Zebra 3 Public Beta Revision 20399

[Urs]

I still can't get over the "where did you get your physical modelling code" question. This thread is like the mélange; addictive glimpses of other worlds, too much is fatal. A few more weird responses and I will have visions of Urs leading armies against the Imperium.

We should start a thread with bogus questions and feature requests... "how does a mapper help me find my way to the top of the charts?"

[ThoughtExperiment]

If you really have 'scoured the entire internet' then you must have spotted the ton of videos for Zebralette 3 which explain how to do exactly that...

I'd be happy if you share a links to the ton of videos where we're not just spinning a waveform in a vacuum, but recreate a real complex instruments

I would, but I'm really not interested in recreating real instruments. Why bother when you can load up a sample or record yourself playing a real instrument?

I'm interested in creating instruments that don't/maybe can't exist. That's exciting.

But you do you

[Synthosynthony]

If you really have 'scoured the entire internet' then you must have spotted the ton of videos for Zebralette 3 which explain how to do exactly that...

I'd be happy if you share a links to the ton of videos where we're not just spinning a waveform in a vacuum, but recreate a real complex instruments

I would, but I'm really not interested in recreating real instruments. Why bother when you can load up a sample or record yourself playing a real instrument?

I'm interested in creating instruments that don't/maybe can't exist. That's exciting.

But you do you

so you recommended 'ton of videos' to me, even though you don't even know if they exist. excellent:)

[Urs]

I think part of the disconnect here is expectations. Zebra isn’t designed primarily as a “realistic instrument recreation” synth, and that’s not the problem we're solving with it.

Its strengths are in abstraction, flexibility, and exploring sound structures rather than imitating existing acoustic or electro-mechanical instruments. The spline/vector approach isn’t meant to map directly to “draw this shape → get a violin,” but to give you a very direct way of shaping spectra, motion, and modulation relationships that would be much harder to express in a fixed architecture.

If the main goal is realism or instrument emulation, there are other tools that are a better fit. Zebra is more about creating sounds that don’t already exist, or arriving at results through exploration rather than reproduction.

[stippenstoh]

This is a great example of using vector drawing for a specific purpose, in this case preventing the Zebra3 bass from colliding with a kick. Bass and this neat trick all done just with one oscillator and only the OSC tools, no filters needed.

[Urs]

Ok, we're getting close to a new beta with hopefully most crash bugs fixed, as well as some more optimisations implemented and most major sound bugs resolved. Maybe next week?

Moving forward, I think we'll start a new thread then, which would be a good place for a FAQ to appear, as this thread has become quite long...

Thank you @ Funkybot's Evil Twin for posting this one, helps a lot already!

This post is intended to help with common questions that have come up repeatedly in this thread. It is unofficial, summarized, and not authoritative. Urs, if it’s not accurate/useful, feel free to use your mod powers to remove it. Or if you want to take it over, modify it, and make it a sticky or something, that works too. Most people won't read the whole thread before asking questions (understandable, not a criticism) so I thought this might be helpful.

Zebra 3 Public Beta (Rev 20399) — Unofficial FAQ
Compiled from recurring user questions and summarized replies from the thread.

General / Status

Q: Is the GUI final?
A: No. The current build is a public beta. The main "Zebra-style" layout is fixed, but colors, fonts, and some parameter/UI details may still change during the beta.

Q: How can all modulations in a patch be viewed at once?
A: A single global “modulation overview” UI is not available yet. A global overview concept has been discussed as something being experimented with internally, with improvements expected over time. The current workflow includes helpers such as focusing/soloing selected sets of modules, but modulation inspection can still require some manual hunting. So expect improvements, but don't expect lots of animations and moving orbs like some other synths.

Q: What is the empty space above the lanes/grid used for?
A: Selection Groups. Selecting multiple modules allows focus on a subset of the patch; a “+” can appear to create groups for quick selection. This is described as a core workflow feature.

