Although not entirely new, Additive synthesis has been one of the few synthesis forms that never quite got off the planning boards. Circa 1989 commercial hardware companies like Kurzweil and Kawai were trying to develop synths that were additive.
Kawai is often noted for it’s dedication to this different form of synthesis and in the early 90s broke through with the K5. It was a Frankenstein cobble of a keyboard, synth engine and lots of push buttons with a tiny lcd interface. To program it you hunched over this keyboard and pressed your way through page after page of menu options with unintuitive terms like partials and frequency range. The output stage of the synth engine was noisy. The sounds the K5 was capable of making were klangerous and bell tone-like. With some dedication and lots of hours attempting to program with buttons and a 2 line by 40 character interface there wasn’t much to get out of the poor K5.
Yet it’s sounds were at times impressive. It excelled at tine style electric pianos, organs, bells and noise style patches that made you cat hide under the bed. At once ugly and beautiful the K5 was doomed in a market that demanded samples of real pianos and basses, drums and guitars, etc. The K5 also had the misfortune of sounding a lot like a hopped up DX 7. And with FM (frequency modulation) synthesis finally on the way out after a incredibly successful run, the music world went back to discussing additive synthesis until in 1996 when Kawai once again developed a full line of additive synthesizers called the K5000 series. These included a hybrid workstation keyboard named K5000W with lots of samples of “real” instruments and synthesized sounds also including additive synthesis minus the K5000s set of knobs and arpeggiator..and sporting a general midi sound set and sequencer. Featuring combinations of additive synthesis mixed with samples,. The true standard bearers, the K5000s and K5000r, a similar rack version (keyboard and rack versions, respectively) which featured additive synthesis aided by small samples of attack partials that were helpful in emulative style timbres. This time Kawai looked at the current market and included many programmable knobs, a few buttons and a larger lcd screen, making edititing easier and making great possibilities for live performances. More importantly this additive synth was at once a hybrid of sample and/or additive synthesis and a wealth of controls available to the user. There were filters, resonance, arpeggiators, and a large effects section. Everything the synthesist could want. The K5000s was the first true additive synthesizer that had enough horsepower to create sounds that weren’t weedy sounding bells, organs and the like, such as it’s cousin the K5 – on the contrary, it could create some very nice sounds. It also had a spectacular synth keyboard action and was a terrific controller. The K5000 was also built like a tank, all metal with a brushed aluminum faceplate and a tiny end cap of heavy duty, industrial strength plastic.
Part of the reason the K5000 series worked had to do with the more powerful additive engine. A robust number of 64 partials per oscillator and up to six oscillators were available. You could dedicate one oscillator to a sample, usually in the attack stage and leave the rest of the sound shaping to the additive engine. Better still, the additive engine created ethereal timbres, big sounding analog style basses and leads, evolving pads and soundscapes, all thanks to the help of a filter system that worked like traditional subtractive synthesizers (analog synths are subtractive, we will follow this track soon). With the additional effects onboard additive had made it’s mark. It was great sounding, at once clear and still with quite a huge sound. More importantly the K5000s, sounded much like the competition at the time, particularly virtual analog., a subset of another exciting development in synthesis.
Physical modeling was based on evaluation of what an instrument was made of. Typically this meant how it was played. Was it blown, struck or plucked? Once that was established the body of the instrument was modeled with the material the body was made of calculated. In the marketplace a fully functional physical modeling synth called the Z1 by Korg attempted the first serious stab at making a multitimbral instrument that could be a lot of things that couldn’t exist in the real world. But eventually the market was driven by the fast developing techno market which craved analog in all it’s glory. The Z1 could do lots of things including analog but lacked that people wanted in order to “feel” like analog based synths.
Why the brief mention of physical modeling? Because a specialized form called Virtual Analog became the predominant form of synthesis. Virtual drum machines, virtual synthesizers, some like the nord Lead and Novation Supernova being incredibly close in sound and more importantly, in feel to the analog synths of old days. Virtual analog was and still is very popular. There was another side to the analog style sound, groove boxes. These instruments featured sampled analog synthesizers and married them to a pattern based hardware sequencer. In particular Roland created a number of these units using the name of their techno essentials, the original analog TB303 bass sequencer and synth and the TR series of drum machines, the TR808 and TR909 . By cobbling together a sample playback system that married the two, calling it the MC303 Roland created a style of music creation that played like analog but was sample based..Roland called the MC303 a groovebox. Most techno artists called it rather unpleasant names as it sounded anemic compared to the real thing, however, it had a bigger brother, the MC505 which used a better synthesis engine and vastly better samples. Along with Yamaha’s RM1x the world of music was ‘groove approved’.
