Zebra 3 Public Beta (final beta)

[jtsterays]

Will the unison section be expanded in the future? I'd love if Z3 has the Swarm mode in Dune 3, it's very unique sounding.

[Gert Wilden]

Hello fellow u-hes,
here are my first impressions, quickly put together:

1) Sound: Truly fantastic! Zebra 3 seems to combine the refined, organic sound of Diva with the powerful definition of Hive and the great flexibility of Zebra 2.

2) GUI: As a user of Zebra 2 and Hive, you'll find your way around immediately. Many well-implemented details (filter type selection matrix, etc.).

3) What I really miss, and I hope this won't be final:
the four XY pads!
The actual four controller buttons can't replace them.
Zebra's flexibility rivals that of many complex modular systems, and you might want to "play" these sounds like a modular performern rather than a „musician“ (using a keyboard).
These four controllers are simply not enough for that. And the really cool Zebra 2 macros were actually perfect. One of their major advantages is that, because they have a defined center position, you can adjust parameters, especially when using encoders, without risking parameter jumps. Please, please bring them back!

4) Yes, some of the new features, especially the integration of mappers with quantized pitches, do enable quite extensive sequencer-like applications. Nevertheless, I wouldn't want to do without an arpeggiator/sequencer. The buzzword here is: generative! That's the direction modern sequencing is heading, and simply letting a sequence run through its entirety gets boring after a while.

A little bit of randomness and probability would be nice.
Some competitors (Pigments) already have cool solutions for this, but I'm sure you can do better!
(Sometimes it would help if you could address the number of steps in the mapper as a modulation target).

5):
It's probably one of the known issues that some presets push the CPU to its limits, for example “Sidewinder” or “Eternity Divides.” They sound great, but don't try them with pads…
Is that the final word in terms of CPU usage?

6): The oscillators are a rabbit hole that you have to delve into longer and more intensively to say more about them than that they basically sound great.

[PieBerger]

Maybe I have a different understanding, of how a semi-modular synth should be, than you. In my opinion the goal should be, not to stand in the way and give the user as much freedom as possible, to realize a sound in the way he is intending it.

Semi-modular just means somewhere between fixed and fully open architecture, it doesn't necessarily mean lots of options. ACE and Bazille are semi-modular synths in that they offer users more flexibility to route audio and or control signals, compared to say a Minimoog emulation, but aren't fully modular like Reaktor/VCV/Eurorack where practically anything goes. ACE and Bazille both have a very limited feature set compared to synths like Phase Plant or Zebra, but they are all semi-modular.

[DCrown]

My first impression is that it has some very good sounds/presets. I have noticed a way higher cpu usage than Zebra 2 though.
Frankly speaking I expect an upgrade to be optimized, it means a better cpu performance than previous version, so I will wait, maybe it will be improved....

Every single module in Zebra 3 is objectively improved in sound over its Zebra 2 counterpart. In the most simple case, modules in Zebra 3 are likely to be oversampled when in Zebra 2 they are not, and they are likely to implement novel technologies for anti-aliasing (if ADAA rings a bell as a term, we use it pretty much everywhere, and we might have the first FM oscillators deploying a version of it). Oversampling would approximately double the CPU.

So in that sense, every Z3 module is expected to at least consume twice as much CPU as their Z2 counterparts. Speaking of FMOs for instance, those now are dual oscillators. So it's intuitively clear that they use more CPU than the single FMOs from Z2.

that said, a few modules will see more optimisation. The texture reverb for instance can use up to 20% on my machine, but a team of devs is on it, just didn't want to risk breaking anything before this beta release.

Thanks for your reply. I will defnetely buy it having Zebra 2 legacy "only". Will it still be possible to demo the officially released full version to check cpu?

[Teksonik]

Name me ONE synth that can do anything in the world OUTSIDE of a truly dedicated modular synth that doesn't take a masters degree in engineering to figure out how to use?......
I'll wait while you count all zero of them.

Nexus 5, VPS Avenger, Parawave Rapid, Omnisphere, Falcon, and so on....All very deep and powerful synths that are dead easy to patch. The first three in that list are synths I always reach for before Zebra because they have the features that are important to me.

