For the last twenty years or so lots of people have asked me questions about how to make fat sounds on synthesizers. Although the rise of the digital synthesizers in the late eighties had severely reduced the amount of these questions, because they had the fat quality sampled or used edicated inbuilt effect processors, with the Nord Modular and it's little but gentle offspring the Nord Micro Modular, the subject seems to have gained interest again. So here it is.

First the meaning of fat in a synthesizer context. The word refers to a quality of a sound generated by a synthesizer. It is a very subjective quality that during the last twenty-five years changed meaning several times. Originally the word was used for an impressive sound. Actually the word meant 'great', 'powerful', 'big', etc. This term was used for other things in life as well, it was not specifically something that had to do with sound. But in the eighties it became a regular term to use for a synthesizer that was able to generate an impressive sound. It was specifically meant to qualify a synthesizer and not a typical sound. In the Dutch language fat is translated as 'vet' and a synthesizer that sounded impressive was a 'vette bak', meaning a 'fat box'. But of course what makes a sound or a machine fat and especially how to create a fat sound on a Nord Modular. For those of you who are new to the field I will start with some background information important to understand this discourse on fatness.

Sound consists of the following three basic parameters:

• pitch or frequency
• loudness or volume
• timbre or 'sound'

The idea is to have one or more boxes that have the ability to independently control these three parameters in time. To be able to control them in time is important in making music, e.g. a melody generally consists of several notes played sequentially, so we must be able to control the pitches of the notes on a time scale. So the basic synthesizer consists of at least three basic modules, a module that generates a pitched wave form, a module that modifies the timbre of this wave form and a module that modified the loudness of the sound, giving it a 'loudness contour'. The module that generates the pitched wave form is called an oscillator, a module that is able to change the timbre is referred to as a 'timbre modifier', and is generally constructed around some sort of filter, and the module that gives a loudness contour to the sound is actually a combination of two separate circuits, a controllable amplifier and a contour generator. Now it is time to mention Bob Moog, as he was the one that had the genius idea of standardizing the way to control these modules using voltages to control the parameters pitch, timbre and volume. Every module had a control input and on this input you had to apply a voltage. The range of these voltages was very well chosen, using 1 Volt as the basic measure. E.g. the pitch of an oscillator was raised exactly one octave if the voltage on the input was raised by exactly 1 Volt. On a timbre modifying filter the so called cutoff frequency was also raised by exactly one octave when the control input voltage was raised 1 Volt. And when raising the control voltage for a controllable amplifier by 1 Volt the volume was raised by 6 dB. Considering the fact that the human hearing range spans more or less 10 octaves, having control voltages between 1 and 10 Volt could span the whole human hearing range, very convenient indeed? So these modules became known as VCO's, VCF's, VCA's and VCADSR's, meaning Voltage Controlled Oscillator, Voltage Controlled Filter, Voltage Controlled Amplifier and Voltage Controlled Attack-Decay-Sustain-Release generator. The nice thing about these modules is that the signals they generated were technically speaking alternating voltage levels spanning a range of 10 Volt, so they could be used as control voltages for other modules. Doing this is called modulating eighties you connected modules by cabling the output of one module to the input of another, and so 'modulating' the other module. Bob used the terms 'modifier' for modules that modify the timbre of an already perceivable sound and 'contour' instead of ADSR, but most other manufacturers used VCO, VCF, VCA and ADSR, so we will also in the rest of this discourse. But the names Bob chose were very clear, so remember them. Now the bad news, the Nord Modular does NOT have VCO's, etc. A little bit silly this remark, but in the digital domain we do not use analog voltages anymore to control the modules. Instead some arbitrary unit is chosen, Clavia choose to have a control range of +/- 64 units, one unit equaling a semitone step, giving a total range of somewhat over 10 octaves, very convenient again. But if you want more, that's okay, you can exceed these limits, 'but at a certain level clipping will occur'. So in the Nord we could call the modules e.g. UCA, Unit Controlled Oscillator, but to remain compatible with the legacy VCO, etc. it is.

