Zebra 3 Public Beta Revision 20399

[ralfrobert]

I look forward to any improvements made for the official release and find it commendable that u‑he is actively monitoring feedback.
I especially hope the release includes many more factory presets, as I think that’s the best way to learn the synth, particularly since I never owned Zebra 2.
The white/blue color scheme from Zebralette 3 would be fantastic as well.
Congrats, u‑he!

A Zebra 3 Cookbook would be nice.

[JoeLowery215]

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)

I’ve been using the LFO, I think at audio rate, with key tracking and it sounds just like cross modulation to me… I mentioned this earlier but it doesn’t sound the same in Z2. Is there something else going on?

It sounds great in Z3 and I’m glad I can get some
Cross modulation sounds.

[JoeLowery215]

I'm trying to find out if this was mentioned but did you get rid of the "soft attack" mode? I can't for the life of me to get a super clicky sine wave sound.

V1 Init, at 100. Of course Attack at 0. You might also have to adjust oscillator Phase so that it starts at an extreme.

Oh great, thanks. Wasn’t sure what that knob was for haha

[Urs]

I especially hope the release includes many more factory presets, as I think that’s the best way to learn the synth, particularly since I never owned Zebra 2.

Thank you. There'll be at least twice as many presets as far as I know... we'll see...

[sarmad5]

So I have 2 Oscs with 2 different Envs, I want both of them routed to a filter, but keeping their separate envs. What's the cleanest way to do this? Atm what Im doing is just chaining all of them to 1 column, then use Env 2 to module the volume knob in Osc 2. But that's still not a workaround because if for example env 1 is a long attack pad type adsr, and env 2 is a short decay pluck, the env 1 is gonna override env 2 and make my 2nd osc disappear.
gagaddggda.png

One solution could be to apply the filter post envelope in Fx/Out section.

[JoeLowery215]

I'm trying to find out if this was mentioned but did you get rid of the "soft attack" mode? I can't for the life of me to get a super clicky sine wave sound.

V1 Init, at 100. Of course Attack at 0. You might also have to adjust oscillator Phase so that it starts at an extreme.

Actually this doesn't work for the FMO sine wave

Maybe be a bug?

[jtsterays]

So I have 2 Oscs with 2 different Envs, I want both of them routed to a filter, but keeping their separate envs. What's the cleanest way to do this? Atm what Im doing is just chaining all of them to 1 column, then use Env 2 to module the volume knob in Osc 2. But that's still not a workaround because if for example env 1 is a long attack pad type adsr, and env 2 is a short decay pluck, the env 1 is gonna override env 2 and make my 2nd osc disappear.
gagaddggda.png

One solution could be to apply the filter post envelope in Fx/Out section.

I tried that, the filter sounded kinda weird because it's not poly. Looks like this simple routing is actually impossible.

[Q-Bass]

So I have 2 Oscs with 2 different Envs, I want both of them routed to a filter, but keeping their separate envs. What's the cleanest way to do this? Atm what Im doing is just chaining all of them to 1 column, then use Env 2 to module the volume knob in Osc 2. But that's still not a workaround because if for example env 1 is a long attack pad type adsr, and env 2 is a short decay pluck, the env 1 is gonna override env 2 and make my 2nd osc disappear.
gagaddggda.png

One solution could be to apply the filter post envelope in Fx/Out section.

I tried that, the filter sounded kinda weird because it's not poly. Looks like this simple routing is actually impossible.

And why not modulate Volume of OSC1 with Env1 and Volume of OSC2 with Env2? Then use the Mixer 2in1 module let's say in the first column and after that plug in the filter.. Or maybe better modulate with those envelopes the Volume knobs in the Mixer, because they have those rectangle "primary" targets, which Urs wrote are always better choice, because they work more precise..
Then, if there is nothing else in the patch, you simply mute the output of second column and set the envelope of the first column to something that preserves shapes of all previously used envelopes.

[jtsterays]

So I have 2 Oscs with 2 different Envs, I want both of them routed to a filter, but keeping their separate envs. What's the cleanest way to do this? Atm what Im doing is just chaining all of them to 1 column, then use Env 2 to module the volume knob in Osc 2. But that's still not a workaround because if for example env 1 is a long attack pad type adsr, and env 2 is a short decay pluck, the env 1 is gonna override env 2 and make my 2nd osc disappear.
gagaddggda.png

One solution could be to apply the filter post envelope in Fx/Out section.

I tried that, the filter sounded kinda weird because it's not poly. Looks like this simple routing is actually impossible.

And why not modulate Volume of OSC1 with Env1 and Volume of OSC2 with Env2? Then use the Mixer 2in1 module let's say in the first column and after that plug in the filter.. Or maybe better modulate with those envelopes the Volume knobs in the Mixer, because they have those rectangle "primary" targets, which Urs wrote are always better choice, because they work more precise..
Then, if there is nothing else in the patch, you simply mute the output of second column and set the envelope of the first column to something that preserves shapes of all previously used envelopes.