Which brings us back to our friend, the K5000s. Kawai did it’s homework this time. It’s synth could do techno. In fact, to this day many techno tracks feature the K5000. It could be phat, it could be clean. It could be a whole lot of things but one thing was certain, it was far more difficult to learn. Additive synthesis as developed by Kawai was just too complex. The bank of knobs were inviting as was the K5000’s advanced polyphonic arpeggiator. But making sound required a new way of thinking, one that while being advanced and ahead of it’s time was also doomed because of it’s difficulty of programming and it’s lack of presets. It took a memory upgrade to get more than 120 patches, in fact, patch data varied by preset.
Virtual analog was king and the Groovebox was a close second. Musicians didn’t want to learn about formants or look at a bunch of bars on a plane that looked more like a sales chart than a waveform. It wasn’t much better with lounge, jazz or rock players. A new synthesis system? Forget it, we’ll stick to our Korg Trinitys and Roland XP80s or MC505, thank you. And Kawai once again stopped development and the K5000 series crashed and burned.
And that was that as they say. Nothing was to be heard in the world of hardware that was additive synthesis based.
Now that we’ve had a little history of commercial sales of additive synths, let’s discuss what they do.
My first notice of additive synthesis was in the mid 1980s. The idea seemed exotic and futuristic. If you take a look at sound in additive synthesis the smallest part is a simple sine wave. In terms of synthesis a sine wave is the most basic of sounds, it is bland, boring, the white bread of sounds, add up enough of them and you can get any kind of sound imaginable.
Sine waves have minimal distinction and aren’t likely to induce thrills if you put a couple together. In analog synthesis you could make sounds using this style of waveform but without bringing other influencing factors such as amplitude envelopes, filters and pitch differences the sound will not be terribly exciting. Sine waves did have a purpose but for most sounds it wasn’t until you used a low frequency oscillator (LFOs for short) with sine waves that sounds became interesting. This is because the sine wave works well for modulating pitch as an example. In fact, many analog synthesizers depended on sine waves in LFOs to make modulation interesting and musical sounding.
The idea behind additive synthesis is sometimes non intuitive to the traditional analog synthesist’s mind. Instead of taking a waveform and modifying it, here was a style of synthesis that said, ‘we start with a sine wave and we add more sine waves until we reach the kind of timbre we want, oh and by the way, when we add them we will change the pitches, places where they get played, and oh yes, it will take a few hundred just to make a sound with any type of character’. And, if that’s not enough, we want you to think about sound in terms of wave after wave smashing into one another to eventually make a sound.
The best way to explain this effect would be looking at a sine wave. It has that familiar up and down, semi circular look. But when you add another sine wave at a different pitch the new wave changes substantially and stops looking like a sine wave. That is part of the additive process, it adds complexity and as complexity continues sound changes as well. If you look at the combination of waveforms in an audio editor you will see the two joined sine waves have become more complex.
Not being a technical person, and this article not meant to be a white paper, more of a ‘let’s sit on the grass and talk about it’ kind of discussion a lot of fairly technical information will not be here. There are important issues such as Fast Fourier Transmission (FFT) that are important in the mechanics of additive synthesis but at a fairly technical level that would require a great deal of discussion. Links will be provided for those who want to read further on the more technical aspects of additive synthesis. To describe FFT in very general terms, it is the ability to analyze sound over time and how to track it, or in the case of additive synthesis, extract the information about the sine waves that make up a sound or waveform.
So here’s where we are.
In theory you need a lot of sine waves to make an additive sound. These sounds can be anywhere from a few hundred to tens of thousands depending on what type of sound you are trying to create. For example, a flute has a specific attack section and afterward sustains at a fairly constant rate. That’s why almost any kind of synthesis can make flute like sounds, they’re easy. But creating a realistic voice requires substantial fluctuations in waveforms and changes a lot over time. Complex sounds with noisier components like woodwind instruments or vocals may require more sine waves than simple harmonic sounds like pianos.
There is a catch here which is sample analysis. While a system that uses sine waves strictly to create sound might take hundreds to thousands of sine waves it is possible to get around this by using sampled sounds and converting them to additive form. This way sounds are now complex additive structures and can be edited in such a way as to make each of the 32 waves considerably more complex and capable of a high degree of sonic information. As we will see later on, there are several ways to implement additive synthesis.
Until recently there weren’t many additive synthesizers because the power of computer processors couldn’t generate the amount of information needed to form a robust synthesizer engine. In 1993 that barrier was finally broken and no less than five VST instrument developers created additive synthesizers
3. Additive Finally Makes a Splash
We are reviewing four additive synthesizers. The remarkable aspect of each synth is how it took additive synthesis and developed a engine that not only enables creation of unique timbres but also familiar timbres one would think were made with a analog synthesizer.
There are new designs for interfaces that allow for intuitive use of this form of synthesis. Some include loading pictures or drawing oscillators. The pictures get interpreted by the synth engine and converted into additive waveforms. There are synths that allow for graphics to be uploaded but without editing in the graphic’s format. We will see new ways of creating sound, in particular most of these synths use bar graphs to represent waveforms. Some have only 32 waves, small considering the typical way of working in the additive format. But we will also see that all additive synths have the ability to import waveforms. Some synths have time limits and require a small amount of preparation while others happily take whatever is given them, discarding the unused information.