Regrettably there is not a controllable parameter called 'fatness', so how the hell do I get those fat sounds. As might have occurred to you when reading the previous section, fatness depends on the modulation we apply. But this is not really straightforward, so read on. First a quick history lesson. As said before fatness refers to the impression that a sound makes. So lets look at what people find impressive in several music styles and we find the trick. Back in the late sixties pop bands began to incorporate classic orchestra sounds in their music. The sound of the classic orchestra is VERY impressive, standing in front of a fully populated orchestra playing an 'orchestral hit', was 'fat'. So we see that the first meaning of 'fat' was an orchestral sound. Such a sound is generated by sometimes more than a hundred instruments playing chords that span the whole human hearing range. As the instruments are not exactly tuned and played with vibrato the overall sound 'thickens'. So to make such a sound make a patch of a hundred or so oscillators and tune them all over the hearing range. Fat indeed, but not very practical. So the trick is how to use as little as possible oscillators. How we do it will be explained later, but it was possible on the big synthesizers walls, used by the symphonic rock bands of the early seventies. Legendary names like Rick Wakeman, Keith Emerson and Vangelis discovered these sounds and used them up to exhaustion. By this time an engineer at the Solina organ factory, whose name I have forgotten, invented a circuit based on a new kind of chip called a 'bucket brigade', that gave a distinct orchestral sound to an otherwise not orchestral sound. This circuit was incorporated in the Solina String Ensemble and it opened the ability to incorporate orchestral type sounds in the music of band without big budgets. Especially the ability to play chords and to be able to take it to a gig added to the popularity of the String Ensemble. But it did not have the versatile sound shaping abilities of the voltage controlled synthesizers. The first development in synthesizers in the end of the seventies was to add polyphony. This was only possible when incorporating the then newly developed microprocessor chips to control the flow of the control signals, and for the synthesizer manufacturers of that period, being analog electronic engineers, was not easy. It actually was the cause of the bankruptcy of some of them. But things developed and as we know the Japanese at that time were very good in taking somebody else's idea and change it into an affordable and reliable product. So inspite of Moog, ARP, EMS and Sequential Circuits it was the Japanese who hit it big time in the eighties, not by incorporating polyphonic multi-VCO designs but by using the 'bucket brigade' circuit, enabling them to make synthesizers with one VCO per voice but thickening the sound with an 'ensemble'-modifier. This developed into the sample player synthesizer where the orchestral quality was either already present in the original sample recording or it was added with effects processors, versatile digital circuits adding 'chorus' and 'ensemble' effects smothered in reverberation. At the end of the eighties and the beginning of the nineties when dance music was gaining popularity, the quality 'fatness' got a new meaning. As dance music relies heavily on synthetic sounds a revival of the early analog synthesizers was to be foreseen, but not until it became a real big market. So now every original synthesizer manufacturer produces some 'virtual' analog music box, but only a few relatively new manufacturers offer them as versatile as the originals. In my opinion only Clavia has done a really good job, offering a true modular synthesizer that is not physically dependent of a computer system to PLAY the instrument. (Sorry Creamware, but I don't want to carry computer towers and monitors to a gig anymore) But dance music does rely on other things than orchestral sounds.

Now it is 'loudness' that is everyone's hype. We want BOOMS, and I mean loud BOOMS. Impressive BOOMS. Hey, there it is again 'impressive'. This loudness thing is a very interesting subject, what bands basically want is to stand out against other bands. This is very important as melodically dance music is not very refined. A single simple melodic theme with a booming rhythm is all it takes. But both the rhythm and theme have to be catchy and loud. Loudness has to do with the contrasts in the volume dynamics of the sound and especially how the 'body' of a sound stands out against the attack phase of a sound. This has a lot to do with how we perceive sounds and the single elements that make up a sound. Making fat sounds in the modern meaning can be compared to cooking. We have a recipe with several ingredients, but several people cooking the same recipe can come up with distinctly different tasting meals. So you have to augment your skills as a good 'sound' cook to come up with fat sounds. As the orchestral sounds are actually quit easy to make, modern fat sounds need a lot more attention to detail.

