Supersaw!

[xoxos]

only read first and last pages.. my 'horizon' is one of probably dozens of synths that do high unison count (max. 32 oscs.. spread panoramically, if so desired). how 'noisy' it sounds depends on the amount of detune. with low detune it sounds like a smoothly modulating oscillator, somewhat like solo bowed string. there are nice, silky textures that usually seem to require at least 8-10 or more voices.

[A.M. Gold]

some good points. chords are important with trance etc..

I actually don't do trance, my stuff is more like melodic dance backing instrumentals, but the same principle applies. I like that sort of "symphonic" synth sound you can get from supersaw chords, but have always been frustrated in trying to mix pieces with heavy supersaw based chord parts. Everything's fine until you try to layer more synth sounds on top of that massive wall of detuned osc's, then things get tricky real fast. Kudos to anyone who is good at doing that and pulling it off---I know they are out there, but I have yet to master it myself.

[parry316raver]

I dont always use 30 saw wave super-saws though.

If i'm doing a string patch i will use less.

they do clutter up the mix space.


i wonder what synth Urs used for the mp3 example earlier ?

[blueghost]

Interesting

I JUST DOUBLED THE NUMBER to 26 and filtered out the freq band that sounded like noise

Where not in Kansas anymore Toto

[A.M. Gold]

What does not-Kansas sound like?

[Urs]

i wonder what synth Urs used for the mp3 example earlier ?

Hehe, I thought that was obvious by now

You hear two synths:

Every second note in turn is a "real" JP8080, every other note is Zebra with an experimental Sawstack module and various filter/eq settings.

http://www.u-he.com/music/supersaw.wav

#---

At the moment I don't really care about what other synths sound like. It occurrs to me that no softsynth available mimics the characteristics of the JP8000's OSC1 section close enough to satisfy those looking for these features (if one has to play some notes an octave deeper or adjust some eq depending on which notes are played to match the JP then that's fail)

Just like the whole analog vs. digital debate, the sound of the JP8000 is often glorified. In my opinion it is a very limited machine with various close-to-lethal design flaws.

If anything then I want to take the glorious bits and slap them onto algorithms that surpass the sonic quality of the JP. I'm not after the same sound and features, I'm after a better sound and a superset of its features. It would be stupid to mimic the aliasing, the modulation routing and the weak filter of the JP. Even the two knobs CONTROL1 and CONTROL2 of OSC1 that form the base of the JP8000's fame are mostly *fail* from a nowadays perspective.

An example for that is the "Wrap" thing that was suggested to me by Sendy and which I recently implemented. It surpasses the JP's wrapping triangle by so many lengths that you won't look back.

[olikana]

"supersaw" has imo 2 meanings , although related.
1)the oscillator present in the Jp8000 and jp8080 series (a 7 saw stack with detune control).
2)a term ,adopted from the JP800 Supersaw oscillator, which is used to describe the sound of multiple stacked detuned saws.

judging by the posts on this forum over the past 10 years i would say definition n.2 is the predominant one here.

if this thread intended to discuss specifically the Supersaw OSCILLATOR....then it should have been titled "Supersaw Oscillator".
with the generic title "supersaw" it was only to be expected many irrelevant posts would turn up....irrelevant to what the OP wanted to discuss that is.

I have no JP, but a Virus TI and i must say, the Hypersaw(the better supersaw^^) sounds better.
But it belongs on the example also which detuning amount you use etc

the hypersaw stacks 9 saws. the supersaw oscillator only 7. and i think pretty much that's the main difference.
and yes URS example is not detuned enough into the so called "sweet spot"...

i actually though the first Supersaw sound came from a juno 60
"famous Dominator Hoover sound"
o wait there, is a hoover a supersaw though. maybe not as it uses PWM
i remember back in 1991 i was listening to rave music (hardcore techno )
which used Detuned saws.

the first type of supersaw-like sound i can find is by John Foxx in 1980

and i guess he just used a monosynth and some chorus/reverb.
on a tipical analog monosynth with 3 oscillators ..a detuned stack of 4 saws can be mimicked. in practical terms :
osc1 pwm pulse (which sounds like 2 detuned saws) +osc2 saw +osc3 saw.
lets say just the barebone of a supersaw sound.

