Zebra 3 Public Beta Revision 20399

[Urs]

We'll see where we take things...

The FMOs have a pretty novel approach to antialiasing. Not every waveform lends itself to do that. If we add more waveforms, we either need to come up with an equivalent method, or do something else. And not every combination of modulator and carrier will easily work this way.

[jtsterays]

I find the max decay in the envs fall a bit too fast, can you increase the maximum value? I know about Stretch but I want to save that for Feedback.

[pdxindy]

I find the max decay in the envs fall a bit too fast, can you increase the maximum value? I know about Stretch but I want to save that for Feedback.

You can use Stretch and then multiply to envelope stages by themselves to do the same thing as Feedback (But with more control)

[jtsterays]

I find the max decay in the envs fall a bit too fast, can you increase the maximum value? I know about Stretch but I want to save that for Feedback.

You can use Stretch and then multiply to envelope stages by themselves to do the same thing as Feedback (But with more control)

Oh yea, forgot about that, I do that in Hive. Thanks for reminding me.

[pdxindy]

I found, the examples that were compelling for "more MSEGs" were often examples where multiple MSEGs were combined to achieve dynamic shapes, like blend two envelopes based on velocity or something. And the result of that is, in Zebra 3 you can set up multiple curves in each MSEG and morph or crossfade between them. This gives you dozens of MSEGs for IMHO the most common need for multiple MSEGs.

The morphing MSEG's are such a welcome tool for me. They allow a precision of modulation not available elsewhere. I've not seen much mention of them so I suspect many users haven't discovered them yet.

Even just 2 curves, and morphing between them with velocity or key tracking is such a useful function.

[JoeLowery215]

We'll see where we take things...

The FMOs have a pretty novel approach to antialiasing. Not every waveform lends itself to do that. If we add more waveforms, we either need to come up with an equivalent method, or do something else. And not every combination of modulator and carrier will easily work this way.

Urs is there a reason I cant get a click sound with the FMOs? Would love to get them as punchy as the DX7.

[Ghul]

Is there a detail I overlooked?

The Mappers are bipolar. They trigger when they go from a value below zero to a value above zero.

In the arpeggiator video there's a Mapper that has the first step all the way up and further steps below zero exactly for that purpose. From 2:31

Thanks, that did the trick!
Now it works.

Much appreciated.

Cheers

[misterandy]

Please, please add the option that the carrier in FM could also be a wavetable oscillator. So basically two oscillators can FM eachother.

If that would be the case, reaching for Serum would be obsolete.

Have you tried the Osc FX DeltaX and set index to Sine * wave? Then the modulator is a sine wave and the carrier is whatever wave you create in the oscillator. You can do a lot of cool FM stuff with this.

[trance_lucent]

Awesome work, U-he!

Zebra 3 is a fantastic synth with phenomenal sound quality. The possibilities are endless, and it’s great to see an increased focus on physical modeling.

That said, I was hoping to see more elaborated tools for modeling body/deck resonances.

The Modal module is a great addition, but it’s only part of the equation. One of the key elements behind convincing physically modeled instruments is body or deck resonance with a fixed frequency response - i.e. resonances that do not key-follow. While it’s possible to set key follow to 0 in the Modal module, this then requires carefully designed resonance patterns for different bodies (violin, piano, etc.), and such profiles are not currently shipped with Zebra 3.

In my own experiments, the Comb filter often sounds more convincing than the Modal module when used specifically for resonance modeling. I do miss some of the older Zebra 2 comb behaviors, though. Also, tweaking the comb purely by ear is not very convenient. A plot of its frequency response would make it much easier to tweak the Comb and and match resonance pattern to reference plots of real instruments. I remember requesting something like this 10-15 years ago, and Urs even replied back then that it might be possible to add something like this.

There is also the Resonator (now part of the EQ module), which is useful for shaping a few main low-frequency resonant bands. However, it doesn’t really work for building a complex resonance image, such as a piano soundboard or instrument body.

[Super Piano Hater 64]

MSEGs are unipolar. That's how they are also usable as VCAs.

There are obviously other ways to achieve this (I've done it many times in Hive) but adding a bipolarize function to the rectifier (if not MSEGs themselves) would make certain modulation patterns much easier to work with, especially when a patch has to be adapted to a new environment without substantially changing the sound.

[jrgillam]

I found, the examples that were compelling for "more MSEGs" were often examples where multiple MSEGs were combined to achieve dynamic shapes, like blend two envelopes based on velocity or something. And the result of that is, in Zebra 3 you can set up multiple curves in each MSEG and morph or crossfade between them. This gives you dozens of MSEGs for IMHO the most common need for multiple MSEGs.

In a few days we'll post an example where we use this to create a drum kit within Zebra. This is something that is very hard to achieve without multi samples or a dozen layers, but it is quite obvious and easily achieved in Zebra. Zebra has no layers, but at the same time it can have many layers in multiple degrees of freedom...

Thanks Urs. I will try to get better at using what you give us

Have you considered adding the vibrato LFOs in the pitches section as mod sources like in Hive? I think that's a cool feature for getting a little bonus randomness and motion on parameters other than pitch. (Also, I think this has been mentioned but can we get a delay on the vibrato).

Kind regards.

[Urs]

Have you considered adding the vibrato LFOs in the pitches section as mod sources like in Hive? I think that's a cool feature for getting a little bonus randomness and motion on parameters other than pitch. (Also, I think this has been mentioned but can we get a delay on the vibrato).

Kind regards.

We have smuggled 4 extra LFOs into the ModMath modules.

One of my concerns is, I want to keep things tight, and that includes the number of modulation slots, and modulators. A common pattern in this thread and others is that people pick something from their favourite synth, which is usually something something from Serum. So let me also pick something from Serum: Serum 1 has a fixed list of 33 modulation sources, and 32 modulation matrix slots. While not obvious, I think there's some beauty in that kind of balance.

I'm wondering how that applies to Zebra 3, because obviously, while Zebra 3 also has 32 slots in its ModMatrix, it additionally has 300 or so direct modulation slots in its modules. Plus, contrary to Serum, it has some modulations hard wired, e.g. key follow in filters and envelopes etc. However, only so and so many modules are used at a time, and I would not be surprised if the averaged number of modules used with the average number of direct modulations slots added up to about the same balance as in Serum 1, but with 50 instead of 32, as Zebra 3 has something like approaching 50 modulations sources.

[ffx]

Good morning Urs,

but could you consider to really add two more MSEGs? Since these are required for any kind of custom ENVs, e.g. if you want to use regular bending of attack, decay, release phases. Then 2 MSEGs are quickly in use onl for amp and filter, and then it would be really helpful to still have 4 MSEGs available for all the other modulation scenarios. I think this still would fit nicely into the GUI, too. And you don't have to use it.

Thanks for consideration!

[Urs]

We're thinking about a lot of things.

That includes slope control for the regular envelopes. Not sure how well this can work, as I do not want to sacrifice the envelope's dynamic behaviour.

It includes adding envelope followers that, like Hive's Function Generators, could act as LFOs, envelopes and modifiers.

It includes reusing MSEG or osc curves somewhere else (e.g. within LFOs), or even sampling different MSEG curves from within the ModMath modules. Or simply just giving MSEGs a secondary output.

However, none of these concepts are quickly done, none will happen before initial release, and some concepts are mutually exclusive. We'll probably have to spend a year or two with Zebra 3.0 before we can confidently decide and design on the best possible solution.

[jtsterays]

Multicore is not supported yet, is that why my CPU usage is really high when doing chords? I did use the Diva-esque filter, but in the actual Diva it wasnt that bad.