So now we know what to do. We need to make a patch where several modules are going to modulate each other in a way that we end up with an impressive sound, either loud, orchestral or both. For this we need to know the oscillators very well, so we start to look at the Nord Modular oscillators first. The most versatile oscillators are Oscillator A and SlaveOscillator A. First Oscillator A. We have the ability to modulate the frequency in a logarithmic fashion on the pitch inputs. The pitch can also be modulated in a linear fashion on input FMA. We can sync the oscillators wave forms to another oscillator and when choosing for a square wave form we can modulate the pulsewidth of the wave form. SlaveOscillator A has the same possibilities except for the pulsewidth modulation but as a very interesting extra it has a controllable amplifier built in, this is the AM input. This input is actually a so-called fourquadrant multiplier, better known as a ringmodulator. When connecting such an oscillator to an output module and listening to the sound, we hear a very static sound. The sine wave is the dullest of all, but tune is to a low bass sound and you will hear that it does have a distinct loud quality in the lowest registers. More than the other wave forms. Try this out and hear the difference. In the higher registers it is especially the saw and the square that sound louder. Here is already the first clue. Add some sine wavelike wave form as a suboscillator to a sound e.g. SlaveOscillator E, give this oscillator its own ADSR with a somewhat slow attack around 35 ms but always slower than the other oscillators, a full sustain and a somewhat longer release than the other oscillators and the sound starts to gain loudness. (Watch your speakers!)

The square wave has a distinct hollow quality to it, when turning the pulsewidth knob we hear the sound change into another timbre. This is the first way to 'liven up' the sound, when connecting a triangle LFO to the pulsewidth modulating input and opening the input knob about a third we hear that the timbre slowly changes in a way that once was called 'fat'. It gives a little bit an orchestral quality to the sound, but doesn't sound like the Royal yet. One of the disadvantages of this type of modulating is that is a bit to regular and you hear the LFO frequency. Mixing two detuned LFO's or using a random generator can make things better. But what I mostly do is to use a AttackDecay generator with a fast attack and a slow decay of 1 to 15 seconds. This does two things. First it makes the pulsewidth modulating very smooth but still very distinctly livening up the sound. Second if the pulsewidth is small immediately after the attack it sounds thinner, more high as the fundamental of the sound is less present, but when the pulsewidth starts reaching the square wave the fundamental gains volume and so the body of the sound seems louder. When using this technique use a separate AR for the pulsewidth and an ADSR for the output of the module. Set the Attack to 4 ms, full sustain and the decay to 170 to 400 ms. It is interesting to look at a very famous sound from the seventies and the eighties, namely the Vangelis sound. Vangelis used this same technique but he was able to increase the effect by using a filter. Delete the sine wave oscillator and it's ADSR, add an AD, add a 24 dB filter (not the classical!), add a crossfade mixer and connect the blue! envelope output of the AD to the pulsewidth input of the oscillator, set the attack time to 0 ms, the decay time to 2 sec and the pulsewidth modulating input knob to 60. Now connect the output of the oscillator to the input of the filter and the output of the filter to the ADSR. Set the ADSR attack to 4,2 ms, the sustain to max and the decay to 170 ms. First open the filter completely so it doesn't have any effect. Play some notes and you hear the effect of the AD in the pulsewidth, experiment with settings between 400 ms and 15 sec. You hear what I mean, the sound is decidedly more vivid but modulation is smooth. With a decay of 15 sec you can hear that directly after the attack the fundamental is more weak that after a few seconds, emphasizing the immediate attack and making the body slightly fuller. It is subtle but in a mix it can subjectively make a difference. Now to the Vangelis part of the story. Connect the output of the oscillator to the input of the AD and the output of the AD to input1 of the crossfade mixer. Connect the blue output of the AD to input2 of the crossfade mixer and the output of the crossfade mixer to the upper modulation input of the filter. Open the filter at 880 Hz and open the modulating input knob to around 30. Set the resonance of the filter to around 20. Assign Knob 3 to the crossfade mixer balance knob. Now play the keyboard while turning Knob 3 from left to right. Interesting, hey? What actually happens in the filter is a bit difficult to explain without thorough knowledge of what filters technically do, but as the filter 1s opened and closed in the same tempo as the oscillators wave form changes only part of every single cycle of the wave form is fed through the filter. When the pulse wave form rises the filter is not only opened but the sharp 'transient' of the rising flank of the pulse excites the resonating character of the filter, when the pulse falls the filter is not only closed but the resonance is damped as well. The ear however perceives the both exited quality and the damped quality of the filter. This is a bit like what happens in an Aural Exciter. But strangely enough the ear does not perceive the phasing character of the changing pulsewidth anymore. When turning Knob 3 you hear that when you open the filter with only the envelope (knob to the right) the phasing is there but the sound seems to sound duller. When turning the knob left, and the filter gets the wave form instead of the envelope the sound get brighter and fuller and so gains more presence in a mix. Very good for basses. I really don't know if Vangelis did it this way, but it sounds the same as the fat synthesizer basses on the albums he produced in the end of the seventies and the early eighties.

