I have not heard of any dedicated vsti's for this type of synth (I hope NI will clone the Kawai k5000 some time); is it implemented in Reaktor? Or somewhere else? I don't know much about it, and would like to play around with it. Any software which will do the job?

Thanks,

e.

I guess most VST instruments do additive synthesis. You're talking about mixing (or adding) waveforms aren't you?

jorgen

I am writing an additive synth. Look for more news this summer.

You can't see me, but I just broke out in a sweat! Give us a tiny morsel of what to expect, who will it come from? There are very few hardware synths I refuse to live without these days, but my K5000 additive is one of them. I just haven't found a real software replacement for it. Could it be that I will soon?
Scot Solida
Christus And The Cosmonaughts

That's interesting, it was the first of my hardware synths I decided to get rid of. I used SCORPION to replace it in most of our songs and it does a really good job. I tried to get into the additive side of the K5000, even bought the Wizoo book, but it never really got my juices flowing.

I had a K5000S but it crapped out on me a couple of times. I loved how it sounded but the one I had was ultra sensitive.

Reaktor can do some limited additive synthesis but it's really the filter and LFO stages where you can make it interesting.

Rainbow synth does a hybid of FM and AM. It's a interesting VSTi. It's not anywhere as flexible or complex as the K5000 but it often reminds me of it timbre wise.

Yeah, I had forgot about Raibow, which I have toyed with. Hope to buy it some day. Thanks for the input, all.

I too am looking forward to Malouin's project. Do you have others? I don't know everyone here.

Jorgen, as I understand it, additive synthesis gets more complicated than at first might appear. It is apparently one of the more difficult types of synthesis to implement, which is why few hardware synths (the Kawai K series being the exception) and almost no VSTis take on the task.

I think it involves adding very many waveforms, each with a separate envelope to produce something which "looks" like a fourier transform diagram. Should be easy for a computer, but for some reason it isn't. I guess I'll go do a search on the subject.

e.

Rude of me to post so many times in a row, but:

http://www.sospubs.co.uk/sos/jun00/articles/synthsec.htm

should help.

Absynth does additive

Klause Schulze recorded "Dig It" (get it?) in 1980 using the Crumar General Development System (GDS for short). Listen to it on CD so that you can hear how that technology trickled down to the Synergy so favored by Wendy Carlos and eventually the Kawai K5000. Personally, I always found the sounds dry and lacking and switched to sample-playback systems w/f/x like the Korg M1. I still own a Korg O1R/W as my only hardware module.

One of the best tools to understand additive synthesis is Windows Media Player 7.1 with the bar setting. I love listening to old spooky radio shows streaming online and watching how speech patterns affect frequencies is amazing!

"Absynth does additive"

And fortunately there is gonna be a windows version...

Aren't drawbar organ simulations such as Rumpelrauch Taips (hope I spelled that right) ZR-1 considered additive synths?

I know this has been said before but: Complex additive on the scale on the K5000 is still too DSP intensive for modern PC's. Even hardware Units like the Nord Modular can't do complex additive. Basic addittive synths (e.g drawbar organ simulations) are not a problem but for complex and involving sounds you need more than a handful of oscillators. It can be quite hard to make decent sounds from scratch (unless you're a maths boffin), but hopefully someone will someday create a more intuitive way of programming and modifying sounds. Absynth looks promising but I think you can only edit the first 64 harmonics of the waveform which is dissapointing.

FM synthesis is in fact additive synthesis. There are several forms of Additive synthesis and simply stated every system that generates sound by adding harmonics to build a sound in any way is Additive.

BigTick's Rainbow 2 was destined to have a form of additive synthesis but it seems the development is halted for now.

I remember the GDS being very close to the actual working of FM because it worked also with sinus waves and there where options to let them modulate each other. I have some information on the GDS somewhere but I have to search through boxes with paper to find it.