Q: What is the safest way to use the beta?
A: A master limiter is recommended (unexpected glitches can occur). Frequent saving is also recommended (rare crashes are still possible).

Q: Will there be another beta?
A: There will be at least one more beta / release candidate ahead of final release.

Q: Is there an FX version / Audio Input?
A: Not yet, but it's being looked into.

Q: Will a Constant Modulator be added?
A: Yes. In an upcoming version.

Controls A–D (Performance controls) / Automation

Q: Can Controls A–D be controlled from hardware now?
A: Yes. MIDI CC can be assigned to Controls A–D for realtime control, and recorded/editable as MIDI automation in the MIDI clip/track.

Q: Will Controls A–D be automatable / exposed as host parameters (not only MIDI CC)?
A: Host-parameter automation has been presented as the likely direction. A quoted reply in-thread: “We’ll probably just make them automatable…”
Practical takeaway: Controls A–D can be driven via MIDI CC now; host-parameter automation has been implied as a future step, but not presented as completed yet.

Performance / CPU / Multi-core / MPE

Q: Is Zebra 3 heavier on CPU than Zebra 2?
A: Yes. A typical Zebra 3 module has been described as often using ~2x the CPU of a Zebra 2 module. Some new filters are described as being in a “Diva-level” CPU ballpark. Care is recommended when stacking many heavyweight modules in high-polyphony patches.

Q: Is multicore support available yet?
A: Not yet (listed as “No multicore support yet”).

Q: What does the Burner module do?
A:It's for performance testing. See the dedicated thread in the U-he forum for more information on why it exists and how to use it.

Q: Is MPE support available yet?
A: Not yet (listed as “No MPE support yet”).

Q: Is Release Velocity supported?
A: Not yet.

Sampling / Presets / Editing

Q: Can Zebralette 3 presets be loaded in Zebra 3?
A: A preset adapter is planned, but described as still work in progress.

Q: Is sample import ready?
A: Sample import is marked as experimental; glitches are expected. Improvements have been indicated as forthcoming.

Q: How reliable are undo/redo and other editing tools?
A: Undo/redo is explicitly described as not fully reliable yet. The curve editor is also described as having quirks and small bugs.

Filters / Levels

Q: Filter types can vary a lot in loudness. Are filters normalized? Will there be normalization?
A: Filters have been described as normalized at Cutoff fully open with Resonance at 0, but different designs behave very differently. Attempts to equalize levels across all filters can create undesirable behavior, and self-oscillation can become extremely loud even when overall output drops.

Known issues (recurring)

Q: Why can the GUI feel slow on Windows?
A: Improving GUI performance on Windows has been framed as a longer-term effort (discussed as a 2026-ish project).

Q: Any other quick “heads up” items?
A: The first post’s Known Issues list is the best starting point (includes sample import being experimental, no MPE/multicore yet, undo/redo not fully reliable yet, rare crashes, and various platform-specific quirks).

Workflow requests (common asks + thread summaries)

Q: Will XY pads come back?
A: Possibly, but if implemented, the expectation stated in-thread is no more than 2. Support across a large factory library was described as a significant effort, and multiple XY pads were described as being difficult to utilize/maintain in prior versions.

Q: Will the arp/sequencer come back?
A: It has been described as on a ToDo list for a later update (mentioned as 3.1 or 3.2), with details TBD. In the meantime, Mod Mappers and the Pitch page can cover some related workflows.

Licensing / Pricing (as stated in the thread’s first post)

Q: What upgrade / pricing options have been mentioned?
A:

• Free upgrade eligibility has been described for owners of both Zebra 2 + The Dark Zebra (from before they became Zebra Legacy).
• Zebra 2 (pre–Zebra Legacy) owners can upgrade for 30€.
• During beta, new purchases have been offered at an introductory 179€ price (noted as 70€ off 249€).