When experimenting further with these patches you will discover that it is all very subtle. Actually in the british magazine The Mix, issue 61 May 1999, page 110, there is a very well written article about loudness.

Now on the orchestral type of fatness. This is mostly done by using more oscillators. E.g. mix the sound of two saw wave form oscillators in an equal amount. When exactly in tune the sound is static. When detuning just a minimum amount the sound starts to phase. This is because as the wave forms drift apart all the overtones start to mix, doubling in energy when in phase and canceling each other out when exactly out of phase. With only a minimum detune you can hear this, but when detuning a little bit more the doubling and canceling out get to complex for our ear to perceive and the sound get a distinct 'orchestral' quality, like several instruments playing together. When detuning even more the start to hear two not very well tuned pitches. So between exactly in tune en decidedly out of tune there are two areas of interest, one area where we have a phaserlike sound and an area where we have a 'ensemble' like sound. When adding another oscillator and detuning the three in the 'ensemble' area, we get a distinct orchestral sound. Do this with three square wave forms that have three differently tuned LFO's on their pulsewidth inputs and the sound gets very rich indeed. One problem is that the higher a note you play on the keyboard the faster the energy doubling and canceling of the overtones occur, resulting in a slow phasing sound in the bass register and a very nervous sound in the very high registers. This is because the detuning is logarithmic. Original analog synthesizers had an option called 'high frequency tracking' that compensated this effect giving a more stable ensemble effect over the whole keyboard range. The Nord Modular however lacks such a control, but it can easily be emulated by making use of the FMA inputs of the oscillators. Do this in the following way. Set all oscillators to exactly the same frequency, add two constant modules, connect one module to the FMA input of the second oscillator, connect the other constant module to the third oscillators FMA input, open both FMA control knobs just a little, 1 is already enough. Now set constant module one to -32 and constant module two to +32 and play the whole keyboard. Now you hear that the speed of the phasing is constant over the whole keyboard range. But now you may find the bottom registers to busy and the high registers to slow. The original trick on the analog synthesizers was not to add a constant linear detune but to slightly linearly detune the oscillators with the keyboard voltage. Figure out for yourself how to do this on the Nord.

So these were the techniques that are mostly used to 'fatten up' a sound, either by giving it more body by boosting the fundamental in the sustain art of a sound, adding a sub octave in the sustain part preferably as sine like as possible using either a discrete suboscillator or some exotic property of some filter types, using pulsewidth modulating and/or using more slightly detuned oscillators giving a distinct 'ensemble' or 'unison' sound. Instead of using more oscillators you can of course use the Stereo chorus module that does the same as that legendary 'bucket brigades' circuit, only now without the noise.

But some more notes should be made before we delve into the subject of sync. If you want to make a sound that stands out by itself, without the backup of other instruments, use several oscillators tuned in registers all over the keyboard range. Using different wave forms can give some special subtle effects. E.g. when using a square wave form and a saw wave form and the square wave form is tuned one octave below the saw wave form the overtones interact in an interesting way. The square wave form misses the even overtones, so it only has harmonics 1, 3, 5, 7 etc. A saw wave form has all harmonics. This means that in this particular case there is no second harmonic in the square wave that can interact with the fundamental of the saw wave form. Detuning the saw wave form slightly means that both the fundamental of the square and the fundamental of the saw are stable in volume. This gives a stable body to the sound, only the higher, but not all of them have phasing. Adding yet another slightly detuned saw wave form oscillator gives a rich sound texture. Try how it sounds with some of the different filter types and especially with the phaser module. This type of sound stands like a rock because in a mix it actually covers more of the frequency spectrum than only the saw by itself.

Well, of course the bag of tricks is almost endless, we haven't mentioned high resonance filters combined with the overdrive yet. This is a very popular trick these days. Try it yourself.