[ 23 March 2002, 19:22: Message edited by: crimsonwarlock ]

Yeah, FM is a basic form of additive, and the DK Synergy actually HAD FM synthesis on it, though they weren't allowed to name it as such, 'cause of Yamaha's patent. But personally, I want to see the REAL DEAL. I'd like to see essentially a K5000 with multi-segment pitch envelopes for each harmonic, as well as amplitude envelopes. Does anyone hear know of something like this? Man, I'd be all over that like Homer on doughnuts! Mmmmmm, additive....
Scot Solida
Christus And The Cosmonaughts

Thanks Scot for a lucid (and compartively short) descrtiption of the idea.

Couldn't a trepiditous vsti programmer set things up so you could "tween" between harmonics/envelopes if you so desired, to reduce the tedium/fun of programming?

Yeah, you could copy envelopes, or use maybe "template" envelopes. These kinds of things reduce the migraine potential on the K5000, certainly. It's still gonna be a tedium to some folks. When I try to explain it to friends, they usually find even the suggestion of working this way too much to deal with. Hey, I like, no, I LOVE analog subtractive (gotta roomful of metal and wood beasties), but working with analog is simple, and you can only go so far. Analog is ideal when you just wanna make a sound right NOW, but additive give you infinite possibilities. When I tell my kid "I love you", that's additive synthesis. All the sine waves combining and interacting over time, and through all the modulations between my lungs and ears. Sure, it's hard to program, but think of what you might do with that kind of power...Like I said, I'm not giving up my other synths, 'cause they allow me to make sounds and write music in realtime. Additive is something for when I'm feeling contemplative.
Scot Solida
Christus And The Cosmonaughts

Scot

It was for Rainbow 3, not Rainbow 2.
And, it ain't halted, but it's going very slow...
Cheers,
'Tick

I've just begun working on this additive synth about a week ago, after several years of 'planning' (random thoughts while working on another project). It will be a commercial synth, running standalone and as a plugin. I plan to have something workable this summer, but the Windows version may be delayed a bit relative to the Mac OS version. Historically there hasn't been much of a market for additive synths, so I'll have to gauge the level of interest and balance between this and my day job.

I owned a K5 and K5m for many years, but this will be more similar to the Synergy and Kurzweil K150. There will be individual frequency and amplitude envelopes for each partial, fixed-frequency and non-harmonic partials, noise components, ability to morph between timbres in real-time, and other things you would expect. The multi mode will be like the K5, which I always liked. Analysis tools for building sounds will have to come later. If anyone has specific feature requests, comments, or suggestions I'd love to hear them. Please email malouin <at> mac <dot> com.

I also wrote ObjektSynth, a real-time modular synthesizer for the BeOS, so I have some experience writing low-latency software synths.

Malouin...Wow, that sounds quite nifty, just what I'm looking for. Well, except for the Mac thing. Too poor for a Mac. The analysis functions WOULD be neat. Always wished for the ability to resynthesize bigger waves on the K5000 via SoundDiver. Any more details you wanna let loose? What about Bick Tick and Rainbow 3? Any details? I just listened to the demos of Rainbow, and it already sounds tasty. Nice EP, by the way. And there's gonna be a free upgrade path to Rainbow three? Hmmmm, may have to start stashing a little bread away...
Scot Solida
Christus And The Cosmonaughts

Hey Scott nice to see you here, you're wealth of info on Scott Garrigus's Sonar forum, has got me out of a lot of jams with that program. Plus it's always nice to have more and more Sonar users in here

wrt platforms, although I prefer Mac OS I am not (overly) religious about these things. Simple time constraints.

