Additive vs. Subtractve

[Husker]

Can someone be kind enough to explain the difference between an additive vs subtractive synthesizer? I see these terms on other physical and virtual instruments.

 

I am just learning about all of this.

[brenner13]

It's all subjective and dependent on proximity of perspective.

 

(Just being snarky 'cause it really all goes way over my head.) Anxious to read some replies of the more learned folks among us...

 

[mate stubb]

Subtractive is the most common and exists in most classic synths. It uses waveforms with rich harmonic content like sawtooth and pulse waveforms, and SUBTRACTS harmonics to create other tones using a filter. This is the mode that Moog, Arp, etc. and other synths like them use.

 

Additive uses lots and lots of sine waves to assemble the fundamental and overtones of the sound. It requires lots of hardware (or software) for control, as each and every harmonic usually has its own envelope (higher harmonics decay faster than slower in real world instruments.) Very few synthesizers outside of computers have ever used additive. I think there used to be a Korg that did this.

 

There is also the so called west coast model, used by Buchla and others where the oscillator starts with a sine wave and creates harmonics by being modulated at audio rates by another oscillator. The resulting waves can be filtered still if you want.

[timwat]

Subtractive starts with an oscillator (or oscillators) producing a waveform - often with harmonic content above the fundamental, and uses a filter (or filters) to subtract harmonic content to taste or intent. Early analog synthesizers like the Minimoog are examples.

 

Additive uses a series of individual oscillators to add harmonic content above the fundamental - a simple example is the series of 9 drawbars on a Hammond organ.

 

Very simplified description, and I'm sure a quick Google of the keywords will provide a far more detailed explanation, and examples.

 

Tim

 

Edit: Moe's post above is better than mine.

 

 

[Dave Bryce]

The real basic version?

 

Subtractive synthesis - the one most folks call analog - is when you take a raw (usually) simple waveform and modify it by running it through some kind of filter that subtracts frequencies/harmonics.

 

Additive is when you take multiple simple audible (usually sine) waveforms and combine them together to make different more complex waveforms. There's (usually) no filter involved - the only way to change the harmonic content is to change the waveform itself. How that's done takes longer to explain, and is less intuitive to the average synthesist than subtractive synthesis.

 

There's a version of additive called FM (frequency modulation) where combinations of inaudible sine waves (modulators) are used to modify audible sine waveforms (carriers) to make some really cool very complex waveforms - the Yamaha DX7 was such an instrument - but that one will really make the average person's head hurt.

 

How's that?

 

dB

 

 

 

 

[mate stubb]

So on a practical note, every synthesist should at a minimum learn to be comfortable with subtractive synthesis. This will give you a foundation to master many instruments.

 

Even ROMplers like a Motif can be considered a subtractive synth, albeit with samples instead of oscillators.

[sunshadow]

it all starts with sine waves....subtractive is when you start with a complex (many sinewave) waveform and then shape it with filters, subtracting "harmonic content"...harmonic content is a bunch of sinewaves all mixed to gether

 

additive, you start with a single sine wave, then you create additional harmonic content (a bunch of sinewaves) using frequency modulation...when you affect the phase and harmonic ratios of multiple sine waves it multuplies the content and makes new sinewaves...

[mate stubb]

I'm not sure that FM synthesis is a type of additive synthesis. To me additive always meant that you are specifying each harmonic yourself, rather than accepting the sidebands you get back from frequency modulation or ring modulation.

 

But I'll leave that argument for the academics.

[timwat]

So on a practical note, every synthesist should at a minimum learn to be comfortable with subtractive synthesis. This will give you a foundation to master many instruments.

 

Even ROMplers like a Motif can be considered a subtractive synth, albeit with samples instead of oscillators.

 

I couldn't agree more with this. Few of us will ever be master sound designers. But even a passing familiarity with subtractive building blocks (and what they do to the sound) will elevate you from being a mere preset jockey.

 

In today's gig world, the ability to come up with an approximate sound to fit the context seems to have become a scarce resource, and thus elevates what I always thought was a mundane job requirement into some sort of wizardry status.

[Dave Bryce]

I'm not sure that FM synthesis is a type of additive synthesis. To me additive always meant that you are specifying each harmonic yourself, rather than accepting the sidebands you get back from frequency modulation or ring modulation.

 

But I'll leave that argument for the academics.

Don't make me call Hal. He'll make everybody's head hurt.

 

Ever play with his K150?

 

http://www.vintagesynth.com/kurzweil/k150synth.jpg

 

dB

[mate stubb]

Was not familiar with the K150. Add in the Kawai K5 and K5000, and additive synths may not be as rare as I once thought.