User Guide

Q: Where is the user guide?
A: The comprehensive guide is accessible by clicking the u-he logo in the GUI. Reading it is recommended before filing bug reports.

[NAD]

This is a great example of using vector drawing for a specific purpose, in this case preventing the Zebra3 bass from colliding with a kick. Bass and this neat trick all done just with one oscillator and only the OSC tools, no filters needed.

Ugh, the way he completely misunderstands the Phase Offset oscfx makes me unreasonably upset. No, it's not like Disperser at all! Even without reading the manual it's quite clearly something entirely different! And he's so gosh darn confident about it, too... Ughhh!! That being said, he still manages to put it to good use, I'll admit he ends up making a pretty decent bass sound.

[Urs]

Yeah, it's great example for Map-o-Matic, but the explanation is often incorrect... I think we have to pre-empt the notion that Osc FX are somehow "Spectral". Some are, many are not. Likewise, whether or not a Curve is a spectrum or just a waveform is not a random thing. I'll eventually have to redo the original Zebralette conceptual explainer video and clear up those common misunderstandings.

[jtsterays]

Pop up hint when hovered over would be great too.

[Urs]

Pop up hint when hovered over would be great too.

Long term project... already logged into our ticket system along with other A11y improvements.

[ThoughtExperiment]

If you really have 'scoured the entire internet' then you must have spotted the ton of videos for Zebralette 3 which explain how to do exactly that...

I'd be happy if you share a links to the ton of videos where we're not just spinning a waveform in a vacuum, but recreate a real complex instruments

I would, but I'm really not interested in recreating real instruments. Why bother when you can load up a sample or record yourself playing a real instrument?

I'm interested in creating instruments that don't/maybe can't exist. That's exciting.

But you do you

so you recommended 'ton of videos' to me, even though you don't even know if they exist. excellent:)

I didn't say I don't know where they are, I just don't feel inclined to help you. I can't be bothered to post yet more links to videos that are:
a)already linked here in this thread
b)easy to find if you're willing to try

If you look at Urs' reply to your comment he's attached one of his (quite old now) videos explaining the process. Start from there if you're serious about learning how to use Zebra3 the way you want to use it.

Here's another tip - don't waste your breath replying to this post, you're now on my ignore list.

[stippenstoh]

Yeah, it's great example for Map-o-Matic, but the explanation is often incorrect... I think we have to pre-empt the notion that Osc FX are somehow "Spectral". Some are, many are not. Likewise, whether or not a Curve is a spectrum or just a waveform is not a random thing. I'll eventually have to redo the original Zebralette conceptual explainer video and clear up those common misunderstandings.

I would love to see a 'clean' series of Zebra3 videos, as short and on point as possible. YouTube influencers will always rush to put out content and, as NAD points out, mask ignorance with confidence, often increasing confusion with complex topics like Zebra3.

I still think the way he draws out the fundamental to create a bass that can be automated to not get in the way of a kick is a great example for the concept of vector drawing, specifically because it tackles an everyday issue that usually needs to get fixed with additional processing after sound design.

From my perspective of a hobbyist, tricks like this are highly intriguing and motivating. I remember you already implied that some useful 'tricks & tips' videos will be coming.

[jtsterays]

Looks like I cant use the noise module to modulate spline Osc's pitch, right? (I'm aware of Modnoise, but the module has some weird modes)

[ffx]

Could each OSC FX also have mix amount? You can see this in Ultra synth, it makes very subtle harmonics possible. Could be a hidden parameter, if there is no space left in GUI. (Sorry posted this one in bugs thread)

https://ultra.audio/start

[Urs]

Looks like I cant use the noise module to modulate spline Osc's pitch, right? (I'm aware of Modnoise, but the module has some weird modes)

You are correct. The spline-based oscillator is not suited for FM, and won't be unless CPUs become much, much faster.

(one bug ticket I have worked on recently is related to a preset with a 7 octave drop of the oscillator in very few samples, which it struggles with and simply can't do cleanly)