VirSyn TERA has an additive osc. You draw into two frequency graphs and morph between the two states. It's not as deep as the K5/K5000 implemetation, but much easier to use.
VirSyn TERA demo

[ 23 March 2002, 12:19: Message edited by: Collusion ]

A drawbar organ is indeed a very simplified version of additive synthesis, though for good additive, you need at LEAST 64 harmonics to work with, each with it's own pitch and amplitude envelopes. The GDS and Synergy beasties mentioned above allowed you to use 32 harmonics each, I believe, which is not enough to fool the ear into believing it hears a natural sound. 64 on Absynth should be able to get a lot closer, but only if you have the envelopes for EACH harmonic. Anyone using it? Can you tell me? Seems like yet another plus for Absynth. You know, I've read a lot of reviews of Absynth, and not one mentioned it's additive capabilities. The K5000 allows you to use either the upper or lower 64 harmonics, and you can pile on up to six groups. Every harmonic has its own amplitude envelope, but sadly, not it's own pitch envelope, those are only available at the oscillator level, not to the individual harmonics. And they are fairly standard envelopes at that. For really good additive, you need a lot of breakpoints in the envelopes. The envelopes in the FM7 would be nice on an additive synth. The K5000 is 1997 technology, so I have to believe that a modern PC would be up to the job. As for it sounding cold or digital, remember that ALL sound is additive, so with the proper implementation it could sound any way you like, if you can stand the tedium of programming it. I can program dozens of patches in a day on any other synth I own (even my dear old Wavestation), but on the K5000, I consider programming more like working on a complex oil painting, and I work on a single patch for months. You know, maybe I'll try building a half-decent additive in Tassman, just to see what kind of processing power it might consume...
Scot Solida
Christus And The Cosmonaughts

Collusion:

Thanks for the URL; it looks like a worthy synth, but out of my price range for now. I hadn't heard of it before.

e.

Hey Malouin if ya need a beta-tester i'll be happy to test uout on my dual 533 g4, running Cubase vst 5.1r1/32

realmarco@videotron.ca

There is a full-blown Additive softsynth already, called Morphine. It has 128 harmonics, 20 oscillators, 10 envelopes and 20 LFOs. It is made for Sync Modular and it's processor usage is very low altough it's complexity (16 voices = 30% CPU usage on pentium 800 Mhz).

dashMORPHINe is 20 osc, 10 envelopes, 20 lfos and 2 delays with an astonishing low CPU use with an additive engine. What does additive means? Usual synths and samplers are "substractive" which means using a filter (high, low or band pass) to substract harmonics at some point, recreating them with a resonance control, starting with 1, 2 or 3 oscillators. Additive synths create unique sounds impossible to get by substractive synthesis based on some mathematics law by Fourier, that says that any wave, independently of how complex it is, can be considered a sum of simple sine waves at different harmonic ranges (or sines which frequencies are multiples ones of the others).

Check it out on Dash Synthesis (there is also MP3 demos):

http://www.dashsynthesis.com/synths/synths.php?synth=morphine

[ 24 March 2002, 09:03: Message edited by: Tarkus ]

Tarkus,
Thanks for the link. I haven't used SyncModular, but this looks like its getting close to what I'd like to see, so I'm gonna check it out. Also, I get the day off today, so I'm going to work on my Tassman additive. If I come up with anything, I'll post it on my website, and let everyone know. Problem is, the Tassman, like way too many modualars, has limited envelope stages. This will allow for some control, but not nearly enough. Collusion: VirSyn looks nifty, too. I went to your site by the way. Nice stuff. I can hear your Sisters influence. I tried to e-mail you, but it got sent back undelivered. My bassist DJs a Goth/Industrial night at a local club, here. I know he's going to want to play your stuff. That OK?
Scot Solida
Christus And The Cosmonaughts

I listened to a few demo-tracks for Morphine and I heard nothing that can't be done with other synths. Most of it has a definite FM kinda ring to it, but not as deep and moving as real FM sounds (especially the ones FM7 is capable of).