 

The one thing they all seem to have in common is that they are unsuccessful in the market!

[Dave Bryce]

Was not familiar with the K150. Add in the Kawai K5 and K5000, and additive synths may not be as rare as I once thought.

 

The one thing they all seem to have in common is that they are unsuccessful in the market!

With the K150, it was more that it was somewhat challenging to get anything resembling a useful musical sound out of it.

 

dB

[Doerfler]

Was not familiar with the K150. Add in the Kawai K5 and K5000, and additive synths may not be as rare as I once thought.

 

The one thing they all seem to have in common is that they are unsuccessful in the market!

 

I remember when the Kawai K 5000 came out 20 years ago. It sounded amazing. No one bought one. :idk

[Dave Bryce]

I remember when the Kawai K 5000 came out 20 years ago. It sounded amazing. No one bought one. :idk

No one buys the weird cool synths. Prophet VS. Wavestation. Z1. I could go on....

 

dB

[ksoper]

The Emax SE (Synthesis Enhanced) and later the Emax II gave you the ability to create additive timbres. From the SE upgrade manual:

 

"Spectrum Interpolation Synthesis: A complete digital additive synthesizer is the basis of this update. At Iast...additive synthesis that is user friendly! Never before have the dynamic timbres of digital synthesis been so easy to create. Simply define up to 24 Spectrums (or select them from the on board palette of 100), place them at any or all of the Time Slice locations, and enter the Synthesize! command. Sounds may be any length (up to the maximum memory) and can be customized by using the sophisticated compliment of edit parameters that are included. After a sound has been created it may be spliced, combined, and processed in the same manner as a sound that has been sampled into the Emax. The Spectrum Synthesizer features Spectrum drawing/editing, Tuning of Partial Ratios, Definable Amplitude and Pitch Contours, Time Slice creation and editing, and above all - Great Sound"

 

What they neglected to say was that after you entered the "Synthesize" command, it could be sometime the next day before the sound was ready to play. And you could spend all that time and not necessarily come up with something usable. I can't say I used the feature all that much. Good thing they included a bunch of ready-made timbres.

[Synthoid]

Is there synthesis that involves multiplication or division? :idk

[Markyboard]

Don't make me call Hal. He'll make everybody's head hurt.

dB

 

I'll see your Hal; and raise you 1 Theo.

 

No one buys the weird cool synths. Prophet VS. Wavestation. Z1. I could go on....

 

dB

 

Certainly some weird (but maybe not so cool) people do. There seems to be plenty of love for these synths mentioned on GS and probably elsewhere. But far far from mass market appeal.

 

Is there synthesis that involves multiplication or division? :idk

 

They all do as in...

Cost/# paychecks = GAS Satisfied (temporaily)

 

[The Real MC]

I'm not sure that FM synthesis is a type of additive synthesis. To me additive always meant that you are specifying each harmonic yourself, rather than accepting the sidebands you get back from frequency modulation or ring modulation.

 

But I'll leave that argument for the academics.

Don't make me call Hal. He'll make everybody's head hurt.

 

Ever play with his K150?

 

http://www.vintagesynth.com/kurzweil/k150synth.jpg

 

dB

 

I played with one in a store. A panel full of buttons is a poor interface for additive synthesis. There was a Mac application but only for the antique model of the day, and there is zero software editors available for this thing. The only good factory sound was the piano but the K250 sampled piano is far better. The rest of the sounds were extremely sterile and bland. The 150 didn't sell me onto additive synthesis.

 

And FM is not additive, it is a synthesis that is dependent on modulation.

[DaveMcM]

Just for fun:

Anybody with a Hammond or other drawbar based organ can experience (albeit to an extremely minimum degree) additive synthesis. Bypass all rotary and C/V effects and use this setting- 008080800. You now have an approximation of a clarinet.

 

As for FM being or not being a form of additive, my 5 cents says it is, sort of. FM starts with basic waveforms and adds them together at varying frequencies and amplitudes which creates additional overtones. And then there is always Algorithm #32.

 

In reference to the point made by The Real MC, the same can be said about straight ahead additive machines in that they also rely on modulation. In upper-end additive, each harmonic will have its amplitude and even its pitch modulated by EG's. Just sayin'. Although I do get your point.

[David Emm]

Can someone be kind enough to explain the difference between an additive vs subtractive synthesizer? I see these terms on other physical and virtual instruments.

I am just learning about all of this.