Whether or not you have the skills or intellect to use them is irrelevant.

So by your reasoning Zebra 3 should be feature locked at this point and should not receive one single new feature? Just because it will make it too complicated for you? I think we see where the entitlement lies.

No one is asking Z3 to "do everything" that's just you jumping to an illogical conclusion. I'm just making suggestions on what I think would make Zebra 3 a better synth and at the end of the day it's a product that has to compete with other products in the market segment. With a full list price (not intro price) of $291 USD if my calculations are correct the competition is pretty stiff in that range.

[Teksonik]

Or use different synths that do EXACTLY what YOU want, or try a different approach with a synth that makes you think about how to get a specific result. Play with those different types of gate triggers, MSEG's and ModMappers..they are awesome.

I do use synths that do EXACTLY what I want and no, the MSEGs and Mappers are not the same as an Arpeggiator. I've already made over 20 patches for Z3 so I'm putting in the work.

The point is do you want people to pay $290 for a synth at full price and then have to constantly reach for other synths? How successful would that product be?

Like I said don't dumb down Zebra 3 just because YOU don't need a certain feature. That's selfishness beyond reason. I welcome any and all changes to Z3 whether I need or want them or not.

I for one love the limitations.

Well you know the old saying.... Simple Things Amuse Simple Minds....

Just use Serum or something else if you think Z3 isn't enough for you..quite easy choice right?

Or, and hear me out, during this BETA period I can give the developer input on what I think will make Z3 a better synth. Whether he implements those features or not is out of my hands but it sure beats blindly defending the status quo just because you can't handle anything beyond limitations.

Anyway I have to go get groceries now so I'll have to catch up on this thread later.....

[Dr.Gunjah]

Our approach to the ModMatrix is ultra flexible, but to make it feasible, it runs on a more sparse granularity than direct modulation slots. The control rate of direct modulation slots is 1/4 sample rate or 10kHz, while the granularity of the ModMatrix is about a sixteenth of that, about 800Hz when the sample rate is 44.1/48 kHz.

So in Hive we tested various ways to increase that granularity, but the audible differences (if any) would not justify the increased CPU hit. At an update rate of 8-16 samples (which is better than many, many synths), the CPU doubles. Not worth it.

That's an interesting aspect actually. Have you considered to make the resolution switchable in the matrix? Besides decluttering the modules even further it would also make it far easier to quickly decide if you want the higher quality or rather save some cycles instead in case there is no audible difference.

[SamDi]

Our approach to the ModMatrix is ultra flexible, but to make it feasible, it runs on a more sparse granularity than direct modulation slots. The control rate of direct modulation slots is 1/4 sample rate or 10kHz, while the granularity of the ModMatrix is about a sixteenth of that, about 800Hz when the sample rate is 44.1/48 kHz.

So in Hive we tested various ways to increase that granularity, but the audible differences (if any) would not justify the increased CPU hit. At an update rate of 8-16 samples (which is better than many, many synths), the CPU doubles. Not worth it.

That's an interesting insight of the reason why this two approaches exist. Even when it's mentioned in the manual somehow, I didn't really understand it.

So if we had once enough processor power to calculate modulations in audio rate in a central matrix, direct modulations would be dropped?

[rickcupertino]

Hi everyone,

I just wanted to give some feedback on Zebra 3 after trying the beta last night:

Pros:

Oscillators, filters, and fx sound fantastic. Low CPU usage on Mac Studio M2 Max.

I am running it in Logic and I’m able to directly compare the Physical Modelling (Modal and Exciter) with Logic’s Sculpture and Zebra 3 is light years ahead, both in terms of sound and efficiency (I am running Sculpture on High Definition).

Compared to Logic’s Alchemy, the sound of Zebra 3 is much more refined, clearer, and higher fidelity and makes Alchemy seem lo-fi. I am comparing patches made on Alchemy (Ultra Quality) with stacked OSCs, unison, comb filters and additive features with Zebra standard quality (Fast).

Cons:

The main grid is cumbersome and perplexing. Why can’t you just click on the cables and route them between the modules freely, like a node-based workflow - as in ComfyUI or Blender?

