If I'm not mistaken, the most powerful additive synth currently available is NI Razor, which came out nearly 15 years ago. Air Loom II could maybe contend with Razor, but isn't necessarily as useful for sound design.

Most other additive synths I've seen only go up to 32 or so partials, which is hardly useful unless you're making more boring bell sounds.

Why is there so little innovation in this area, is there some technical limitation I'm not aware of?

 

It's tedious manipulating every partial. You can do it in Alchemy, Vital, Serum (512 bins), Fathom, VPS Avenger, MSoundFactory, many others. That's what makes Razor and Loom II so good, they have ways of changing a lot of partials at once that have useful results. So any innovation is in coming up with other ways of doing that. MSoundFactory is nice since you can capture partials from samples then modulate each one independently (can't remember is it's 32 max or more).

Partials reach Nyquist very quickly. For instance, concert A (440 Hz) hits it by the 50th or 54th partial depending on 44.1/48 kHz sampling rate. 512 partials like Serum is pretty good, it will be less than Nyquist only if the overtone series starts below 43 Hz.

Wavetable synths are often fundamentally additive. The waveform is (simplifying here) broken into partials for band-limiting. That's why a lot of them show the partials; they already did the math. Wavetables are a lot easier to use, you just sweep and possibly morph between additive states, and some synths allow you to warp each frame. That's two modulators vs however many partials you want to manipulate.

What would be innovative, and I hereby allow anyone to use this idea, is to figure out a way to present an additive synth interface that allows manipulating multiple inharmonic frequencies in a way that is useful. This is where wavetable synths also fail, as they cannot have inharmonic frequencies (there is research doing this via other means so it's coming). The thing is, there are other approaches that do this better at this time (e.g., physical modelling, chaotic oscillators, Waverazor's waveform chopping thing).

 

Well since Hammond Organ or Synclavier there isn't much room for radical innovation the same as with Subtractive synths which in their core are all Minimoogs.
The most recent Additives are UVI Falcon, Arturia Pigments, Musicdevelopments Syne (20000 partials!). I like MPower Synth by Melda.

 

Last edited: Nov 24, 2021

Helluva first post!

 

So, a couple things confuse me in this response, I was wondering if you wouldn't mind clarifying? Firstly, the number of partials required to hit the nyquist frequency. I was under the impression, that for bass sounds at least, you need ~500 partials to fill the spectrum up to 20khz. There is a tutorial Au5 posted that shows how to increase the number of partials in Razor from 320 to 1500+, and there is a noticeable difference in high end clarity between 320 and 1500. Not to say that 1500 partials is always necessary, as it would just be overkill in many situations, it is far more than 50 or the 24-32 that most additive synths provide.

Secondly, regarding wavetable synthesis. I was aware that synths like Serum/Avenger/Vaporizer use FFT data to modify wavetables, but I thought the wavetables themselves were fundamentally samples of actual audio with each frame being point in time/duration of the sample. And I think that is a bit different than having the root of the waveform sourced purely from an arrangement of sine partials. Of course wavetables are easier, because they're already made and you don't have to do anything, but by that philosophy you might as well sample everything because that is easier. Most wavetables also sound bad, imo, very rare to find usable ones that aren't derived from the basic waveforms.

I agree that finding useful algorithms for placing partials is the most challenging part to additive synthesis. But at the same time, additive synthesis is likely the only way to design sounds with no limitations, and that is why I'm excited about it. I've used waverazor a little here and there and I think it sounds awful for the most part, so I'm not too hopeful on that front.

Suppose I did come up with an idea for an interface that facilitates useful control of harmonic and inharmonic partials (It is something I've thought about quite a bit), what should I do with it? Not like I can code a cutting edge piece of software by myself.

 

The pigments additive engine is okay, bit it's 1/4 as good as it could be if the developed it more and made it into it's own synth.

MPower Synth is a good mention, I'm still learning my way around that one.

Also, I haven't heard of Syne before, but I'm checking it out now. Looks interesting.

