ADSR repeat if enabled) and a triangle. In the Moctave, I substituted
a sine wave for the triangle, since it gives a better sounding pulse
width modulation and smoother vibrato and otherwise doesn't matter on
my Cat patches. I think the original Cat designers would have used a
sine wave here if they could have generated it inexpensively.
Coarse tuning is on knob 4, fine tuning is not exported but can done
in the editor by tuning the VCO1 master osc and the 3 slaves. The
slaves are not phased accurately to the real Cat which means you can't
really do that goofy Cat note-sequencing thing accurately, but I've
always thought that that's more interesting than useful. As long as
the phases are wrong, I slightly detuned the slaves because I like
that sound better. The Cat's VCO1 keyboard control switch isn't fully
implemented. If you bump the Moctave polyphony down to 1 voice, then
the Moctave emulates the Cat's mono mode, and if you turn off
keyboard tracking in the "VCO1 pulse" module, it's like turning this
switch off. I don't thing the Nord can really emulate the Cat's weird
poly mode, since that would require a bit of logic processing on the
uncooked keyboard status and the Nord really only forwards processed
keys. The VCO1 modulation routing switches are not exported, but they
can be set in the editor. The first switch selects LFO sine (my
default), LFO square, or sample&hold as the modulator and is labeled
"+VCO1 mod1". The second switch selects ADSR, the AR envelope, or
VCO2 (default) as the modulator, and this one is labeled "+VCO1 mod2".
The Cat has a single knob to select the amount of modulation for each
of the mod sources, and these knobs are in the "VCO1 modamt" module
and exported as knobs 2 and 5. The VCO1 mod amount module also mixes
in a bit of weather modulation since my VCO1 is not entirely stable.
I did not have enough free modules to accurately model the effects of
weather on my Cat though! The Cat allows you to either set the amount
of VCO1 pulse width modulation or to dial in the width of the pulse.
This functionality is not exported but is found on the "+PulseMod
switch" module. Unlike most of my switches, I do need to set the
knobs on this module corresponding to the selected switch. Finally,
VCO1 supplies 4 waveforms: a suboctave square wave (1 octave down), a
sawtooth, a triangle, and a pulse. You set the mix level for each of
these waveforms on knobs 3, 6, 9, and 12. These mix levels determine
both the sound of VCO1 and the degree of FM modulation of each VCO1
waveform on VCO2 if you have dialed in FM on VCO2. (That's knob 11 and
the default selection of switch 3 on the "VCO2 mod2" module for those
of you who have read ahead...) Similarly, the levels of the VCO2
waveforms will affect the amount and sound of FM in this VCO1 section
if you left the default routing alone and dialed up knob 5. By the
way, if the knob assignments aren't obvious (and I don't expect that
they should be yet) there's a simple organization to them that I will
explain near the end ofthis document.
VCO2 coarse tuning is exported on knob 7. VCO1 can be synced to VCO2
with the switch in the "+Sync VCO1" module, though this doesn't really
sound like the Cat's sync, and I haven't looked into why. On the Cat,
you can get a nice percussive sound by using AR as the main envelope
and using a quick ADSR envelope with 0 sustain to modulate the synced
pitch of VCO1. I haven't tried this yet on the Moctave Cat, though.
This section largely mirrors the VCO1 section, so "+VCO2 mod1" selects
LFO sine, LFO square, or S&H as a VCO2 modulator, and "+VCO2 mod2"
selects ADSR, AR, or VCO1 as a modulator, and the mod levels are set
in "VCO2 modamt" and exported as knobs 8 and 11. VCO2 supplies 3
waveforms: suboctave square, square, and saw, and their mix levels are
exported as knobs 10, 13, and 16.
