KVR Audio

kmonkey · Post by **kmonkey** » Thu Oct 28, 2010 9:05 pm

Will some of you guys stop rambling nonsense. Urs asked about JP8000 supersaw. He is trying to do it in Zebra. No, he (most probably) won't make new Dance synth with just super sissy saw and no this is not another Viral marketing thread. If you want to get deeper then read that nOes thread. For christ sake stop talking about how it is different then VirusTI or any other synth be it Roland or not. I guess we'll have to wait when someone post actual sample from actual hardware

JP8000 - not something other..

Urs · Post by **Urs** » Thu Oct 28, 2010 9:07 pm

Kriminal wrote:
Urs wrote:
Kriminal wrote:doesnt sound thick/fat enough to me....
To be a JP8000? Or to be a phat Supersaw?

Note that this is all about the JP8000's Supersaw...

both, i always imagined it to be more detuned...but im only going on the sound clips i have heard, not 'real' experience (of a jp8000)

One caould detune of course it to taste. The "technical magic" of the JP8000 is however best demonstrated at a moderate detune setting, where the individual saws are still somewhat related to each other (e.g. other supersaw implementation suffer from phasing phasing artifacts etc.). The, hmmm, let's say "smooth" operation of the JP8000 Supersaw is IMHO achieved by a clever sawtooth algorithm, clever phase relation of each sawtooth component and a afaik trivial detuning scheme.

I believe that once these technical issues are solved good enough to deceive the listener (e.g. by no-one noticing that an a/b comparison actually consists of two synths, the emulation and the original), then getting stronger detune (or no detune at all) right is a trivial issue. It occurrs to me however that the deception is way easier than I thought, as my algorithm is far from perfect yet (but maybe good enough, as asked for in the thread title)

Urs

aciddose · Post by **aciddose** » Thu Oct 28, 2010 9:20 pm

the distance between each step follows a power law.

distribution = 1/2 is nice.

for fraction 0 to 1 in n steps
result[n] = pow(12throot2, detune depth * pow(fraction, distribution))

try the unison saw in xhip, for example.

i've never been able to remember how i came up with this, but i believe it was while reading some properties of various trig functions i discovered that a power of a power has blah relationship to blah in the case of blah - something about maximizing lowest common denominator. i'm not sure if what i came up with is the perfect function, but it sure works well.

you'll also want to center the output so you don't detune your unison.

result[] -= (average(result[]) - 1) / 2;

i think.

- however, that isn't what the jp8~ uses. it uses some super simple lookup based method which gives it a very distinct timbre, but non-optimal lcd property.

UltraJv · Post by **UltraJv** » Thu Oct 28, 2010 9:24 pm

Believers in the JP Supersaw mythology say its due to the "feedback oscillator" I interpret that as chorus/phasing effect. Maybe this will help :

http://homepage.mac.com/synth_seal/html/ds_jp8080.html

antto · Post by **antto** » Thu Oct 28, 2010 9:28 pm

wtf is that "feedback oscillator" anyway.. smells like marketing crap or something..

can anyone describe it?

UltraJv · Post by **UltraJv** » Thu Oct 28, 2010 9:36 pm

antto wrote:wtf is that "feedback oscillator" anyway.. smells like marketing crap or something..

can anyone describe it?

"As fourth oscillator we have Feedback Osc. With control 1 it is possible to vary the sound of harmonics, while with control 2 we adjust the amount of feedback to adjust the level of harmonics."

Urs · Post by **Urs** » Thu Oct 28, 2010 9:37 pm

UltraJv wrote:Believers in the Supersaw mythology say its due to the "feedback oscillator" I interpet that as chorus/phasing effect.

The feedback oscillator is simply a sawtooth oscillator followed by a comb filter with variable (positive) feedback and a tuning range of 0-12 semitones above the oscillator tune. That's a delay of a maximum length of about 1/50th second.

