Additive Synth VST/AU Thread

[aMUSEd]

Alchemy was already mentioned - Akoustic looks interesting but appears not to have been developed for some time since they are still mentioning it being Catalina ready, which I guess also means it's not Apple Silicon ready or available as VST3

[martiu]

oh yeah, i have Akoustic, it is amazing for percussive sounds, sometimes it feels like you are using physical modeling synth

[carrieres]

dune3 has an additive tool with 32 partials (afair, will check the max number)

[j wazza]

In bitwig, or a modular like reaktor, vcv, vm, you can create partials with midi notes and then send them to an mpe polysynth to use polyphony for the harmonics. Can make really cool sounds. By chaining multiple of these midi note generators you can add harmonics on to harmonics to multiply them. It also means you can use per voice modulation to control lots of partials at once. Can get really tough on cpu though

[Nug Wrangler]

Sytrus can do additive synthesis

[carrieres]

and Ableton Operator

[j wazza]

You can do a lot just with standard waveforms. Eg a sine + a saw an octave up gives you all the even harmonics. A square is odd harmonics, so crossfade between them and you have an odd/even knob to make a morphing spectrum, and you can do that with analog synths

[N__K]

Sytrus can do additive synthesis

and Ableton Operator

Only for setting a static waveform for their oscillator(s). In those synths, properties of individual partials cannot be modified later in the signal chain.

In other words, they only offer what can be called "static additive synthesis".
That's what the whole concept of additive synthesis is still often misunderstood as: only creating the initial waveform.
In past 30 years, that was indeed mostly the extent of implementations, due to lack of processing power.

Nowadays I'd expect an additive synth to offer "dynamic additive synthesis" - that is, being able to change the parameters of partials while the sound is playing. Technically it is done by having all (or most) of the signal chain be metadata for changing parameters of partials, instead of a waveform to do signal processing on.


***


Image-Line's (Maxx Claster's) Morphine is an example of "dynamic additive synthesis", while limiting the scope to providing envelope breakpoints (or "frames of animation") for parameters of the additive oscillators. In other words, it's an implementation of the basic theory of additive synthesis with "low level" direct control, so to speak. Making sounds that way often gets laborious.

In comparison, Image-Line Harmor (programmed by gol) also offers "dynamic additive synthesis", but with more abstractions to control its processes at a higher level*.
Its "frequency-domain" modules [red-tinted in GUI] are designed to simulate many common methods in synthesizing sounds, and so offer quicker ways to control the properties of sinewave partials.
The parameters of those modules are extensively user-adjustable, have their own envelopes etc., and for majority of them, maintain the possibility to change the properties of the additive synthesis as the sound is played.

That's why Harmor was/is so much of a paradigm shift, and still seems largely misunderstood.

*it must be noted that Harmor also allows direct low-level control over properties of partials via "Shaping" > "Timbre # harmonic level" etc. in ENV tab, and animation possibilities in IMG tab. It is comparable to what can be done in Morphine, but somewhat more laborious due to design of Harmor's interface (as we know it from the year 2011).

[Nug Wrangler]

Sytrus can do additive synthesis

and Ableton Operator

Only for setting a static waveform for their oscillator(s). In those synths, properties of individual partials cannot be modified later in the signal chain.

In other words, they only offer what can be called "static additive synthesis".
That's what the whole concept of additive synthesis is still often misunderstood as: only creating the initial waveform.
In past 30 years, that was indeed mostly the extent of implementations, due to lack of processing power.

Nowadays I'd expect an additive synth to offer "dynamic additive synthesis" - that is, being able to change the parameters of partials while the sound is playing. Technically it is done by having all (or most) of the signal chain be metadata for changing parameters of partials, instead of a waveform to do signal processing on.


***


Image-Line's (Maxx Claster's) Morphine is an example of "dynamic additive synthesis", while limiting the scope to providing envelope breakpoints (or "frames of animation") for parameters of the additive oscillators. In other words, it's an implementation of the basic theory of additive synthesis with "low level" direct control, so to speak. Making sounds that way often gets laborious.

In comparison, Image-Line Harmor (programmed by gol) also offers "dynamic additive synthesis", but with more abstractions to control its processes at a higher level*.
Its "frequency-domain" modules [red-tinted in GUI] are designed to simulate many common methods in synthesizing sounds, and so offer quicker ways to control the properties of sinewave partials.
The parameters of those modules are extensively user-adjustable, have their own envelopes etc., and for majority of them, maintain the possibility to change the properties of the additive synthesis as the sound is played.

That's why Harmor was/is so much of a paradigm shift, and still seems largely misunderstood.

