Zebra 3 Public Beta (final beta)

[synaesthesia]

Well, these are just words again, I would like to see examples, or rather hear them.

Well, look, I didn't want to hijack a thread about the Zebra3 beta release with promoing of my own Zebra programming, that wouldn't be very professional of me.

But, if you want to check out my , lots of examples of Zebra2/HZ sounds on there designed for big, cinematic action scoring etc. Including a trailer for my new soundset which is possibly the sort of thing you're looking to hear... maybe.

[kraster]

So by publishing a beta version, we are fully aware that the first impression may not be the best, and we will lose a whole bunch of potential customers. Many people will not change their minds once they get to try the polished final product. I've seen this many times, it's a thing. Some peeps won't come around even if their niggles are resolved.

However, the feedback from those who constructively take part in this process is invaluable. We sacrifice a few sales in favour of making a better product.

A few years ago I would probably have taken on the argument and tried to sway some minds. Nowadays... if it's not for them, it's not for them

Alright so seeing as this is probably directed at my comment, I’ll try to explain it a little further.

I’m not reacting to the idea of a beta, nor am I making a judgement about the musical intent or feature set. I’m purely talking about the sonic behavior I’m hearing under specific conditions.

In my experience, what my ears usually push back against in digital synths isn’t the core algorithmic idea, but secondary digital artifacts. Things like oversampling strategy, interpolation, phase behavior, or other numerically related side effects. That’s also why I personally don’t like working at 44.1 kHz at all; most of those issues tend to calm down for me at 88.2 kHz, which is where I do almost all of my serious work.

That’s also why I brought up ZebraHZ specifically: at 88.2 kHz it has a density, depth, and mono behavior that, to my ears, feels much more “resolved.” With the current beta, even at 88.2 kHz, I’m still hearing something flatter and less cohesive, and it doesn’t seem to collapse to mono in the same way.

To be clear: I’m not saying “the algorithms are bad” or “this will never be for me.” I’m saying that what I usually react to negatively in digital instruments is often not the musical design, but the digital implementation details, and those are exactly the kinds of things that often evolve during a beta.

So my comment wasn’t meant as a final verdict, just as a description of what I’m hearing right now and where, based on past experience, I’d expect improvements to potentially land. Usually going into these details doesn't work on forums but seeing as THE Urs Heckmann replied I probably should.

If it ends up not being for me, that’s fine but I wanted to clarify that my feedback is coming from that angle, not from a knee-jerk dislike of early software or a resistance to change. You can be sure I will try the release demo again and will very much hope I will like it!

Have you tried experimenting with the oscillator calculation and curve calculation settings in the oscillator rack panel?

They control the frequency of the wave calculation and the precision with which the curves are calculated respectively. Higher settings do eat up CPU though.

Personally I think Zebra 3 is U-he's best sounding synth in terms of raw sound but it also could be the case that you just simply prefer the sound of Zebra HZ and maybe because it has more artifacts relative to Zebra 3.

[Urs]

Alright so seeing as this is probably directed at my comment

Not really, as obviously you enjoy Zebra2/HZ.

I had just read a few comments that question the mere existence of Zebra 3, not just in this thread, people who inform themselves on Youtube (a common theme IMHO that to me comes across as if they need others to tell them what to think). As if, after a few days, the verdict gets told on a video blog and that settles it then.

In contrast, we have some of the best sound designers on board, who have not explored Zebra 3 in all its depth after 5 intense months. But their work still exceeds what we thought was possible. So much stuff, so many alleys of sound design to explore.

Hence, if people give up "trying to like it" this easily, in this early stage and with this little motivation, why not just let them pass?

[billinder33]

I finally got to spend some more time with it last night and I really like the sound of the Zebra. The Oscs and FMO have an inherent raw, midrange punch that I find appealing. Adding the Distortion and Folder to the chain gets some really, agressive, biting tones. It kind of reminds me of the old Nord Lead when it first came out. Polarizing opinion I'm sure.

Unfortunately that's pretty much where it ends for me. The workflow is brutal. Why a separate routing interface for both Oscs and FX, with the same things in both, instead of just one matrix where unlimited instances of any module can be added? Why the fixed (slider-based) envelope positions? The popup filter configurator/matrix makes absolutely no sense. Also, why not just show us graphically what the filter is doing?. You can't sweep the EQ filters en masse without individually assigning them (or alternatively there's supposedly workaround hack that can be done with the pitch?).