I use the Nord's filter D for the Cat filter, since this one is
closest to my Cat's sound. By the way, my Cat is the original model
with discreet components, later Cats like the SRM have completely
different filters. The Nord's D filter is the one that comes closest
to the Cat's high end sizzle, though its frequency range doesn't go
nearly high enough; The Cat filter has a magical sound as it emerges
from super-audible frequencies that no digital filter that I have
heard can match. Also the D filter has the closest behaviour in that
magical hi-Q zone near oscillation, but again a digital filter and
midi controller really can't do justice to an analogue filter and a
pot. Finally, at hi-Q, the D filter thins the bass frequencies the
least and again, this behaviour is closer to the Cat sound than the
other Nord filters. There is a bummer to filter D, though; Its
modulator input is control rate rather than audio rate, which cruds up
the filter if you attempt to FM it, and filter FM is yet another key
to the Cat sound. I don't know what to do about this yet. Note that
I left an E filter in there as well; If you dial that one in with the
"+Filter picker" module you can get a decent Minimoog-like bass if you
dial the other parameters in right. The Cat switch for VCO1 audio out
is not exported but can be set by turning off knob 1 in the "+VCF
Input Mix" module; this turns off its audio without disabling its FM
abilities. Filter keyboard tracking is not exported but can be set in
the "VCF" module. Like the VCOs, there are two switches in the filter
section for mod routings, and they are not exported. "+Filter Mod1"
selects between LFO sine(default), LFO square, and S&H to modulate the
filter cutoff frequency, and "+Filter Mod2" selects between ADSR, AR,
and VCO1 (default) to do the same. On the Cat, you can get some great
nasty but still musically useful sounds by FM'ing the filter
frequency. A good modification for the Cat would be to use an
envelope to just FM the filter frequency on the attack. The filter
modulation amounts are set in the "Filter Mods" module, exported as
knobs 14 and 17. This module also includes a non-Cat mod routing: the
third mixer sets the amount that ADSR modulates the cutoff frequency
and this level is assigned to the main modulation controller (my mod
wheel). This allows me to use both FM and ADSR on the filter, which I
always wanted on the Cat. Finally, filter frequency and resonance (Q)
are exported on knobs15 and 18.
Still reading? Well we're almost done since the Cat is almost out of
knobs, and the Nord is completely out of knobs... The switch for
amplitude level is not exported, but it can be set in the "+AEnv
choose" module. The selections are ADSR, AR, and bypass, and by
default I use the ADSR envelope to determine the output amplitude. On
the Cat, you often use AR as the amp envelope when you are using ADSR
for effects like a filter sweep or transient pitch weirdness. Bypass
mode makes for a great drone. And for the final knob, I don't export
the Cat's noise control but you can set it with knob 3 in "+VCF Input
Mix". A worthwhile modification to the Cat would be to route the
noise through an envelope, but that's not how the Cat does it.
There's a method to the knob organization that makes it reasonably
easy to remember, especially if you're used to the real Cat. The
first vertical block of 6 knobs is loosely dedicated to VCO1, the
second block to VCO2, and the third block to the filter. The middle
row is dedicated to modulators, so by default knobs 2 and 5 modulate
VCO1 by the LFO and FM respectively, and knobs 8 and 11 do the same
thing for VCO2, and knobs 14 and 17 do the same thing for the filter.
The 4 knobs in the lower left select the output levels for the 4 VCO1
waveforms, and the upper right 3 knobs do the same thing for the 3
VCO2 waveforms. That leaves us with 3 knobs in the upper left which
are LFO speed, VCO1 tuning and VCO2 tuning, and 2 knobs in the lower
right which are filter cutoff and resonance, just where they are on
the Cat.
A lot of the Cat's nasty sound comes from all those rough edges in the
waveforms. You can smooth the Cat a good deal by changing the
suboscillators to sine waves. This will lose a bunch of high
frequencies from the sound; it can sound very good but it won't sound
like a Cat. Many of the weirder Cat sounds are had when you do
bi-directional FM, where VCO1 modulates VCO2 and vice-versa. You can
route a signal for FM into either an oscillators pitch input or its FM
input. These inputs have very different ranges and somewhat different
responses, so I let you pick whichever sound works better for you;
pick the ones you like in each of the two "+FM type"modules.
Figure 15.2. A Nord Modular emulation of the Octave Cat
(S. Streeper).
15.4 Arp Odyssey
by Dave Peck
The first synth I ever bought was an Arp Odyssey, in 1973. Although the
Minimoog was more popular, and although many feel the Mini had a better
sounding filter, I felt the Odyssey's versatility more than made up for any
sonic shortcomings. The Odyssey had pulse width modulation, oscillator hard
sync, a ring modulator, and a versatile sample/hold section, none of which
were included in the Minimoog. Also, the sample/hold section had a dedicated
input mixer, and the output of this mixer could be used for audio-range
modulation of the filter, also not possible on the Minimoog.