The modulation features are confusing. It took me about thirty minutes to work out how the direct modulation works - you have to hold down on the two triangles and drag down. On the filter module, if you want to modulate the cutoff with an LFO, you can also create an LFO on the right side of the screen and drag the square to target the cutoff - this seems confusing for someone who has never used Zebra before.


Conclusion:

Great-sounding and complicated synth - it does need some improvements in the UI and workflow. Has huge potential though, in fact, I am comparing it to my favourite synth Omnisphere 3 and I can say the sound quality is right up there with Omnisphere 3 and that uses a 60GB sound library to achieve most sounds, this can do it purely using synthesis which is amazing if you think about it.

[Urs]

Will it still be possible to demo the officially released full version to check cpu?

But of course, all our plug-ins have fully functional demos (with just some annoying thing added that people may wish to go away.)

[Urs]

Our approach to the ModMatrix is ultra flexible, but to make it feasible, it runs on a more sparse granularity than direct modulation slots. The control rate of direct modulation slots is 1/4 sample rate or 10kHz, while the granularity of the ModMatrix is about a sixteenth of that, about 800Hz when the sample rate is 44.1/48 kHz.

So in Hive we tested various ways to increase that granularity, but the audible differences (if any) would not justify the increased CPU hit. At an update rate of 8-16 samples (which is better than many, many synths), the CPU doubles. Not worth it.

That's an interesting aspect actually. Have you considered to make the resolution switchable in the matrix? Besides decluttering the modules even further it would also make it far easier to quickly decide if you want the higher quality or rather save some cycles instead in case there is no audible difference.

It might be a bit too nerdy that way... but who knows?

We do have different concepts we can explore. Global switches, per module switches, individually fixed parameter resolutions... These are things we need to do at more quiet times though.

[Urs]

Our approach to the ModMatrix is ultra flexible, but to make it feasible, it runs on a more sparse granularity than direct modulation slots. The control rate of direct modulation slots is 1/4 sample rate or 10kHz, while the granularity of the ModMatrix is about a sixteenth of that, about 800Hz when the sample rate is 44.1/48 kHz.

So in Hive we tested various ways to increase that granularity, but the audible differences (if any) would not justify the increased CPU hit. At an update rate of 8-16 samples (which is better than many, many synths), the CPU doubles. Not worth it.

That's an interesting insight of the reason why this two approaches exist. Even when it's mentioned in the manual somehow, I didn't really understand it.

So if we had once enough processor power to calculate modulations in audio rate in a central matrix, direct modulations would be dropped?

There's another aspect. Direct modulations use a lot less screen estate than ModMatrix slots. Naturally so, because they live within their context.

I'm trying to imagine a few of my patches with only ModMatrix slots. I think the number would double in some cases.

Now if that is so, patches with 30+ slots would become patches with 60+ slots. Who's gonna keep an overview here, and what would have to change in the UI to make this happen?

So I think there are pros and cons to each concept, there are pros and cons to sticking to purely one concept or having a mixed approach.

Apparently I'm a mixed approach guy. We have MSEGs and Envs. We have Oscillators and FMOs, LFOs and looping MSEGs. We have small editors and big editors, we have Racks and Tabs. So it seems that my design philosophy for Zebra 3 is based on balancing polar opposites, as if to weave a pattern of black and white stripes. Not sure I'd ever stray from this.

[wagtunes]

Name me ONE synth that can do anything in the world OUTSIDE of a truly dedicated modular synth that doesn't take a masters degree in engineering to figure out how to use?......
I'll wait while you count all zero of them.

Nexus 5, VPS Avenger, Parawave Rapid, Omnisphere, Falcon, and so on....All very deep and powerful synths that are dead easy to patch. The first three in that list are synths I always reach for before Zebra because they have the features that are important to me.

Whether or not you have the skills or intellect to use them is irrelevant.

So by your reasoning Zebra 3 should be feature locked at this point and should not receive one single new feature? Just because it will make it too complicated for you? I think we see where the entitlement lies.