Maybe this kind of workflow appeals to someone like Hanz Zimmer who has all day to experiment with different routings options that ultimately produce relatively similar results, but for people who have a general target sound in mind and want to get there fast, Zebra is like the anti-workflow. It's a shame. I spent 3 hours getting lost into building 2 patches and spun a few mudpies out of each, but that's like a deathmarch pace of work. I'm trying to get out of the business of spending hours lost in sound design, so at this point, I'll probably put this one aside and maybe check in on this thread from time time to see how things are progressing.

But I'd like to suggest to U-he a follow-on / separate product based on the Zebra oscs....

Take the OSC, FMO, Noise, (maybe the Exciter and Modal as well) and add the Distortion, Burner, Folder, Filter/EQ (combined), and Envelope to each, creating singular units with a fixed internal routing. Basically merging the fun things you can do to torture the oscs and putting it all in one place/frame, each with the ability to FM one to the other (as per the Serum paradigm). 2-4 oscs total, all running in parallel, stacked in the UI similar to Omnisphere's FX racks. Make sure the combined Filter and EQ share the same visual feedback.

This will allow you to ditch the Osc routing matrix entirely. The visual focus of the UI will be on the combined Osc/Dist/Burn/Fold/FilterEQ/Env frames where 75% of the real work happens. Also ditch the FX router and just have a single FX lane, below the oscs as you have now. Make the FX units collapsable (like Bitwig/Phaseplant). Keep the current modulation system and maintain everything on a single page with no tabbing, since that's a core concept to most things U-he. Also, remake the comp/limiter into some kind of OTT/Smasher so you don't have to stack compressors and limiters to attain that kind of result. Prior to beta, make sure any reasonably experienced synth user can easily build a lead, mid-bass, neurobass,or wub in under a minute without needing the manual.

Give it a clever marketing name like Zebraletto or better yet Ze-Brat (get it? Zebra's sassy, bratty child), charge $120-$140, representing something between Zebralette and Zebra. Market it to the wub wub and DJ crowd as an entry level product into the world of U-he.

Good luck with the upcoming release of Z3, I hope you do well with it.

[NAD]

Take the OSC, FMO, Noise, (maybe the Exciter and Modal as well) and add the Distortion, Burner, Folder, Filter/EQ (combined), and Envelope to each, creating singular units with a fixed internal routing.

Lol, yes! In fact the burner alone deserves its own spinoff plugin! I can use the heat from my CPU to make a nice pot of coffee...

[NAD]

Why a separate routing interface for both Oscs and FX, with the same things in both, instead of just one matrix where unlimited instances of any module can be added?

On a more serious note though, it might clear things up if you realize that the main grid is for per-voice stuff and the FX grid is for post-FX, so it makes sense (at least to me) to split it up that way.

[Funkybot's Evil Twin]

Alright so seeing as this is probably directed at my comment, I’ll try to explain it a little further.

I’m not reacting to the idea of a beta, nor am I making a judgement about the musical intent or feature set. I’m purely talking about the sonic behavior I’m hearing under specific conditions.

In my experience, what my ears usually push back against in digital synths isn’t the core algorithmic idea, but secondary digital artifacts. Things like oversampling strategy, interpolation, phase behavior, or other numerically related side effects. That’s also why I personally don’t like working at 44.1 kHz at all; most of those issues tend to calm down for me at 88.2 kHz, which is where I do almost all of my serious work.

That’s also why I brought up ZebraHZ specifically: at 88.2 kHz it has a density, depth, and mono behavior that, to my ears, feels much more “resolved.” With the current beta, even at 88.2 kHz, I’m still hearing something flatter and less cohesive, and it doesn’t seem to collapse to mono in the same way.

To be clear: I’m not saying “the algorithms are bad” or “this will never be for me.” I’m saying that what I usually react to negatively in digital instruments is often not the musical design, but the digital implementation details, and those are exactly the kinds of things that often evolve during a beta.

So my comment wasn’t meant as a final verdict, just as a description of what I’m hearing right now and where, based on past experience, I’d expect improvements to potentially land. Usually going into these details doesn't work on forums but seeing as THE Urs Heckmann replied I probably should.

If it ends up not being for me, that’s fine but I wanted to clarify that my feedback is coming from that angle, not from a knee-jerk dislike of early software or a resistance to change. You can be sure I will try the release demo again and will very much hope I will like it!

If you hear a difference and think there's opportunity for improvement somewhere, why not try to recreate a Z2HZ patch in Z3 (they're obviously not patch compatible, but you should get close for most things) and post the 88.2khz audio along with the patches to demonstrate what you're talking about?