Figure 15.3. A Nord Modular emulation of the Arp Odyssey
(D. Peck).
OdyXWarmXPWM.pch
OdyXTrillXDrone.pch
OdyXSquareLead.pch
OdyXFilterXPatterns.pch
OdyXFilterFX.pch
15.5 Casio CZ Series
by Rob Hordijk
Just looked up Phase distortion in Curtis Roads's
excellent computer music tutorial - and there is this bit on the subject:
"Phase distortion synthesis is a term invented by the casio
corporation to describe a simple modulation technique
developed for several of its digital synthesizers. PD
synthesis uses a sine wave table lookup oscillator in which
the rate of scanning through the oscillator varies over the
cycle. The scanning interval speeds from 0 to PI and then
slows down from PI to 2PI. The overall frequency is constant,
according to the pitch of the note, but the output wave form
is no longer a sine....."
I think Casio used two (digitally simple and DSP cheap) tablescan
oscillators with both the same basic frequency and synced to each
other. Then when adding a bit of the output to the tableindex of the
other the phasedistortion occurs. This would mean a sineoscillator's
output to an FMB input on a SlaveFM oscillator, setting both to the
same frequency and syncing the SlaveFM to the sineoscillator. But I
remember vaguely from the frontpanel of a CZ-100 (it was called that
way, wasn't it? only knew one guy that had one of those Casio's in the
eighties) that it could be done with other waveforms than a sine as
well. Anyway, I like to do it as in the attached patch. Using the
multiwave slave gives a more harsh sync-like sound, more like the
CZ-100 as I remember. The sinevalue here is not added to the
'tableindex' as with the SlaveFM (FMB), but lineair to the
frequencyvalue (FMA), so not proper I suppose. But what's proper
anyway? Additionally the AM input on the slave oscillator is used as a
ringmodulator, suppressing small glitches that occur at the beginning
and end of a waveperiod on the NM's oscillators when using sync, due
to waveform interpolation (beginning) and FMA modulation (end).
Figure 15.4. A Nord Modular emulation of the Casio Phase-Distortion
synthesis technique (R. Hordijk).
15.6 Hallsey Mark 1 and 2
by Wout Blommers
In the late sixties and early seventies my musical friends and I were
really POOR!
Most of the gear was fake... The bassplayer could paint the Gibson
lettering so well,
he's now a typographical designer.
In those days we build our own Synths, yes sir!
For the Hallsey Mark 1 we used the oscillator of our Farfisa organ.
The Synth sound was made by distortion and it sounded like distortion...
Figure 15.5. A Nord Modular emulation of the famous Hallsey Mark 1
(W. Blommers).
15.7 Other Emulations
Emulating favorite synthesizers seems to be a popular pastime with
Nord Modular owners. Here is a list of some vintage synthesizer
emulation patches.
ARP Omni, by Andy "Buckshot"
ARP Omni, by Rob Hordijk
ARP Axxe, by Kevin Thomas
ARP Axxe, by Mr. Marko
ARP Solina, by Bergmuller
Roland Juno, by Roland Kuit
Roland Juno-6, by Bergmuller
Roland Juno-2, by Kendall Jackman
Korg Electribe EA-1, by Douglas R. Kraul
Korg Mono/Poly, by Rob Hordijk
Korg MS-20, by Wan Kemper
Korg MS-2000, by Rob Hordijk
Korg Polysix, by Bergmuller
Korg Poly61, by Keyboardslover
Korg Poly800, by Klaus Ludwig/Peter Gorges
Moog Minimoog, by Bill Barton (there are a number of
other patches available based on this template)
Moog Rogue, by James Maier
Roland Jupiter-8, by Bergmuller
Roland JX-8P, by Roland Kuit
Roland SH-101, by Rob Hordijk
Roland SH-101, by Bergmuller
Sequential Circuits Prophet-600, by Bergmuller
Sequential Circuits Sixtrak, by Rikard Latvala
SID, by Andreas Wickman
SID, by Rob Hordijk
Yamaha CS-40M, by Bergmuller
[Prev (Spatialization)]
[Table of Contents]
(c) James Clark 2003