favorite supersaw synth?

[D-Fusion]

SuperSaw, HyperSaw, gee, I wonder what's next? It's absurd...

This one also have Supersquare and super sine + supertriangle

This is alraedy possible with e.g. the "Übermod" feature in Waldorf Nave (could be used with Saw, Triangle, Square, variable Pulse and/or PWM and also with Noise), some of the the Ultrasaw modes in Electra 2.1 (e.g. "Hypersquare", "6x Pulse", "Hypernoise") and also with DUNE 2 (actually DUNE 2 could use the "Supersaw" feature with all waveforms including wavetables).

Anyway the terms "Supersaw" and "Hypersaw" seem to be specific implementations by Roland (e.g. JP-8000) and Access Music (e.g. Virus TI) and it would be correct to use those terms for those or emulations of them (for example Electra 2.1 contaisn emulations of both). The actual Hypersaw and Supersaw have a specific sound to them and are not just the same word for the same thing (nd teh actual Spersaw is limited to using up to 7 detuned Saws while Hypersaw is available with different amounts from 2 up to 9).

For other implementations taht are no specific emulation a term like e.g. "Multisaw" or "Ultrasaw" (or "Überwave" like in Nave) would be maybe better or more precise (and is already used in some plugins).

FWIW i also do not think that a typical Unison is a proper replacement for the actual Supersaw and Hypersaw, both in terms of sound and CPU use (the dedicated Spersaw and HYpersaw usually do not have a higher CPU use than other standard waveforms, opposing to an actual Unison).



Just a few thoughts...



Ingo

Didn't know that it was Überwaves in Nave.
How are they compared to the real deal (Supersaw)?

[fluffy_little_something]

Hypernoise eh? I don't need a synth to make that

Yeah, I know those terms mean something rather specific. I was just wondering what's next and whether whatever it will be still makes any sense, except maybe in certain freakish genres...

I guess the implementation i nDUNE 2 is alraedy one of the next steps:
- works with all waveforms including basic ones, wavetables and FM mode
- Up to 32 detuned waveforms (64 when using 2 oscillators)
- 9 tuning modes for spreading those in different ways
- Amount control which controls the amplidue ratio of the central waveform and detuned ones (comparable to a similar feature in teh Roland Supersaw)
- Stereo spread for the detuned waveforms (not available in the original Hypersaw and Supersaw)

As in DUNE 2 each Unison voice could act as full layer (with all parameters available) you could actually create up to 32 x 2 x 8 = 512 detuined waveform with a single patch (and a single voice, more when played polyphonic).

Personally i do not think that having more than 32 (or 64 with 2 oscillators) detuned waveforms makes much sense so what i mentioned above seems to be an "overkill" without much practical use.

Yes, overkill is the right word. The more waves there are, the less a good filter matters. Even on something as basic as Sylenth (basic relative to Dune 2, that is) the nice sound of the filter goes away as one increases the number of waves beyond 3 or 4.

[chk071]

The more waves there are, the less a good filter matters.

Where is the correlation?

[Ingonator]

Hypernoise eh? I don't need a synth to make that

Yeah, I know those terms mean something rather specific. I was just wondering what's next and whether whatever it will be still makes any sense, except maybe in certain freakish genres...

I guess the implementation i nDUNE 2 is alraedy one of the next steps:
- works with all waveforms including basic ones, wavetables and FM mode
- Up to 32 detuned waveforms (64 when using 2 oscillators)
- 9 tuning modes for spreading those in different ways
- Amount control which controls the amplidue ratio of the central waveform and detuned ones (comparable to a similar feature in teh Roland Supersaw)
- Stereo spread for the detuned waveforms (not available in the original Hypersaw and Supersaw)

As in DUNE 2 each Unison voice could act as full layer (with all parameters available) you could actually create up to 32 x 2 x 8 = 512 detuined waveform with a single patch (and a single voice, more when played polyphonic).

Personally i do not think that having more than 32 (or 64 with 2 oscillators) detuned waveforms makes much sense so what i mentioned above seems to be an "overkill" without much practical use.

Yes, overkill is the right word. The more waves there are, the less a good filter matters. Even on something as basic as Sylenth (basic relative to Dune 2, that is) the nice sound of the filter goes away as one increases the number of waves beyond 3 or 4.

I guess in DUNE 2 when using many detuned waveforms the included stereo spread feature could be quite useful. Also those 9 detune modes (which allow a different "distribution" of teh detuned waveforms) seem to make sense depending on hw many waveforms you use.

[fluffy_little_something]

The more waves there are, the less a good filter matters.

Where is the correlation?

There is one to my ears. With two waves and slight detuning I can still hear that fine pleasant sizzling character. With more waves, the whole sound as such becomes smeared somehow.
It's a bit like the difference between a violin and a cello playing the same, and a huge string section of an orchestra where the individual instruments kind of fuse together.

[mystran]

The more waves there are, the less a good filter matters.

Where is the correlation?

Much of the character of a filter typically comes from non-linear interactions with the input waveform. With more oscillators mixed to the filter input, the more noisy the waveform tends to look and as a result the delicate interaction between the waveform and the filter becomes more or less random as well.

Alternative view is to consider distortion in a filter as IMD products between the different input harmonics (and sometimes the filter resonance). As you increase the number of distinct harmonic series in the input, the IMD products become increasingly complex and it doesn't take very long before the result could just as well be noise.