With your method, I just need to use 1 column, no need for the mixer (so column 1: osc 1 > osc 2 > filter). But the problem that it still goes through the final VCA at the bottom, and since Env 1 & 2 are already used, now I have to use Env 3 which is overly complex and wasting another env.

[Urs]

Hehehe, putting the filter behind the VCA is never a good idea for a software synthesizer.

Polysynths with dynamic voice processing need a place that determines "ok, the voice is quiet now, I can put it to sleep". That place needs to be the last element in the signal chain, before the individual voices are mixed for output or effects processing. In Zebra, much like every synth that puts its voices to sleep, the combination of VCA and envelope does this job.

No exception comes to mind, but I do know of some synth plug-ins that process all voices forever. But even those have the VCA as last element in the voice processing chain.

(note: I am careful to use the word "soft synth" and "plug-in" because hardware does not necessarily put voices to sleep)

For the problem at hand, the only reason for wishing otherwise that I can think of is that the filter character is emphasized in a certain way, based on oscillator gain. If the filters in the FX section don't cut it, then it must be a poly patch. So it must be done before the VCA.

Two solutions come to mind:

a) using 2 envelopes and applying them on parallel VCAs (adding up) after the filter output while applying them differential on the respective oscillator volumes, in order to get the mixing right while preserving overall volume.

b) Do just the same thing you originally did, but use an MSEG with a squarish shape as VCA source that keeps it open long enough until the voice has gone quiet. (Once we add envelope followers in a subsequent update, this can be done dynamically)

(if b is impossible because the filter is ringing forever, you've accidentally made a case for as to why filters should not be placed behind VCAs in a voice)

[wilx]

There'll be at least twice as many presets as far as I know... we'll see...

I hope The Unfinished gets involved - Matt's presets are the first ones I go to for ZebHZ

[jtsterays]

Hehehe, putting the filter behind the VCA is never a good idea for a software synthesizer.

Polysynths with dynamic voice processing need a place that determines "ok, the voice is quiet now, I can put it to sleep". That place needs to be the last element in the signal chain, before the individual voices are mixed for output or effects processing. In Zebra, much like every synth that puts its voices to sleep, the combination of VCA and envelope does this job.

No exception comes to mind, but I do know of some synth plug-ins that process all voices forever. But even those have the VCA as last element in the voice processing chain.

(note: I am careful to use the word "soft synth" and "plug-in" because hardware does not necessarily put voices to sleep)

For the problem at hand, the only reason for wishing otherwise that I can think of is that the filter character is emphasized in a certain way, based on oscillator gain. If the filters in the FX section don't cut it, then it must be a poly patch. So it must be done before the VCA.

Two solutions come to mind:

a) using 2 envelopes and applying them on parallel VCAs (adding up) after the filter output while applying them differential on the respective oscillator volumes, in order to get the mixing right while preserving overall volume.

b) Do just the same thing you originally did, but use an MSEG with a squarish shape as VCA source that keeps it open long enough until the voice has gone quiet. (Once we add envelope followers in a subsequent update, this can be done dynamically)

(if b is impossible because the filter is ringing forever, you've accidentally made a case for as to why filters should not be placed behind VCAs in a voice)

Thank you, b worked. I also added looping to the MSEG so I can hold the notes indefinitely (1 osc is a pad type: full sustain, long release adsr) while still having the release stage (gate VCA just cuts it off immediately)
I'm too stupid to visualize A in my head (also amateur english) so I didnt try it.

What do you mean about the env follower? Will it be a VCA source that I can choose so the MSEG wont be needed anymore?

[Urs]

What do you mean about the env follower? Will it be a VCA source that I can choose so the MSEG wont be needed anymore?

An envelope follower could determine silence with a threshold, and then switch the MSEG curve from "100%-for-a-long-time" to "done" once the signal is quiet.

[einei]

Sorry of this a noob question, but I really can't figure this out, also tried to read the manual. How do I get an Lfo or a Mseg to modulate a parameter all the way from "top" to "bottom". The modulation visual indicator "ring" doesn't cover the full range and the pitch drop stops in the middle. Sorry for my lack of proper terminology here.

[Urs]

So the ring only represents the modulation depth parameter. If the modulator is bipolar (which the LFO normally is), the modulation extends to the negative side in the same depth. So what you see is already what you wanted to achieve.

We are working on a way to show the bipolar aspect with a differently coloured arc on the opposite side, hopefully for the final release version.

For some background, the polarity of a modulation source is determined by multiple factors, which are often dependent on context. So the knob does not necessarily "know" the extent of the modulation. E.g. whether a Mapper is bipolar or not depends on the values drawn into it. Hence, I always found that just painting a default bipolar arc for the modulation may be a bit misleading, hence I wanted to focus on the depth parameter itself.

However, people frequently ask for this, probably because it's what they know from other software, and even though I'm still not the biggest fan of the idea, we'll just do it.