Oscillator synchronization is subject not very well known by people what it actually does, so some explanation. Traditionally there were two types of sync soft sync and hard sync. First hard sync, as this is the type that is available on the Nord Modular. Imagine what a square wave form does with your speakers. On the rising flank of the wave form it very quickly pulls the cone of the speaker out of the box towards you until the cone is in the maximum position and hold it there, until the falling flank, where the cone is pulled into the box and is hold there until the next rising flank (this is if you have a very, very good box) So a square wave form makes that the cone has only two positions, in the box and out the box. There is however a very short time when the cone travels inside or outside, the shorter the time it takes your cone to travel that distance the better box you have. This very short period during the flank is called a transient, which means a very short pulse with an enormous power, theoretically of infinite duration and of infinite power. Why so much power? Because in this very short moment all the physical mass of the cone and the air pressure in the box and outside the box have to be moved in preferably no time. Both a saw wave form and a square wave form start their wave form cycle with a transient, a sudden change from maximum negative to maximum positive, the difference being that a square wave stays stable during the half of the cycle and than has another transient, but the saw 'slowly but steadily' goes negative again until reaching the maximum negative value where another transient starts the next cycle. To understand what sync does is to understand that a sound wave form of an oscillator consists of a series of cycles. If by some means you can restart the cycles before they reach their natural end you do something very drastically to the sound. E.g. we have two analog saw wave form generators far out of tune. When connecting the output of one oscillator to the sync input of the other it is the energy that is in the transient, the rising flank that forces the other oscillator to restart its cycle, no matter where that oscillator has in its own cycle. This means that the second synced oscillator syncs to the pitch of the first oscillator, every time the first oscillator restarts its wave form, the second oscillator has also to restart its wave form, but as that wave form has a different pitch it gets 'broken' cycles. Although original pure analog oscillators actually do sync on the energy that is in the transient, e.g. sine waves did not have enough energy to be able to sync another oscillator, in the Nord Modular it is slightly different. The process of restarting the cycle is the same but the oscillators are synced on the moment that the first oscillators wave form crosses the zero line from a negative to a positive value. This means that we can use any wave form to sync an oscillator that has a sync input. So where do we use sync. We can use it when we modulate two wave forms together in some sort of patching where the phase shift between those oscillators influences the sound, but that influence is undesirable. In that case we can sync on the keyboard gate. If we do this we can be sure that both oscillators start their cycles at the same time. And one thing about the Modular's oscillators in contrast to traditional analog oscillators, when tuned to the same value they do not drift apart in frequency, enabling very sophisticated synthesis techniques like FM, that were not possible on the old dinosaurs as they already drifted when you looked at them.

When syncing two oscillators traditionally you do it like it is described in the Nord's Manual, but you can go way beyond. Traditionally the synced oscillator is tuned higher than the syncing oscillator giving a train of, lets say, one and a half cycles. Changing the pitch of the synced oscillator gives a screaming sound when all these cycles are torn in half and expressing their pain. That's why there existed something called soft sync, in that type of sync the wave form was not restarted but the direction of the was reversed, so if a wave form was gradually rising it started to gradually fall. As this is not present on the Modular I will not explain this further. On the old analog boxes not only was the moment of sync the transient but also the moment when a wave form started was the transient. This resulted for saw wave forms that you got a rather smooth sweep sounding like pulsewidth modulating only a little bit more aggressive. On the Modular however the cycles start on the zero line and the sync moment is the zero crossing. As a saw wave form on the Modular travels slowly up and the transient goes from positive to negative, the syncing process when using saw wave forms is not on the transient but halfway on the slow way up. This results in a far more harsh sync sound, the sweep is far from smooth. There is a little trick that can soften up things a bit. Use a SlaveOscillator A as the syncing oscillator and add a constant module, connect the output of the constant module to the AM input of the SlaveOscillator A and set the constant to a value of -64. Now another oscillator will sync on the transient. This you can use if you mix the output of the syncing and the synced oscillator together. Also use SlaveOscillator A as the synced oscillator and connect the output of the syncing oscillator to the AM input of the synced oscillator, this suppresses the transients a bit. Then connect the output of the syncing oscillator to the FMA input of the synced oscillator and the fun really starts. If you would have owned a Korg MonoPoly you would have known all of the above. As the effect of synchronizing oscillators gives a very extreme alteration in the sound, impressive to most of us, some people call this type of sounds 'fat'. So we are back at the beginning again. Time to stop this discourse and wish you all luck, and especially a very good time making music?