It is very unlikely that the Supersaw in the JP is done by chorusing (which is a modulated comb filter effect). Chorusing would mean that the peaks could "wander" in two directions from the main sawtooth, or each pair of them would "meet" at the same phase from time to time to sawap their positions.

As the JP has a "master saw" and 6 "sub saws" that can be mixed at different levels, it is easy to follow the motion of the "sub saws" relative to the "master saw". I havn't yet made out a regular scheme between pairs or triples, so I'm quite sure they're using free running sawtooths, not any modulated phase trickery or a delay based approach.

(Add to that, the Feedback oscillator only works in mono mode, so I guess the delay for one voice eats up the memory of all voices, hence there wouldn't be ennough memory or computing power for a 6-voice chorus)

Urs

UltraJv · Post by **UltraJv** » Thu Oct 28, 2010 9:43 pm

There have been many threads on the JP Supersaw, it would be cool to see it finally nailed and I feel URS is up to the challenge

parry316raver · Post by **parry316raver** » Thu Oct 28, 2010 9:45 pm

Urs · Post by **Urs** » Thu Oct 28, 2010 10:07 pm

parry316raver wrote:I think Mr Urs SuperSaw is very good.

Thanks & good stuff!

The interesting point of the JP is though that it's got something different to it, which IMHO can't be achieved by normal sawtooth oscillators.

The background for this whole quest is, I spent months on filters that distort nicely (e.g. *not* like a Memorymoog, but maybe like a Minimoog). Normal supersaw oscillators such as the ones in Zebra however don't play nicely with distortion because of inharmonic buzz you get from intermodulation. Thus I found that common supersaw isn't onyl cheap but also cheap sounding in comparison to real unison voices that have a filter for each one or two sawtooth oscillators.

However, the JP8000 is different because the intermodulation distortion is pretty much as tame as only, say, 2 sawtooths and thus by all means bearable with screaming filters. Looking closer I found various reasons for this, and hope to come up with a model that works.

I will however never understand why Roland hasn't spent a bit more juicy on the JP8000's filter. They could have.

antto · Post by **antto** » Thu Oct 28, 2010 10:13 pm

erm, so the supersaw on that JP8000..
each saw goes into a VCF (like unison)
or
all the saws go to the VCF?

Urs · Post by **Urs** » Thu Oct 28, 2010 10:19 pm

Cyforce wrote: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

I just did an a/b listening test between Virus TI and JP8000 and I must say that while the Virus is admittedly more brilliant and aggressive sounding, the JP8000 has way less aliasing, in a magnitude of db. Thus I find the JP8000 to sound better on raw oscillator output, albeit maybe less spectacular. The filters speak a whole different story though.

Urs · Post by **Urs** » Thu Oct 28, 2010 10:23 pm

antto wrote:erm, so the supersaw on that JP8000..
each saw goes into a VCF (like unison)
or
all the saws go to the VCF?

The Supersaw is a mode of the OSC 1 that produces a stack of 7 sawtooth-like waveforms which can be detuned. These are mixed with the ouput of OSC2 before they go into the filter.

It's not unison as in stacked voices with a filter each.

parry316raver · Post by **parry316raver** » Thu Oct 28, 2010 10:28 pm

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.

everybody and their dog uses SuperSaws in UK Hardcore etc....

Maybe its time to come up with a new form of SuperSaw.

MegaSAW

zerocrossing · Post by **zerocrossing** » Thu Oct 28, 2010 10:31 pm

Sendy wrote:Why not go one better and make a 'Metasaw', a saw so powerful that it turns all your intstruments, the overall waveform of your mix, and your face into a fractally-layered, infinitely regressed saw wave, ultimately causing the destruction of the fabric of space time, because of the infinitely sharp edges of the saw tearing away at it and creating singularities. (Which could perhaps create whole new baby universes where the laws of physics are dictated by saw waves, and all dimensions are folded into a saw shape).

I'd buy that for a dollar

How do you know we're not the result of an event exactly like this?

Supersaw!