Tone2 will release Icarus - 3D WaveTable Synthesizer

[OneOfManyPauls]

The JP-8k has an internal high pass filter (like most VA's i believe), to avoid aliasing in the higher registers, so that will already alter the shape of the waveforms, depending on pitch, AFAIC.

I think you mean low pass filter as it's high frequencies that cause issues with aliasing.

This is why i craeted single cycles of the single waveform that in combination with the Supersaw mode could sound like the real Supersaw or at least close to it.

I'm still not getting why a single cycle of a supersaw is useful.

if each stacked saw is running at a different frequency - which it is because they're detuned - then their cycles are different lengths, so a single cycle waveform is just a moment in time with regards to the phasing of those saws, and every time the sampled single waveform cycles, it is going to be cutting short the cycles of the original saws.

Not sure the point re HPF - that is just there to cut out the bottom end so it sits better in the mix and is obviously easy to recreate. HPF'ing a single saw sample and stacking that isn't going to be the same as using a HPF after the stacked raw saws.

To me it sounds close enough to a real Supersaw and teher is really more than enough "movement" there.

it sounds dull to me - which makes sense as your supersaw isn't made up of "full" saws for the above reason.

to be clear, when I talk about "movement" in waveforms - I'm talking about the "chatter" at the margins that you get in analog gear which creates additional frequencies and leads to character. you can see this with an oscilloscope - it doesn't look like much in terms of waveform variation, but with a frequency analyzer you can see where the character comes from. you can't create this by interpolating waveforms.

[Ingonator]

I'm still not getting why a single cycle of a supersaw is useful.

As i tried to explain in several posts the Sawtooth in the Supersaw simply does sound like a normal Sawtooth and teh most simple way to get the right basic sound is using a singlce cycle of the original waveform (you could get the single waveform without detuning from the real thing, JP6K and also Diva).

With the proper waveform and the Supersaw Unison mode in Icarus you could come much closer (or very close) to the real sound than with using a standard Sawtooth.
Of course without that Supersaw Unison mode used it will not really sound like the Supersaw at all. Anyway that Unison mode works with all waveforms so it does not include the original Supersaw waveform by default.

You could simply compare the sound of the Supersaw waveforms i posted with the standard Saw in Icarus (e.g. taht from teh INI menu in the editor) and the difference should be obvious.

[OneOfManyPauls]

added: oh, you've edited your post a bit while I've been typing this out.

A real supersaw is a stack of sawtooth waveforms that have different cycle lengths due to their detuning - you can't just take a single cycle of that and say it represents anything useful.

think of it like this: if there were just 2 saws and they were detuned by an octave, then the lower frequency saw would need twice as long as the higher frequency to complete a cycle. To keep it simple, lets say they are triggered in phase simultaneously rather than being free-running. Now if you were to take a single cycle of that at the high frequency rate, it would only contain half the low frequency cycle, meaning that when played back as a wavetable, the low frequency would be a stunted saw. Now multiply that by 7 different cycle lengths, that aren't all neat doublings and triggered in phase, and the problem should be obvious.

Your 14s sample sounds like a (dull) supersaw in the audio clip you posted because of Icarus's processing - ie stacking and detuning, but it's stacking and detuning a defective source as it's just repeating a single moment in time of a sound that's constantly changing.

of course there's a difference between the init saw and your "supersaw" waveform because your waveform is a snapshot in time of the multiple saws, but the cycles are all out of whack because they all get reset at the frequency that Icarus restarts the cycle.

I'd expect that a raw jp-8000 saw (not supersaw) cycle combined with Icarus's supersaw mode and a HPF to cut out bottom end (and maybe a LPF to reduce any aliasing) would give the best results.

[OneOfManyPauls]

I don't have a jp-8000 saw waveform cycle, but I've done this quick patch with the jd-xa analog saw wavetable:

oscillator 1 - jd-xa analog saw wavetable
oscillator 2 - icarus init saw waveform
oscillator 3 - Ingonator supersaw waveform

the patch has eq, fx and arp set to off - but I have set it with the default trance ARP and delay+reverb if you want to hear it in context.

patch: https://www.dropbox.com/s/ogk6achxkaa6p ... s.fxp?dl=0

wav file preview: https://www.dropbox.com/s/uipv1n8458gbe ... p.wav?dl=0
(right click and open in new tab/copy link address)

the preview is 3 x single notes in order of oscillator followed by 3 x arp+fx in order of oscillator

bear in mind there's going to be a degree of louder=better bias - but these were all done with exactly the same setup, just different oscillator wavetables.

[Ingonator]

I'd expect that a raw jp-8000 saw (not supersaw) cycle combined with Icarus's supersaw mode and a HPF to cut out bottom end (and maybe a LPF to reduce any aliasing) would give the best results.