as u say though very similar sounds to the roland supersaw were achieved towards the end of the 80s and early/mid 90s(for example the leads in italo, eurodance and rave music) thanks to the introduction of 3x unison in some midi analogs (mainly roland synths) which allowed up to 9 saws.
but until roland's "supersaw " came along there was no catchy name or widely recognized term to describe those detuned saws. they were referred to in so many different ways....many americans for example used to call those detuned saws "eurotrash leads"

it is safe to say the sound of the roland supersaw wasn't really anything new or unheard ...
but the supersaw oscillator did allow quite a few more saws to be stacked than it was ever possible before. i think in so called dual mode the jp8000 allowed 14 saws (x2 supersaw oscillators).
from then on supersaw sounds were continued being made also by other popular va's synths like the VirusC or the WaldorfQ which ,although having no supersaw oscillator whatsoever, allowed x 8 unison and over 32 voices of poliphony ...3osc x 8 = 24 saws.
then softsynths took over which allowed almost an infinite number of detuned saws to be stacked...but past 30 saws like someone allready mentioned the sound barely changes and if anything it gets deteriorated so there's no much evolution for the supersaw left ahead....not in terms of numbers anyway...just in terms of quality if that's even possible.

[elxsound]

i wonder what synth Urs used for the mp3 example earlier ?

Hehe, I thought that was obvious by now

You hear two synths:

Every second note in turn is a "real" JP8080, every other note is Zebra with an experimental Sawstack module and various filter/eq settings.

http://www.u-he.com/music/supersaw.wav

#---

At the moment I don't really care about what other synths sound like. It occurrs to me that no softsynth available mimics the characteristics of the JP8000's OSC1 section close enough to satisfy those looking for these features (if one has to play some notes an octave deeper or adjust some eq depending on which notes are played to match the JP then that's fail)

Just like the whole analog vs. digital debate, the sound of the JP8000 is often glorified. In my opinion it is a very limited machine with various close-to-lethal design flaws.

If anything then I want to take the glorious bits and slap them onto algorithms that surpass the sonic quality of the JP. I'm not after the same sound and features, I'm after a better sound and a superset of its features. It would be stupid to mimic the aliasing, the modulation routing and the weak filter of the JP. Even the two knobs CONTROL1 and CONTROL2 of OSC1 that form the base of the JP8000's fame are mostly *fail* from a nowadays perspective.

An example for that is the "Wrap" thing that was suggested to me by Sendy and which I recently implemented. It surpasses the JP's wrapping triangle by so many lengths that you won't look back.

I'm not familiar with the debate or the synths (forgive me, I grew up as a drummer) but I'm always happy to hear of new (possible) additions to Zebra

[parry316raver]

great write up olikana



so , have we reached the final stage of Supersaws now then ?

what is left to create with them. Soundwise, not Musically.


best

Rob


have we now heard every single Supersaw ever created ?

[Kaine]

the hypersaw stacks 9 saws. the supersaw oscillator only 7. and i think pretty much that's the main difference.

Hmmm.

To do list: item 87.

Build synth that takes it up to Eleven.

We shall call it a SpinalSaw.

[bwwd]

I think Urs wants to have JP-8k supersaw behaviour mimicked for zebra and is not interested with other synths/other supersaws ,and its good approach,i cant wait to hear some melodies ,single notes dont tell much and we dont know how different settings were from octave to octave on his emulation.

[aciddose]

...but past 30 saws like someone allready mentioned the sound barely changes and if anything it gets deteriorated so there's no much evolution for the supersaw left ahead....not in terms of numbers anyway...just in terms of quality if that's even possible.

that's not actually true. our perception of harmonic density is exponential just like you'd expect. as long as the detuning distribution follows a power law, the difference in timbre between 32 and 64 is about the same as 1 and 2. 64 and 128, 128 and 256. by which point there is so much noise is becomes difficult to distinguish the difference from white noise - and in fact statistically the result _is_ guassian white noise.

if the distribution or it's delta doesn't follow a power law the result is that you get more and more metallic timbre with higher unisons due to the statistic being more gaussian for whether specific harmonics will line up or not as you add more unison layers.

[bmrzycki]

Build synth that takes it up to Eleven.

[Urs]

3x unison in some midi analogs (mainly roland synths) which allowed up to 9 saws.

Which ones would that be? I don't seem to remember any analogue Roland synth with 3 oscillators and 3-part unison.

until roland's "supersaw " came along there was no catchy name or widely recognized term to describe those detuned saws. they were referred to in so many different ways....

It was widely known as unison sound.

[brok landers]

until roland's "supersaw " came along there was no catchy name or widely recognized term to describe those detuned saws. they were referred to in so many different ways....

there wasn't any supersaw before the one of roland. everything before was unison, something entirely different, technically and soundwise. true unison stacks complete voices, not only osc's. therefore, if unison is 2x, you end up with 2x osc count you used without unison, 2x filter, 2 times lfo's, 2x times env's, etc... it sounds completeley different if you stack osc's, then feed them into a filter (if the filter has any feedback saturation), than if you stack voices, where all the individual voices have their own seperate signal path. that's actually why most analog synths that featured unison were monophonic then.