Another beast:


https://zynaddsubfx.sourceforge.io/zyn-fusion.html

 

VirSyn - CUBE 2 - Spectral Morphing Resynthesizer
www.virsyn.de/en/E_Products/E_CUBE/e_cube.html

Software Description: CUBE 2 is a musical instrument and the ultimate tool for sound designers who wish to have instant and intuitive access to all principal components every imaginable sound is build from. Every sound can be decomposed in a sine spectrum representing the harmonic and inharmonic partials and a residual noise spectrum for the noise components. The spatial character of a sound can be represented by a pan spectrum. The temporal evolution of these three spectra is a very convenient way to describe every sound. CUBE 2 is built around this model and allows you to manipulate the structure of a sound with all its components in a comfortable and musical way.

VirSyn CUBE v2.2 (2003-2007) [SoftSynth, Demo]

 

Hi ArcticStorm, you are absolutely right, apart from Albino we also have:

LinPlug - Spectral is the result of more than 14 years of synthesizer building experience. Someone being familiar with subtractive synthesis will feel right at home with the Spectral, because one can choose an oscillator waveform, choose a filter shape and go on from there. But Spectral delivers far more options than one would expect from the well sorted user interface. Both Oscillator Waveform and Filter Shape can be edited in depth and created from scratch. This unique audio engine is capable of delivering many sounds never heard before.www.kvraudio.com/product/spectral-by-linplug

ImageLine - Sytrus VSTi
Sytrus is a powerful and versatile synthesizer featuring six customizable oscillators (operators). It can perform FM (Frequency Modulation), RM (Ring Modulation/Amplitude Modulation), Subtractive and Additive synthesis. It includes 3 filter modules, an effects module with chorus, three delay lines and unique, per-voice programmable Unison envelopes. www.image-line.com/fl-studio-learning/fl-studio-online-manual/html/plugins/Sytrus.htm

 

I tried to make my teeth on this one a little bit, but damn, the control and ergonomics just seem to be an non existent concept at the time this was created... a shame because it brings some funny cheesy sounds, telling a kind of a 90's-2000's story. But I think I need a least one plain hour of non fun to begin to get how things flow within this one.

Additive / partial re/synthesis should be pushed and developed, there's a very lacky and empty space to fill with new ideas I completely agree with you @Noizey98.

However you shouldn't underestimate gol's works, those (Harmor, Harmless, Ogun, Morphine, among its most famous "FL Studio"...) were truely ahead of their time when around there were only shitty sounding virtual analog stuff.

Even today I don't find them sounding shitty at all

 

Recognition and praise from me too!

At that time there was hardly anything good and then the prices. Today, thank God, we have way too much and therefore cheaper prices. Whoever wants to sell a virtual synthesizer today has to offer something and come up with something.

The last one I'm playing was the Bolt from Nektar with a new harmonics synthesis. Not perfect but the instrument sounds balanced. The Vital Synthesizer was actually the surprise last year. Rob Papen's B.I.T is not bad either, but it cannot prevail. The Europa from Reason was good with a great Karplus Strong Wave. Very balanced and good sounding. With 2 modifiers and a spectral filter.

It was surprising what the company Togu Audio Line (TAL) brought on the way, TAL-Mod and NoiseMaker. Adam Szabo brought the Viper as Virus TI clones. The Surge Synth Team did a great job reviving the Surge, it's free even now. HY-Plugins - shine and surprise with the HY-Poly. The companies Spectrasonics, Korg, Arturia, U-HE and Waves have brought all their plugins up to date, even 4K monitor capable.

Other highlights are the Parawave - Rapid Synth, KV331 - SynthMaster One, Tracktion Software - F-em, ZynAddSubFX - Zyn-Fusion, Vengeance - Avenger, UVI - Falcon and Wavesequencer Virtual Instrument - Hyperion Synth.

 

Last edited: Nov 25, 2021

Haven't seen this one before, I'll have to check it out.

Cube 2 / CubeSynth Pro actually produces some of the best sounding examples of additive synthesis I've heard. It's too bad the dated interface and steep price tag are such a turn off.