Based on my comparisons to the JP-8000 oscillator in Diva i still seem to get the best results with the way i already used so far which means combining the Supersaw single cycle wave from the JP-8000 with the Supersaw Unison mode in Icarus. Actually those could sound almost identical. With a blind test it could even be difficult to guess which is which. Sometimes i had to check if i am really playing 2 differnt synths.

Using other Saws with an addional HPF (i used the HPF with the variable slope from 6-48dB) works but is not really closer than what i used before searching for the proper Supersaw waveform.

Of course i also need the LPF because i need this for the patch i am currently doing with this (including Cutoff modulation with the envelope).

At the moment i will keep my patch the way it already was as for me it seems to sound graet that way, also when directly compared to Diva.

The way you proposed could maybe work for certain other patches but if it is closer to the real Supersaw is another story.


UPDATE:
The original JP-8000 oscillators seemed to do aliasing. You could emulate this with the Aliasing Osc Morph modes in Icarus.

[OneOfManyPauls]

At the moment i will keep my patch the way it already was as for me it seems to sound graet that way, also when directly compared to Diva.

not trying to force you to do anything - just find this an interesting topic. I love the flexibility we have with icarus.

[Ingonator]

nvm

[chk071]

The JP-8k has an internal high pass filter (like most VA's i believe), to avoid aliasing in the higher registers, so that will already alter the shape of the waveforms, depending on pitch, AFAIC.

I think you mean low pass filter as it's high frequencies that cause issues with aliasing.

No, i do mean high pass filter. See here: https://www.nada.kth.se/utbildning/gruk ... _10131.pdf

The shape of the oscillators, are found to be actual
sawtooth waveforms, however, there is a high pass filter at the fundamental harmonic of the
waveforms which is “pitch tracked” that is, it follows the frequency of the first harmonic of the
oscillator. To prevent the oscillators from aliasing at higher frequencies, the high pass filter
removes the unwanted harmonics that are folded back into the audible spectrum below the set
frequency.

[OneOfManyPauls]

Got it - apologies and thanks. I was confusing it with ADC anti-aliasing filters, which are low pass:

https://www.maximintegrated.com/en/app- ... mvp/id/928

Anti-Alias Filters
With this background, we now move to anti-aliasing filters. When selecting a filter, the goal is to provide a cutoff frequency that removes unwanted signals from the ADC input or at least attenuates them to the point that they do not adversely affect the circuit. An anti-aliasing filter is a lowpass filter that accomplishes this. How does one select the right filter? The key parameters that need observation are the amount of attenuation (or ripple) in the passband, the desired filter rolloff in the stopband, the steepness in the transition region and the phase relationship of the different frequencies as they pass through the filter.

[OneOfManyPauls]

UPDATE: The original JP-8000 oscillators seemed to do aliasing. You could emulate this with the Aliasing Osc Morph modes in Icarus.

is it me, or does the Icarus aliasing morph mode actually look like bit reduction?

[exmatproton]

UPDATE: The original JP-8000 oscillators seemed to do aliasing. You could emulate this with the Aliasing Osc Morph modes in Icarus.

is it me, or does the Icarus aliasing morph mode actually look like bit reduction?

I had the same feeling about that.

[chk071]

UPDATE: The original JP-8000 oscillators seemed to do aliasing. You could emulate this with the Aliasing Osc Morph modes in Icarus.

is it me, or does the Icarus aliasing morph mode actually look like bit reduction?

Well, that is actually what is happening with aliasing, right? At least a similar effect.

[OneOfManyPauls]

I'm getting out of my depth on it, but it seems low bit rate/frequency does cause some form of aliasing - but I can't see how this form of aliasing is useful in the context of the supersaws as it just leads to them sounding like they're poorly sampled.

As I understood it from the various analysis things I read, it's (fundamental HPF'd) fold-back aliasing beyond the nyquist frequency that is useful in the context of the supersaw.

[Ingonator]

I tried using the Aliasing but then removed it as it did not really fit to the patch i created.
Not everything that is possible in terms of emulating teh real thing must also be done, at least not in every case.

i also started doing small adjustments to the Supersaw waveforms (including doing a mix of 2 waveforms and small edits in the free draw waveform editor and the Spectrum editor) i got to get a kind of my "own" Supersaw wave. This doe not mean that the original waveform sounded bad but i want to get it closer to my liking (and the changes ate not really big yet).
To avoid misunderstansdings (which seems to sometimes happen here...) with "Supersaw waveform" i mean a single Sawtooth from a Supersaw (in the real thing and some emulations this is done by setting the "Mix" and "Width" knobs to minimum) that has to be used in comination with the Supersaw Unison mode in Icarus to sound properly (this creates 7 detuned Saws like in the real thing).

[Ingonator]

Tone2 just released "Electronix", the first official soundset for Icarus:
https://tone2.com/html/electronix.html
https://tone2.com/html/shop.html

This includes 213 patches done by Markus Krause and Troels Nygaard from Tone2.

I got te soundset too (a free copy) and IMO it is a nice soundset but of course this is a matter of taste.