This is also why "full" unison sounds quite different with a non-linear filters: if you duplicate the whole audio path (or at least including filters) many times, then each individual only sees a small number of oscillators and it's character is preserved much better as you don't get a similar explosion of IMD product. Personally I highly prefer full unison over just mixing more oscillators (which in comparison sounds very flat and lifeless to my ears), but it arguably tends to cost a lot more CPU.

[mystran]

.. noob hit reply instead of edit ..

[chk071]

Alright, thanks for explaining. Never noticed that.

[Ingonator]

Didn't know that it was Überwaves in Nave.
How are they compared to the real deal (Supersaw)?

The "Überwave" only works with the VA oscillator, not the 2 wavetable oscillators.

You could use the faeture with all waveforms in the VA oscillator like Saw, triangle, Square, Pulse, PWM (when the Pulse is modulated) and also noise.

In the VA oscillator (just called "Oscillator" while the wavetable Oscs are called "Wave 1 and "Wave 2") you could switch on "Überwave", choose a Saw waveform, "Density" at 7 (= 7 detuned waveforms) and Spread at a higher amount it could actually sound very close to a Roland Supersaw.

The global Unsion (at the "Control" tab) works with all oscillators at once and could add additional detuning.

[D-Fusion]

Didn't know that it was Überwaves in Nave.
How are they compared to the real deal (Supersaw)?

The "Überwave" only works with the VA oscillator, not the 2 wavetable oscillators.

You could use the faeture with all waveforms in the VA oscillator like Saw, triangle, Square, Pulse, PWM (when the Pulse is modulated) and also noise.

In the VA oscillator (just called "Oscillator" while the wavetable Oscs are called "Wave 1 and "Wave 2") you could switch on "Überwave", choose a Saw waveform, "Density" at 7 (= 7 detuned waveforms) and Spread at a higher amount it could actually sound very close to a Roland Supersaw.

The global Unsion (at the "Control" tab) works with all oscillators at once and could add additional detuning.

Cool
Guess i have to test out Nave soon then.
I didn't see a need for it earlier since i have Largo and thought it would be some overlap.

[Ingonator]

Cool
Guess i have to test out Nave soon then.
I didn't see a need for it earlier since i have Largo and thought it would be some overlap.

actually Nave and Largo are quite different. Besides the "Übermod" feature Nave could use custom wavetables (either done with teh 3D editor or by importing a WAV file as a wavetable), has a text to specch feature (which creates a new wavetable), has different filters etc.

The "Spectrum" feature in teh wavetable Oscs could be used to change the harmonic content of the current waveform (or the whole table when this is scanned trough) and the Spectrum knob could be modulated too.
Scanning throug ha table while the Spectrum is modulated too could lead to some crazy results.


Finally also PPG Wave 3.V is quite different in terms of features and basic sound so none of the 3 major Waldorf plugins actually is a full replacement for the others.
Besides other features PPG Wave 3.V includes 100 wavetables not included in other synths, could do sample playback, includes the whiole Waveterm sample library, has an emulation of the analog filter in the PPG Wave 2.x hardware synths and some really nice effects /that do not seem to be similar to those in Largo and Nave).
PPG 3.V is up to 8 times multitimbral which could do some complex layering and/or splits. For each part you could set e.g. the midi channel, audio output, panning, volume, key zone etc.
Of course you could also use up to 8 different tracks wit ha single instance of the plugin.
When using 8 patches with different samples also doing a patch with multisamples is possible (when adjusting the key zones for the parts). A multi setup could be saved as a fxb bank file.


Largo seems to be quite verstaile and the best for "typical" VA stuff (compared to Nave and PPG 3.V) as it could do e.g. Osc Sync, FM (in all 3 Oscs you could use Oscs 1, Osc2 and Osc3 as FM sources) and other things not possible with those other synths. Largo is also the only Waldorf softsynth with a dual multimode filtrer, 2 Sub-Oscs, Notch filters and Comb filters. Larg has an Unison with up to 6 voices and Stereo spread which in that way is not included with the 2 other synths (PPG 3.V has teh QUAD mode as an Unison and the Basis knob for stereo spread).

[dioxine]

self-maded supersaw "synth" made with max4live (actually MaxMSP 7 and ~gen) using this tutorial
and with some additions.
this is simple supersaw generator with only note+velocity reaction. no envelopes, filters, modulation matrix and so on now i just love make all synth sounds from the real beginning

but talking about existing and real synths:
on the first place i set Xfer Records - Serum,
second - Urs Heckman's - Hive
third - Synapse Audio - Dune2

[Ah_Dziz]

zebra if I want it clean. Synth 1 if I want aliasing.

[Ingonator]

zebra if I want it clean. Synth 1 if I want aliasing.

Synth 1 seems to have an Unison and no dedicated Supersaw/Hypersaw oscillator. Different thing IMO.

[Mutant]

zebra if I want it clean. Synth 1 if I want aliasing.

Synth 1 seems to have an Unison and no dedicated Supersaw/Hypersaw oscillator. Different thing IMO.

Use a normal saw oscillator and move the detune knob, more saws (8) will appear.
And OFC you are right that it is not a Roland Supersaw or Access Hypersaw emulation, but still it is what we understand as "supersaw".