*it must be noted that Harmor also allows direct low-level control over properties of partials via "Shaping" > "Timbre # harmonic level" etc. in ENV tab, and animation possibilities in IMG tab. It is comparable to what can be done in Morphine, but somewhat more laborious due to design of Harmor's interface (as we know it from the year 2011).

ah i see. Morphine is a better choice then

[zerocrossing]

You can do a lot just with standard waveforms. Eg a sine + a saw an octave up gives you all the even harmonics. A square is odd harmonics, so crossfade between them and you have an odd/even knob to make a morphing spectrum, and you can do that with analog synths

I was joking around at one point and said, "isn't all synthesis spectral?" It was a joke, though. It's not stated, but when we talk about spectral or additive synthesis, we're specifically meaning that one has arbitrary control over partials in a way that one doesn't with an analog synth.

[j wazza]

You can do a lot just with standard waveforms. Eg a sine + a saw an octave up gives you all the even harmonics. A square is odd harmonics, so crossfade between them and you have an odd/even knob to make a morphing spectrum, and you can do that with analog synths

I was joking around at one point and said, "isn't all synthesis spectral?" It was a joke, though. It's not stated, but when we talk about spectral or additive synthesis, we're specifically meaning that one has arbitrary control over partials in a way that one doesn't with an analog synth.

Why did you post this?

And that's not what additive synthesis means. There's a difference between additive synths and additive synthesis, but that's not what either means. There are analog additive synths like the verbos harmonic oscillator. It's also possible to do DIY additive synthesis in synths that aren't marketed as additive synths, including analog ones, and get some of the same control over the partials as you can get with an additive synth. Not all synthesis is additive, filtering a saw wave is subtractive synthesis. Additive synthesis means adding harmonics. Additive synths are designed to do this, but you can also do this in other synths

[j wazza]

Lots of additive oscillators limit the spectrum to the harmonic series, but real sounds usually have frequencies outside of the harmonic series. With falcons additive osc, you can add dissonance and stretch the harmonics. Unison detuning is another way to add dissonance, but it usually adds the same detuning to all partials, and the relationship between the frequencies is important. You can get a richer sound by layering oscs with different partials and different amounts of detuning. Or even adding different detuning to each sine individually.

I also like to add inbetween partials that are still harmonic, but not in the harmonic series (frequency ratios 1,2,3,4,5 etc)
Layering oscs at a ratio 1.5 can do this (up a fifth), or 2.5 etc.
Or layering subharmonics like 0.5 (an octave down) with its own harmonic series will add harmonics 0.5, 1, 1.5, 2, 2.5 etc. Or layering 0.3333 etc.

These partials often appear in real sounds. I sometimes like layering these ratios, but removing the fundamental and first few harmonics and just using a few of the upper harmonics made from a fifth up or an octave down etc. This avoids changing the fundamental pitch that we hear, and avoids turning it into a power chord or octave chord, so it will still sound like a single note at the same pitch, but with a more complex spectrum. And again, adding a different amount of detuning from the main oscillator can make a richer, less sterile sound.

Phase plant is perfect for this as it has additive oscillators, you can layer as many as you like at different ratios with their own unison.

[zerocrossing]

You can do a lot just with standard waveforms. Eg a sine + a saw an octave up gives you all the even harmonics. A square is odd harmonics, so crossfade between them and you have an odd/even knob to make a morphing spectrum, and you can do that with analog synths

I was joking around at one point and said, "isn't all synthesis spectral?" It was a joke, though. It's not stated, but when we talk about spectral or additive synthesis, we're specifically meaning that one has arbitrary control over partials in a way that one doesn't with an analog synth.

Why did you post this?

And that's not what additive synthesis means. There's a difference between additive synths and additive synthesis, but that's not what either means. There are analog additive synths like the verbos harmonic oscillator. It's also possible to do DIY additive synthesis in synths that aren't marketed as additive synths, including analog ones, and get some of the same control over the partials as you can get with an additive synth. Not all synthesis is additive, filtering a saw wave is subtractive synthesis. Additive synthesis means adding harmonics. Additive synths are designed to do this, but you can also do this in other synths

No. Filtering out a partial with a resonant filter isn’t what anyone calls additive synthesis.

[KBSoundSmith]

Tracktion F'em is quite good as an additive synth, if you use all four layers. If you limit yourself to just the traditional operators (exclude the noise, sample, and filter operators), then at 8 operators per layer, you get 32 partials. Each partial can be tuned however you want, and each has their own dedicated envelopes and LFOs (two of each, per operator), plus the other LFOs and modulators (of course, the "dedicated" modulators can be assigned to everything else too). Each layer also has its own FX chain, Arp, EQ, modmatrix, unison and other voice settings, etc. Very powerful.

[j wazza]

Not all synthesis is additive, filtering a saw wave is subtractive synthesis. Additive synthesis means adding harmonics.

No. Filtering out a partial with a resonant filter isn’t what anyone calls additive synthesis.

I didn't say that, I said the opposite. So much time could be saved on here by people reading before replying.

I also like the bogaudio additive oscillator in vcv rack