 

Another one is discoDSP vertigo.

 

Yes - "discoDSP updates Vertigo additive synth to release 4.2"
https://www.kvraudio.com/news/discodsp-updates-vertigo-additive-synth-to-release-4-2-53403

 

Maybe this one - http://www.musicdevelopments.com/syne.html

 

Harmor can generate 512 sine wave partials per note, per unison voice and modulate them in real time to resynthesize any continuously evolving sound.

 

Yes, but if I'm not mistaken you can only control the amplitude of the partials with a filter/harmonic mask. You can't easily add to or manipulate the actual harmonic content, unless you use the oscillators and prism feature, which is great but still limited even compared to Razor.

 

You'd be right, Alchemy lets you modulate the partials too.

 

"Syne is a new, fully modular additive synthesizer capable of rendering more than 20000 partials on a recent CPU"

 

Im partial to additive synthesis.

 

I've seen some recent innovation, albeit in hardware:

The Panharmonium. What's cool: It's a realtime processor.

 

It wouldn't be bad if someone made a proper emulation of a legendary additive synth Kawai K5000S. It's notoriously hard to edit sounds or make new ones from scratch even worse, so the emulation should focus on easier editing.

I just don't want Arturia to do it. Please don't do it! u-he or Synapse could do a crazy good emu of K5000S, I'm sure of it.

 

Image Line - Morphine VSTi
www.image-line.com/fl-studio/plugins/morphine/

 

Perhaps it's because I'm a bit on the younger side, but I just don't understand people's obsession with emulating vintage analog synths. I want new technology, not rehashed technology from decades ago. I get that companies can make a larger profit on something they invest 0 R&D into, but I don't know why customers eat it up like they do.

If Morphine is so fresh and innovative, why is it being used by literally NO ONE??

You would think a company like Image-Line would be able to back a capable product until it becomes widely adopted.

But nope, everyone's still using Razor from 2010.

Also, I've never heard anything made with morphine that didn't sound absolutely awful.

 

You're right - it doesn't sound great, that's why it didn't prevail. I think this is very good, it sounds really great.

AIR Music Technology - Loom Classic - Modular Additive Synthesizer

As an additive instrument, Loom can create sounds that extend far beyond the scope of subtractive synths.

Loom provides a vast array of unique sounds while providing total control over every aspect of the instrument.
Sound creation begins with 30 sound modules, each with a distinct function—moving filter, octaver, repeater, second tone, etc.

These modules can be freely combined in up to 10 cells. Spectral modulation, extensive envelope shaping, and multiple LFOs deliver deep control:

• 30 distinct sound modules, freely combinable in 10 cells.
• Point-and-click sound morphing with powerful control.
• Over 350 patches.
• Smart Sound Randomizer to quickly design unique sounds.
• Economy mode reduces the load on the host processor with minimal impact on the sound.

Was $99.00; Save $60.00 until 8 Dec 2021!
www.kvraudio.com/product/loom-classic-by-air-music-technology
https://airmusictech.com/product/loom

 

Last edited: Nov 25, 2021

You mixed up things that is better not to mix up.

Razor is a synth where you synthesize sounds from scratch. From nothing.

In Morphine you resynthesize sounds. You load a wave file and use it. You can load some wave files and attach them to different zones. It's like a sampler but based on a resynthesis. Is it possible in Razor?

Yes, overall sound of Morphine isn't the best. More or less muddy. But this is a capable and useful instrument. It's a pity that IL stopped to develop it.

It's very hasty to say that it sounds "absolutely awful". How many songs have you heard? One, two, three? Do you know sources of all sounds you hear in music?

I defenitely don't feel anything special regarding Razor. These "harsh" leads... Where a could use it? Nowhere. They are are right just to amaze teenagers. Yes, its pads, bells are useful but there are many synths that aren't worse in it.

P. S. Morphine is very old. From 2007. https://www.kvraudio.com/news/image_line_releases_morphine_additive_synthesizer_7492