Zebra 3 Public Beta Revision 20399

[machinesworking]

After listening to these presets, I think we need to change the One Synth Challenge to a One Oscillator Challenge. This is incredible.

Yeah I made a patch today of one osc, filter and a reverb, crushed
my M2 studio cpu. Hopefully it does get optimized, but to your point, sounded huge with just those thee elements

[pierb]

A small improvement I would like to see are the bipolar modulators actually showing up as bipolar in the target param. Right now you can see the modulation amount but the range always shows up a unipolar.

[digitalmy]

They definitely don't sound "closer" to any originals than Diva's. But I did certainly tweak a lot of them to sound their best possible. That's Cascade, Ladder, Impossible (which has no reference) and Expanse.

Yellow is indeed based on the original Wasp designs, but simplified in Zebra 3. We have loads of hardware references (originals and clones), we even build our own testing hardware to get proper measurements. But again, we decided to simplify to make it sound good and save CPU. (No joke the full fledged model uses even more CPU... but that's also because it needs to be 16x oversampled to avoid aliasing)

SVF is a plain model of the Oberheim SEM, very similar to some really good open source synths, but with an iterative solver for correct behaviour.

Vanilla is a simplified ladder with simplified but very efficient non-linearities.

SVF and Vanilla should be used in high poly patches whenever Linear does not cut it. Or the revamped Zebra 2 filters (Excite, Mid Drive, Old Drive), which were added for nostalgic reasons but improved for audio rate modulation.

The "analogue" ones - Yellow, Impossible, Expanse, Ladder, Cascade - should be used in solo sounds or with low polyphony. Whenever character really matters.

Thanks for those insights. Is there also a real-world counterpart of the „expanse“ filter?

[reppinghades]

No trigger option for MSEG?

[wintoid]

I have no issues with the idea of a bipolar-ish Ratio knob.

However: it would be very handy while editing the Ratio parameter if Zebra 3's data display also showed the Mod:Carrier and actual ratio values, in addition to the 25.00–25.00 number.

I'd be very happy with that.

[Iores Pyadzee]

Hi again, just wanted to ask two things. One is that when I tag the presets from scale 1 to 8 to my liking and when the original release will be available, will them be gone? I remember such tagging files exist in somewhere for Logic Pro Alchemy synth but do not know whether this is the case for Zebra3 too. The other question is, when dragging parametric eq bands, it always jumps between very weird dB values with decimals. Zebra3 is still in beta phase, I wonder whether full version will be available with integer values or will it be implemented maybe on version 3.1 or will it be not considered as an issue at all? I am asking it because I am a math student and always has been very precise and symmetric person, I always automate parameters in DAWs to integer and type in values even if it takes much time. I can’t stand the odd integers(except the ones ends with 5) or decimals not ending with 5 or 0. When I use built in eq in Zebra3, I always feel bad because of that and I know people may not be like me but this is not the case for other U he synths and I wonder whether Zebra3 snapping values will become at least more symmetrical as well.

[the_eraser]

Is there a way to disable the automatic mouse-wheel-scrolling-changes-values?
This makes impossible Z3 to navigate with the MBP trackpad (so many accidental changes!).

thanks

[Urs]

please gave us much more quality presets so that as zebra 2 legacy would be more motivated to buy zebra 3

Thank you for your kind words!

I that vein, let me add a few thoughts. Not directly related to what you said, but in general.

So there's been a bit of back and forth about whether Z3 should have been compatible to Z2. There are a whole lot of reasons it isn't, but if we set those aside and only look at the option of "should Zebra 3 be backwards compatible to Zebra 2?", and if we even neglect the thing where this would typically mean the end of life of Zebra 2, here's what I think:

Zebra 3 is 20 years of progress over Zebra 2, if not in all modules then certainly in oscillators, filters, FM, MSEGs.

I do think that filling these NEW modules with the same sounds that happened in the old modules would be, hmmm, disappointing.

So the disadvantage is, at the beginning, Zebra 3 will come with, dunno, 500? 800? presets. Maybe 1000 if we're lucky.

On the other hand, we have something that has not happened often in this industry: Zebra 3 starts out with maybe 40000 - 50000 users due to the upgrade scheme from Z2 prior to Zebra Legacy. This is a very huge number.

This is a major incentive for 3rd party sound designers. There is literally no competition yet, for a pretty big target audience. This is very different from upgrades of popular synths that are backward compatible, where the market is already watered down with hundreds of sound banks for the old version to compete with.

So even if we only have a few hundred presets, they just need to be really good. If we set a high bar, I'm convinced professional sound designers will come through, and soon.

And, if that works out, we'll soon have a lot of presets, a lot of examples, and with that, a lot of fresh sounds, that utilise the progress made instead of merely using all that for something that Zebra 2 already did and still does for years to come.

[Urs]

Could you guys replace the curve knob in Pitches glide section to a draggable curve, something like this in Serum? Would be a bit more intuitive.

Ours works/looks a bit different, but sure, we can eventually do a visual representation. Just needs to be rendered to an image animation.

[Urs]

How to import a .uhm?

It's going to be merely the other way round. Curves made in Zebra are copied as .uhm so the can be used in Hive later on. But that functionality is not yet implemented.

[Urs]

Both Folder and Dist should have an output knob, gain matching is objectively good practice, even in a sound design context.

We're reluctant to add volume knobs to filters/distorion because there is already an exceptional load of gain staging going on.

We hope that it is sufficient to occasionally add a Utility module if more control over gain is required between two subsequent modules.

[Urs]

How do you import a .wav wavetable?

It was so buggy on larger files, we had to restrict the .wav size to prevent most crashes and deadlocks. We had incidences where even a small file with 20 cycles would be analysed for an hour or two, or ran out of memory.

I will rewrite the import code once I find the time, hopefully early January.

Until then we recommend to stick to single cycles when importing .wav.

[Urs]

Congrats to Urs and the U-He team on the Zebra3 roll out. It sounds *very* nice.

Apologies if this has already been suggested but, would it be unreasonable to include an option to float the oscillator edit windows?

Some degree of back and forth page flipping would be significantly reduced if extended screen space could be utilized to display the more critical and often accessed programming modules.

Opening external Windows has been on our wish list forever. Maybe next year we can tackle it, and then we'll be able to do floating editors.

[Urs]

However: it would be very handy while editing the Ratio parameter if Zebra 3's data display also showed the Mod:Carrier and actual ratio values, in addition to the 25.00–25.00 number.

Well, yeah... we'd have to find a clever way to represent it for the fractional values in-between.

[Urs]

Thanks for those insights. Is there also a real-world counterpart of the „expanse“ filter?

Kind of. The filter model is based on the CEM 3320 filter chip that we also modelled for Repro, but without the code required to model a particular specimen. The different responses are created by mixing the 4 poles and the dry signal, much like how it was done in the Matrix 12 and Xpander. So this filter uses a "shortcut" to create the different responses, unlike all other filters which change the inner model to do so.