Sample use is important in one synth we will look at. In another waveform analysis is available while it’s core engine is capable of timbres impossible to create using other instruments. In it’s case the sample analysis is a nice addition to a strong feature set. Lastly, we will see a synthesizer that uses samples as it’s initial content. It analyzes the sample into additive waves and allows for simple additive synthesis as well, often combining them. The unique area of this synthesizer is it’s ability to literally have up to four analyzed waveforms or combinations of user developed additive creations and morphing the additive waveforms into actual new instruments that can morph while being interpreted. These are not your father’s synthesizer.
In addition on several synths there are multiple filters, LFOs, modulation matrixes, formants (which are capable of creating speech and voice type sounds). Formants are filter like elements that can be changed over time. In fact, all the reviewed synthesizers are based on timbre over time. There is even one synth that offers eight step sequencers for one timbre.
The above paragraphs are a little odd since you are not naming synthesizers directly but you are talking about features on them. It’s kind of confusing and may be better to save this kind of info for your reviews. Maybe it would be better to say something like:
Some are based around importing samples, some are based around visually drawing waveforms. Many even allow you to import images. All vary in the number and type of partials that can be used and whether those partials can be freely tuned or whether they are limited to only doing harmonic sounds. The synths also vary in their modulation options. Some offer lfo’s modulation matrixes, and formant filters. One even offers a step built in step sequencer. These design decisions have a large impact on the type of sounds that can be produced and the way you work with the synths. These decisions also make each one unique and capable of different kinds of sounds.
You may want to think about removing some of the following paragraphs too.. This is the kind of stuff that would be better to have in your review when you can talk about specific synthesizers and what they have on them, rather then throwing it all out now without being able to associate it with specific synths. The way it is, it’s kind of difficult to visualize what you are talking about since you are talking about all of them at once.
These new breed of synthesizers at times use subtractive synthesis to effect the spectral qualities of the additive oscillators. They feature from two to four oscillators and some synths pair spectral data and in one case noise as part of an additive oscillator. Effects are also included as are multi-segment envelopes that enable pitch, amplitude, modulation and filters to be automated. A new form of editing is found on some of the synths we review. Once envelopes are set up they are visible on one part of the synthesizer and are capable of being edited or manipulated with midi data sent. Think of a box with an x and y axis, if this makes no sense think of a box with a line going up and a line in the middle of the box going across it. Being able to move multiple points creates an environment that enables fast changes to many elements of the sound.
Still don’t get it? Think of a drawbar organ. As you play it the frequencies you alter make the sound change. This is not too different from how additive software interfaces work. Again, using the organ. Consider what it would look like in that x/y box. Each drawbar would be a point with a line. Each point when moved would change the organ’s timbre. But in this case it would also change how the timbre related with other timbres over time. Give it a few seconds to visualize. The possibilities are extraordinary.
And while this may sound complex to the point of your head being ready to burst, as well as those that can’t wait to dig into this array of capabilities, the casual user can work with these synthesizers as they all have graphical interfaces that create substantial changes to timbre and alteration of how the sound changes over time. In reality, a determined synthesist that has no background in additive synthesis or how it works, which this article hopefully helped readers, almost no understanding of synthesis is required, just a inquisitive mind.
Then again, there are so many features and so much diversity that one could easily purchase each synth for it’s specific features. Also, because each synth reviewed has it’s own way of using the additive waveform at the interface level understanding how each synth works will help in the ability of creating sounds that simply can’t be created with other forms of synthesis. This is part of the allure of additive. In theory it is the only kind of synthesizer you need although with due respect to all synthesizers reviewed at musicFAQ none are at that point yet. Neither are analog, frequency modulation (FM) synthesis, which we did not cover here but in several ways is similar and makes an appearance on some of the reviewed synths.
The idea of combining parts of traditional synthesis is new. There are many synthesizers that use additive in part but do not fully develop additive as it was traditionally envisioned. The synthesizers here come as close as anything yet developed and available. And best of all none of these synthesizers are currently over the $200 street price, or online price. Prices start at $70, at the VST instrument pricing range this would fit nicely into a affordable solution. So do the others and as you go up the prices you get more. Sometimes in features and in sample libraries leased from high end sample producers.
Unless you are unwilling to try something different or think the only sound worth hearing is made from an analog or in the world of VST instruments, virtual analog, you will be impressed with the depth of sounds and in some cases the astonishingly high quality of sound, sound that rivals or exceeds anything available in hardware.
On to the reviews…..
(editor’s note, unfortunately we were unable to obtain a copy of discoDSP, Vertigo in time for this review, it is a promising synth although it was impossible to review properly using it’s demo)
ConcreteFX Adder $70.00
WhiteNoise Additive Synthesizer. $99.99
VirSyn CUBE $199.99
Camel Audio Cameleon CA5000 $199.99