Or why not just have capable additive synths? To my knowledge, the main difference between filters in a subtractive synth vs. and additive one, is that the subtractive filters cause slight distortions in the phase/frequency of the effected harmonics, whereas additive filters do not because they are manipulating the sinebank data at its source. That is why additive "filters" typically sound sterile in comparison to subtractive/VA ones. However, all it would take would be for someone to create an additive filter function that emulates the distortion caused by the subtractive filters, and you would essentially have the best of both worlds in one synth. And that is just one example of untapped possibilities.

I would actually agree with you that those IL products were ahead of their time, and aren't bad instruments. But, their time was 10-15 years ago.

Think about how much wavetable or granular synthesis has changed in the last 10 years.

We haven't had a "Serum of additive synthesis" (for lack of a better term).

 

The fault lies in the system, why does a company like LinPlug or Camel Audio have to go bankrupt.

It is hardly possible for you to get together and ask the users how much money they would give and what they would like. Imagine
the LinPlug Albino would now be updated and 4K Ready or Camel Audio - Alchemy would be adapted to the times and modernized.

The company TubeOhm made the Pure-D16 24Rack virtual software synthesizer, 32 bit and then it was over again ...
and so the blood trail of the collective destruction of software synthesizers runs through our lives.

We finally need a new, sustainable business model.

 

Well you can partially blame it on the market. Survival of the fittest.

But yes, I do agree a better business model would be a win-win for everyone. Because obviously no one HAS to pay anything for software these days, the best thing would be either a donation-based system or a tiered system that charges amateurs/aspiring professionals less than actual pros instead of just pricing them out of the market.

 

Harmonic partials are multiples of the fundamental. So divide the sample rate by two (for Nyquist) then again by the fundamental frequency to get the minimum number of partials required to reach the highest frequency before Nyquist.

Similarly, to determine the lowest note a number of partials can reach the highest note with before Nyquist, divide the sample rate by two, then by the number of harmonic partials, then look at a note frequency chart and find the first note above that frequency: https://pages.mtu.edu/~suits/notefreqs.html

For example, with 512 partials, F#1 and above are fully represented to Nyquist at 44.1 kHz.

With 320 harmonic partials you must play C#2 or higher to reach the highest frequency before Nyquist (@ 44.1 kHz), but any note above that there is no difference. Look at the partial display in Razor. Start from C0 and work your way up. You will notice the highest partial will increase to the right before hitting a wall. 1500 partials guarantees notes above 14.7 Hz, so C0 and above.

Razor implemented the partials limit to accommodate CPUs at the time. You can also increase polyphony. It's important to note Razor deals only with harmonic partials. Any oscillator with a name indicating inharmonicity doesn't (e.g., noise pushes all harmonics to max amplitude and modulates it).

Like I said, I simplified a lot. This is a really succinct explanation of the problems with and better solutions than naive wavetables (what you are describing): https://ccrma.stanford.edu/~stilti/papers/blit.pdf

Basically, the current best practice is to decompose a wavetable frame into harmonic partials and filter everything else. There are older wavetable synths with much different (non-additive) approaches that have some kind of problem.

Spectral synthesis is related. It stores audio into frequency bins, so with enough bins you can store sounds. You can see a commercial approach in Iris 2. The Sampleson instruments are built this way, but those developers have done the work of analyzing and selecting how the frequency bins change over time algorithmically as opposed to storing data.

I took another look a MSoundFactory and the additive model is great. While there are 64 partials per module you can stack up to 90 for a total of 5760 partials. Also, each partial can be tuned and amplified however you like, meaning it's not locked to harmonic frequencies. You have 4 states, A-D, which you can interpolate between as corners on an XY pad. It can analyze a sample to determine partials. It's still a good amount of work to create good, workable states.

If you want it made, write a paper or article and perhaps a proof of concept in Reaktor or whatever and give permission to use the ideas, and maybe someone will (see how The Infinite Drum Machine resulted in Atlas and XO, and how the paper I linked previously resulted in massive improvements across all current wavetable synths).

 

Wow, never heard of this before but I have seen another Virsyn product, just don't remember what it was. Thanks for bringing this to my attention, I need to check this out. I love resynthesis and resampling. Can't believe this is the first I hear of it.

 

I may try to make a reaktor or synthedit prototype, however there is one problem. A central component of a useful interface is how synthesis process is visually represented and a proper forward-thinking additive synth would need some type of realtime 3d render to show the frequencies/amplitudes/phases of the output harmonics. Seems simple enough, but when you consider inharmonic frequencies and phase cancellation between harmonics, it really is quite complex. Given there exists any program that can accurately visualize/render that type of data in realtime (I currently cannot find any), I would need to find a way to fully replicate it or embed it within reaktor or synthedit, which I assume is likely not possible.