Perhaps it's the filters, perhaps it's an oscillator setting, perhaps your eyes and brain are hearing things that aren't there or others don't hear. Maybe there's some quirk with a partcilar model at the sample rate and just not a lot of people run 88.2k so it slipped through the cracks.

I think it's important to listen to people who are trying to be constructive and give them the benefit of the doubt, but unless we can hear what you're hearing, it's hard to tell if you might have a point. It can also be the case that confirmation bias has your brain hearing things that aren't there. It's so easy in audio to think X is better than Y because everything's subjective and our eyes, ears, and brains just love to trick us [like this persistent ringing I've been hearing for weeks!]. Anyway, just a suggestion.

[Urs]

I had my own difficulties recreating some patches from Z2 in Z3. I would not expect anyone to have to go through this. And I'm sure it'll be a lot more difficult the other way round. One Z2 patch sounded really good due to a happy accident with a glitch. Z3 does not have that glitch, so it can't sound the same.

But that's why Zebra 2 does not go away.

[Funkybot's Evil Twin]

Fair enough.

I was just thinking, do something simple, comapre. If we're all like "OMG, there is a heft there we can't recreate in Z3" that maybe there's something to that. Or if everyone says, "dude, you're hearing things" then we know there's nothing to it.

But if they're truly that different that even "mono saw patch into filter" wouldn't be comparable, then no point. I've also been around plugins long enough to know that stranger things have come up in betas than "hmm...yeah, it does sound a little weird at 88.2khz". So something silly like, "doh, you're right, it's actually a bug in X module at [or the SR conversion to/from] 88.2khz" wouldn't surprise me either.

[Funkybot's Evil Twin]

Sometimes, the boys who cry wolf actually get eaten by one!

[NAD]

One Z2 patch sounded really good due to a happy accident with a glitch. Z3 does not have that glitch, so it can't sound the same.

I was actually just wondering about something like that today. I noticed the FMO self feedback sounds completely different in Z3, no matter how I tweak it. Do you mind providing some insight about that?

[PieBerger]

I've never used Zebra before, but after trying the Zebra 3 beta, I found the sound to be very subtle and simple, compared to synth monsters like Dune 3, for example. Where a saw would cut right through the mix, here everything is so subtle and simple. I thought, of course, that the synth is complex and I probably don't understand everything 100%, but I went to YouTube to look at samples of Zebra 2 and Zebra 3 and couldn't find any strong, powerful, or ultimate sounds.
I've read a lot of rave reviews about the sound design and all that, but I completely don't understand where it is in Zebra. I don't mean to offend anyone, but on the contrary, please point me in the right direction to see the sound design and all that wonder everyone is talking about here.
Also, while I was writing this message... I don't even know what kind of music this synth is suitable for? In an age when there are huge libraries of live instrument samples, vocals, granular synthesizers, and so on... I probably didn't understand the main thing—why this synthesizer is needed. I'd be very interested to know.

What are some strong, powerful or ultimate sounds to you? To use one of your follow up replies against you (sorry!), these are just words, can you show us some example patches (YouTube links are OK, they don't have to be your own patches) of what you mean specifically.

[Urs]

One Z2 patch sounded really good due to a happy accident with a glitch. Z3 does not have that glitch, so it can't sound the same.

I was actually just wondering about something like that today. I noticed the FMO self feedback sounds completely different in Z3, no matter how I tweak it. Do you mind providing some insight about that?

The original DX style feedback averages two samples. This way one can get a pretty good sawtooth from it, before it goes into a chaotic state. During Zebra 2 development I did not know about the averaging, so it goes into a chaotic state quite a bit before the sawtooth is really crisp.

Here, if we had some examples A/B that show that Zebra 2 offers a really cool advantage, I'd be happy to make the averaging thing an option.

[NAD]

Here, if we had some examples A/B that show that Zebra 2 offers a really cool advantage, I'd be happy to make the averaging thing an option.

One of my favorite factory presets in Z2 is "HS Fuzzy Feeling", so I decided to try and replicate it in Z3 but did not get very far. If you turn off all the fx and isolate the FMO just by itself it already sounds pretty cool (IMO). The FMO in Z3 doesn't really have that sound in its range (AFAICT), despite having all sorts of other cool stuff. Then once you add in the comb module (which is an entirely different beast in Z3) you realize it's hopeless and give up. I'm sure I'll have new favorite factory patches but that made me a little bit sad.

[Urs]

Hehehe, I'm sure that one can be done closely in Z3, and probably with fewer modules...