CW,

It's really down to some terminology issues. You seem to categorise everything as belonging to additive (whereby you enhance the spectrum by any means) or subtractive (whereby the spectrum content is reduced). This way, you divide the whole world of traditional synthesis techniques into two extremely unequal classes. In fact, all synthesizers use some kind of distortion that generates additional harmonics, should they be called additive synthesizers? I don't think so.
The classic sources define several classes of synthesis approaches, among them subtractive, additive, nonlinear, abstract, model-based. I don't know what do they teach in Amsterdam, but if they call additive any form of nonlinear, it is contrary to many classical books that I know.
The classic definition of additive assumes strict control of explicitly generated harmonic partials. It is not the case of FM, whereby one has very crude control on the resulting spectra.

so what would you call a synth that uses an oscilator into a filter, a second oscilator into a second filter, then mixes the 2 signals?
i would hardly call that additive synthesis.

your just trying to confuse people

Actually, I'd call that a supremely crappy additive at best! I think you're missing the point that while a filter can be ADDED to an additive, a good additive doesn't need it. By it's nature a filter is subtractive, and as soon as you touch it, you've left the path of additive synthesis. Technically, yes, an additive CAN be made of a whole lot of analog sine wave generators, but the amount of oscillators has until recently made it impractical. However, I believe the original subject of this thread was the question as to whether anyone was developing a "classic" additive softsynth. I understood this to mean additive in the style begun with the GDS or Synergy. Personally, I'd like to see the development of a hybrid in the style of my K5000 with an additive oscillator structure, but with subtractive filters added on. As we've seen with the FM7, these kinds of hybrids, if presented with an intuitive interface, could be very powerful indeed. And no I'm not claiming the FM is "additive", though it's a lot closer than the suggestion in the previous post!
Scot Solida
Christus And The Cosmonaughts

Telxoss,

I don't want to confuse anybody. Remember the DX7 (maybe you have one, I do). It has NO FILTERS, just six oscillators only capable of generating a sine-wave. FM at it's purest form. Now if we look at the definition given by MB (and quoted by me in my previous message) then PURE FM is indeed additive synthesis.

And yes, there are several synths that use Additive and Subtractive techniques in a mix. And no, you can not call these real Additive synthesis machines. But we are talking FM here. The FM7 from Native Instruments is a hybride system because it introduces complex waveforms (having rich harmonic content and therefor no mathematical base to calculate the outcome) and filters (pure subtractive synthesis) into the concept of FM.

FM on it's own is Additive synthesis by all definitions given so far (mine and MB's). If you want to oppose that statement, give a definition of Additive synthesis that excludes FM

My take on it: Fact is fm is a synthesis *technique* and it in no ways implies stacking plain oscillators at all necessarily. It is most common to output the volume of only one operator on an FM instrument, or all, but just because in Fm synthesis you can add together the oscillators dosent mean FM IS additive. Fm is obviously frequency modulation, which IS a differant type of synthesis from just layering sounds which are not having complex musical interactions together.
In additive synthesis, you add ONLY pure fundamentals (sine) to build the sound. That's the type of additive being discussed here. Sure you can take 6 operators on an fm synth and make them not interact with each other and stay pure sine waves mixed like a drawbar organ, but that is not when using fm (or even just oscillator feedback because that changes the waveform drastically).

The second you start modulating any of the oscillators with each other you change the wave coming from those oscillators from sines into extremly complex, brite rich waveforms! Doesnt it seem obvious that this is the seperation between fm and additive? With additive, you are purely mixing funamendal tones, with fm just the output of one sine osc modulated by another sine osc can have one of the most complex metallic sounding wave you can think of, it's not the *addition* of those two sine waves together, but rather their musical *influence* upon each other under frequency modulation synthesis.
In fm, the FM creates the waveform, not the bare oscillators themselves, unless there is no fm and then they are pure sines, and sound pure, which is additive, or could be just the one osc.

My point is, that two osc's in an analog synth are not additive synthesis as what this thread ia talking about, nor is fm synthesis just because it's calculated using two to several pure tones. Fm is it's own type of synthesis, like subtractive or physical modeling, you dont need to add the output of two tones, all can be muted except one, thus it's the fm creating the charactor of the tones. If it's not then we could say that subtractive synthesis is also additive synthesis, because it uses the output of one or more osc waves too. That's my interpretation added to the pile of em'

CW,

The fact that the output spectra of FM is precisely given by the series of Bessel functions does not imply you have a control of the spectra. It only tells you what the output spectra will be, given the FM parameters, but it does not allow you to program a sound of a given spectra. That is because Bessel functions are not reversible and no analytical solution exist for inverse transformation series. Oh, and BTW FM was invented by prof. John Chowning at CCRMA, not by Yamaha, who later developed it in form of commercial synthesizers.