And to create the algorithms in the additive engine I would need lots of complex mathematics and coding which I am also mostly ignorant in.

 

MSoundFactory and other Melda stuff are so underrated. Of course it needs some efforts to get through the interface, but if you succeed you'll be rewarded with most powerful tools on the market.

Edit: BTW besides the additive module the MsoundFactory has an Oscilator Module which is additive too

 

Last edited: Nov 26, 2021

The Loom 2 has the power you need. I had to look again loom 2 is working:

AIR Music Technology - Loom II VSTi - Concept

The so-called "additive synthesis" is used as a synthesis process for the Loom II. The process is called additive synthesis because the overall sound structure is created by adding individual components (partial) and not, as is the case with classic subtractive synthesis, by attenuating spectral components in which the oscillator (s) are followed by filters. Good old Mr. Fourier, who may be known to one or the other from his studies or mathematics class (Fourier series), sends his regards.

This is why one sometimes speaks of the Fourier synthesis. This synthesis process has been around for a very long time, but initially had more theoretical / scientific importance. It was only when the available computing power of the PCs increased that they dared to put it into practice. The well-known Synclavier was one of the first synthesizers to also master additive synthesis. By the way, Loom also offers filters that can be looped into the virtual signal path, so if you will, Loom II offers a combination of additive and subtractive synthesis.

With additive synthesis, a sound is generated by individual sinusoidal oscillations with different frequencies and amplitudes. It is important to know that the sine waves above the fundamental tone do not necessarily have to be harmonics of the fundamental tone. The dynamics in the course of the sound are made by changing the individual amplitudes of the individual sinusoidal oscillations and, in the case of Loom II, by dynamically changing many other parameters.

Source: www.proaudio.de/de/tests/22296-air-music-technology-loom-2.html

 

I add to the list: Camel Audio - Alchemy & Logic Pro - Alchemy, LinPlug - Spectral, LinPlug - Octopus, Image Line - Sytrus

LinPlug - Octopus
- Octopus is a 12-way polyphonic dual matrix synthesizer from the Berlin developer LinPlug. The plug-in contains two independent sound generation modules. On the one hand, FM (frequency modulation) is used as a synthesis method, on the other hand, samples can be integrated. Octopus is equipped with eight oscillators. Each oscillator has 32 harmonic overtones, the additive waveforms can be created by the user. The plug-in is also equipped with two multimode filters. These can also be used as FM sources.

LinPlug - Spectral
- Each oscillator can be activated by right-clicking on the selector switch and can generate or load two different waveforms (A / B) that can be cross-faded; there are ready-made ones in the library. If you click on the wave display, the associated additively working machine room opens with up to 256 freely editable partials, for which amplitude and phase position can be set. The octaves are marked light / dark.

Image Line - Sytrus
- Sytrus is a powerful and versatile synthesizer featuring six customizable oscillators (operators). It can perform FM (Frequency Modulation), RM (Ring Modulation/Amplitude Modulation), Subtractive and Additive synthesis. It includes 3 filter modules, an effects module with chorus, three delay lines and unique, per-voice programmable Unison envelopes.

Camel Audio - Alchemy
- Alchemy is described by Camel Audio as "the ultimate sample manipulation synthesizer".Alchemy features additive, spectral and granular synthesis and resynthesis, sampling, and a very capable virtual analog engine with unison and PWM. You can morph or crossfade between sources. You can import your own samples from SFZ, WAV or AIFF files. A wide range of analog modeled filters are included.

Links:
1.) Additive Synthesis Explained – What is Additive Synthesis? https://musicproductionnerds.com/additive-synthesis
2.) An Introduction To Additive Synthesis www.soundonsound.com/techniques/introduction-additive-synthesis
3.) Learning about additive synthesizers www.musicianonamission.com/additive-synthesis-subtractive-synthesis/

 

Last edited: Nov 26, 2021

Bankrupts? No. Why? Linplug just ceased operation (they just lost their will to do it, but I must to say that it's my impression only based on their announcement), Camel Audio was bought by Apple.

 

.

 

Last edited by a moderator: Jun 7, 2024