 

Analog is pretty easy to grasp, but additive takes longer to embrace, because its generally anything but grab-&-go.

 

What I call "Academic" additive is a unique form of harmonic exploration with esoteric goals generally grasped by only the in-crowd of that arena. Because of the vast number of partials required to get a more traditionally recognizable sound, that end of additive is usually the realm of odd soundscapes and ambient washes with an edge. Its less for casual listening and more for explorations in alternate tunings or studies in off-center harmonic clustering.

 

OTOH, "Player's" additive is very well presented by Native Instruments' RAZOR, where numerous useful *groupings* of harmonics give you a head start under an analog-like GUI. Several 'filter' models are on tap, which give you more ready access to certain desirable additive behaviors that would otherwise require untold hours of microscopic tweaking per partial. Its been a populist EDM favorite, but its a broad synth whose demos you should explore. It'll answer a lot of questions. You can achieve many glassy, additive-like pads and bells with the informed use of single-cycle waves, but its easy to contrast them and decide which suits the job best.

 

There are other additive soft-synths to be had, but essentially none to be found in hardware because of the processing power required. Kawai took an honest shot at it with the K5 and K5000, before there were affordable commercial CPUs that could crunch those numbers for real. Then personal computers finally acquired the power to go at it more seriously. I find additive amusingly named, because for every more obvious bell or cutting bass I might draw on, there are many more cases where a small bit of it under a more analog or sample-based patch lifts the whole mix. Analog can be learned by laying hands to it for a while; additive is bit more, um, labor-intensive. Listen to some YT examples of it and painting with it will seem less daunting.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

[timwat]

Another early commercial effort at Additive was the Synergy.

 

Here's some info on it.

 

As memory serves, it was originally derived from the massive Crumar GDS system (and originally required one to program from scratch). Early champions included Wendy Carlos (!) and Donald Fagen - I thought I remembered seeing an interview with Fagan back in the day in which he stated he hated synths because of their intonation, preferred mechanical instruments but also like the Synergy.

 

At least that's what I remember...

[rod76]

A rare early example of an additive synth but interesting nonetheless.

 

 

[video:youtube]

[Synthoid]

That was interesting. And quite fascinating pitch bending as well.

 

A better view of the RMI:

 

 

 

http://www.spheremusic.com/userimages/Img11160.jpg

 

[Husker]

Thank you all. It's amazing how little I know!

[Tom Williams]

The Kawai K3 wavetable synth had one "user-definable" waveform, which was additive.

 

The fully-additive Kawai K5 was an amazing synth, another of those things I loved but couldn't afford at the time. It allowed you to define either one oscillator per voice with 126 harmonics, or two oscillators with 63 each. You could choose groups of harmonics within an oscillator and modulate (at least the volume of) a group at a time.

 

[Doerfler]

A better view of the RMI:

 

and only a 4 octave keyboard. I truly can't believe anyone actually bought one of these.

[Joe Muscara]

A better view of the RMI:

 

and only a 4 octave keyboard. I truly can't believe anyone actually bought one of these.

[mauriziodececco]

Back in the good old times in Ircam, we called what is usually references as subtractive "Source/Filter" synthesis; essentially, most synth of the subtractive family can be considered structured as a sound source using some other synth technique (basic waveforms, fm, samples, waveshaping, wavetables, whatever) followed by a filter that subtract the unwanted part of the spectrum; by the way, once you crank up resonance, the term 'subtractive' do not describe what happening any more.

 

I would not consider FM part of the additive family; the main characteristic of the additive family is the fact the you add components together, with adding as the linear add operation. FM is part of a larger family of synthesis technique that you could call "functional" synthesis, where you apply a mathematical function to a signal (in the FM case, the sin function itself).

 

Just for fun, there is a technique that use just addition and multiplication, starting from a sinusoid, is the polynomial synthesis technique; for example, consider a function written like x*x+x (second order polynomial); apply this function to a sine, and you will get a sound composed by the sine and its second armonic; ok, not especially useful, but possible :->

 

To conclude, it is interesting to note that additive synthesis is just like frequency domain sampling, and have the same fundamental problem of a PCM based engine; with a spectrum, like with a sample, you can represent any sound that can be reproduced by a digital instrument; the problem is to choose an interesting spectrum, as it is choosing an interesting sample.

[Dr88s]

Nothing to add to the above, other than to point out that we now have a Tusker, a Fusker, AND a Husker.

[Synthoid]

http://cdn1.viewpoints.com/pro-product-photos/000/025/012/300/1271439582-43036_full.jpg