No one is asking Z3 to "do everything" that's just you jumping to an illogical conclusion. I'm just making suggestions on what I think would make Zebra 3 a better synth and at the end of the day it's a product that has to compete with other products in the market segment. With a full list price (not intro price) of $291 USD if my calculations are correct the competition is pretty stiff in that range.

I'm sorry. I own all those synths and they do not do everything. I can name plenty of things they can't do simply because they are not fully modular and don't have every possible module under the sun.

As for my skill set, I can use any synth. Been doing this for 50 years. I'm talking about the average user. Not everybody wants a Softrube Modular or even a Cherry Audio Modular. These are not easy synths for the average user to program. And the proof of that is Softrube Modular is my best selling library by miles. It shouldn't be. It shouldn't have that big a market share. But it sells that well because nobody knows how to program the damn thing. Conversely, my Legend library sells very little. I'm sure more people own The Legend but who the hell needs a patch library for a synth that's so simple? And my sales reflect that.

And as far as Zebra 3 not adding anything new, I never said it shouldn't. All I'm saying is if it DIDN'T add anything new, it's fine the way it is. It doesn't have to have everything under the sun.

Having said all that, you are perfectly within your rights to have complaints about it. I'm not happy with everything about Zebra 3 either. But I am grateful for what it does have whereas the impression I get from you is that because it doesn't have everything you want, you're not happy with it. Again, that's perfectly within your right. You're the consumer. You determine whether or not you buy something. Not the company who makes it.

Again, Urs has to decide if it's worth the man hours to stuff Zebra 3 with even more features. It's not like he can just snap a finger and it happens. So as far as competing with other products in Zebra 3's price range, well, I guess the market will be the one to determine that. But I'm willing to bet my money that Zebra 3's sales aren't going to suffer significantly just because it doesn't have a dedicated ARP (something I too wish it had) or whatever else it doesn't have that you want.

Again, I am just grateful for what it does have.

YMMV

[Urs]

I just wanted to give some feedback on Zebra 3 after trying the beta last night:

Great feedback, thank you!

So we have not yet even begun to properly showcase Zebra 3 in videos and tutorials. We always play the long game, it'll happen in time. Also, I think we're about half the company size of Spectrasonics, and we do not even have a proper marketing department.

Our current approach is to make Zebra 3 for people who already know and love Zebra 2, before we try to captivate a larger audience. So maybe we should've done more to pick up others from where they are, but we weren't sure either how Z3 would be perceived. Expectations are high.

Anyhow, Zebra was always controversial, in the same way as you and others describe. Over the years we have seen a pattern emerge though, where some people either "click" with it or not. I think that a large part of next year will go into providing materials that help people click.

The reason it is like it is has something to do with my general design philosophy. Some designs need to be self explanatory, easy to grasp and immediate. Ticket vending machines for instance. Other designs can't afford the luxury of "easy to use", because they need to be operated quickly, from muscle memory, where constant guiding would get in the way. Cars are like that, in a way. There's no "let me show you how to set gear" all the time. You figure it out once, then forget about it.

So I think that Zebra is one of those concepts that are very fast to work with once an initial hurdle is taken. After some time, what seems convoluted, becomes intuitive. I often speak about "tasks" in sound design. Like, "now I make a waveform" or "now, let's mix these elements!" or "ok, let's make this thing a but faster and deeper". That's where things become more clear maybe.

The point of Zebra as a concept is that I think that

- sound design tasks often involve multiple modules
- ideally all modules involved in a task are accessible at once
- humans are extremely capable at learning abstract concepts
- the general overview must never be further away than one click

I have no immediate way to go into more depth, but yeah, we might have to lay all those thoughts out in time.

[Gert Wilden]

I just wanted to give some feedback on Zebra 3 after trying the beta last night:

Great feedback, thank you!

So we have not yet even begun to properly showcase Zebra 3 in videos and tutorials. We always play the long game, it'll happen in time. Also, I think we're about half the company size of Spectrasonics, and we do not even have a proper marketing department.
So I think that Zebra is one of those concepts that are very fast to work with once an initial hurdle is taken. After some time, what seems convoluted, becomes intuitive.


God gives us everything in exchange for effort. (L.da Vinci)