'scuse me for my penn'th,
but isn't the exciting thing about additive syenthesis, that it opens the way to resynthesis?
i.e. the possibility of taking a sound snd breaking it down, using FFT into it's constituent sines (plus noise). I for one would relish the possibilities this opens for morphing in time between sounds. Seems to me that this is a perfect app for a vsti (plus off line FFT processing). Anyone up for it?

Contrary to what has been said here, FM is NOT additive synthesis. It's a nonlinear technique that may be considered as a tricky way of waveshaping, but it's not related to strictly controlled mixing of harmonics partials, which describes a linear scheme of additive synthesis.

MB,

Sorry to rain on your parade but FM is most definitely additive synthesis. I will clarify (thought this was common knowledge but hey...):

Additive synthesis is the opposite of Subtractive synthesis. Those two designators are not linked to any specific way of creating sounds.

Subtractive synthesis: Shaping a base waveform (albeit complex or just a sine) into the desired sound by REMOVING harmonics (therefor SUBTRACTIVE) by use of filtering.

Additive synthesis: Shaping a sound by ADDING harmonics to the base waveform (therefor ADDITIVE) to create the wanted sound.

FM is NOT filtering but is ADDING harmonic content to create the final sound.

Oh, and before you ask: yes, I do have formal training in synthesis techniques (Rotterdam Conservatory)

Yep, it's also called fourier synthesis. K5000 has something like 64 parts for 1 voice, and I'm not sure if my PC could do this in real-time?

Lowering the number of parts/harmonics could help.

jorgen

On the K5000, each oscillator has 64 partials, and you can choose between the upper and lower harmonics. You have to use more than one oscillator per voice to get more. The DK Synergy only had thirty-two partials, but it certainly seemed quite capable of producing rich material. For me, the limitation of my K5000 isn't that there's not enough partials, it's that they haven't got individual pitch envelopes for each. They DO have amplitude envelopes for each, but they lack the amount of breakpoints required for anything too complex in the time domain. Sure, there are a lot of really convoluted sounds that do some neat stuff over time, but that's due to the formant filters and what not, so that part aint additive. Screw the filters, I want the envelopes ('cos ya see, it just aint tedious enough programming the K5k as it is)!
Scot Solida
Christus And The Cosmonaughts

crimsonwarlock:

quote:

---

FM on it's own is Additive synthesis by all definitions given so far (mine and MB's). If you want to oppose that statement, give a definition of Additive synthesis that excludes FM

---

"elemental sound components (such as sine waves) are added together in time-varying ways. Moore, _Elements of Computer Music_.

"Production of sound by direct summation of component frequencies." Dodge & Jerse, _Computer Music_

Perhaps we can agree that FM is not 'textbook' additive synthesis :-)

nuffink:

quote:

---

Anyone up for it?

---

Stay tuned.

anyone can download SynC Modular and try these
http://digilander.iol.it/luisi/modules/+16tri.zip http://digilander.iol.it/luisi/modules/+32sine.zip
feedback is appreciated

ciao

Thanks epiphanius for the article link, very useful.

malouin,

quote:

---

"elemental sound components (such as sine waves) are added together in time-varying ways. Moore, _Elements of Computer Music_.

---

This still does not state the actual method of "adding". FM uses "elemental sound components such as sine-waves" and add's them together in "time varying ways".

quote:

---

"Production of sound by direct summation of component frequencies." Dodge & Jerse, _Computer Music_

---

Hmmm... I like that one, by this definition I'm indeed wrong.... but by the first definition I'm right. Those two definitions show clearly there is not one "actual say it all" definition.

Let's just agree to disagree