Zebra 3 Public Beta 3 Revision 20977

[ThoughtExperiment]

I didn't see anything about a sequencer in this list. Is that not going to be implemented in the final release?

As I just wrote elsewhere, we think long term.

Like I think I might have mentioned in the , the current means to do sequencing in Zebra 3 are vast, as also the drum kit video and many presets show. So the idea is that we initially release Zebra 3 without any dedicated arpeggiator or sequencer. Then we see what people can do with it. And only then can we try to figure out how to make an arpeggiator and sequencer that integrates with the available concepts. Who knows, it may look quite familiar or maybe totally different than one would expect. Maybe it is not necessary at all, and all we need is a meta-layer that transforms sequences into Zebra 3 patches. Dunno, we'll see.

A real addition (and not some gimmick) would be an arp/sequencer per pitch (pitches 1 thru 4). With per arp/seq at least one mod output

Up to now, I've been fairly ambivalent about a 'traditional' arp/sequencer in Z3. I hardly ever used the seq in Z2, not because I didn't like it, but because I often want to share an arp pattern between two or more instruments so it's more convenient to use my DAW's toolset or an external sequencer.

There's also my feeling that I ought to at least respect the effort that's gone into Z3's approach to sequencing by making the effort to get my head around it

However, this idea is really intriguing, plus it retains that point of difference between Z3 and every/any other synth, keeping it unique. Very interested to see Urs' view on this.

[Sampleconstruct]

The combo of MSEGs with Mappers plus all the various trigger options for envelopes and OSC FX is more interesting than most sequencers I've come across, so whatever you do, don't touch it

[exmatproton]

The combo of MSEGs with Mappers plus all the various trigger options for envelopes and OSC FX is more interesting than most sequencers I've come across, so whatever you do, don't touch it

i agree...those current options should be available and not be replaced by...

[Urs]

So, sequencing has something to do with Pitches and Gates. Commonly, there's one pitch and one gate per voice. Gate goes high, voice starts, pitch is applied and when gate goes low, the voice fades out.

So just focussing on the Pitch modules falls short of integrating gates. No matter how one changes perspective, there will always need to be some kind of gate or event that starts a voice, and ends it.

Many people suggested different sequences for each lane of Zebra's grid. But as signals can cross lanes, the grid itself needs to be tied to a voice. Hence, there is also no way around having one kind of gate or event that starts (and ends) the complete grid at once. In other words, the modules in the grid are always part of the same voice.

So we developed a different perspective, and with it a different prospect. While there is only one Gate for the complete voice, there can be many triggers. One can already see what that makes possible in Zebra 3. It's probably my favourite improvement over Zebra 2.

Another thing we added are the various mixers and dynamic routing options, as e.g. shown in the arpeggiator video as well as the drum kit one.

These things together open up options to create polyphonic, multitimbral and multi/polyrhythmic sequences and arpeggios, from just one sequencer/arpeggiator ( -> one source of gate events). The open questions evolve less around the basic concepts as merely around usability aspects. It'll take a long time to develop this into something fun and controllable. We need to try different things and figure out what's best.

[exmatproton]

So, sequencing has something to do with Pitches and Gates. Commonly, there's one pitch and one gate per voice. Gate goes high, voice starts, pitch is applied and when gate goes low, the voice fades out.

So just focussing on the Pitch modules falls short of integrating gates. No matter how one changes perspective, there will always need to be some kind of gate or event that starts a voice, and ends it.

I am not sure what the issue is here. Since the 4 different pitches already start with 1 gate event, this gate event can be used to start a arp/seq. Am i missing something?

[Urs]

So, sequencing has something to do with Pitches and Gates. Commonly, there's one pitch and one gate per voice. Gate goes high, voice starts, pitch is applied and when gate goes low, the voice fades out.

So just focussing on the Pitch modules falls short of integrating gates. No matter how one changes perspective, there will always need to be some kind of gate or event that starts a voice, and ends it.

I am not sure what the issue is here. Since the 4 different pitches already start with 1 gate event, this gate event can be used to start a arp/seq. Am i missing something?

hehehe, you can't just put arpeggiators/sequencers into the pitch modules as the pitch modules are part of the voice, and the voice needs to be started by the gate from the sequencer. So if the sequencer is inside the voice, it can't trigger the voice. The sequencer needs to be outside, before the voices.

Otherwise we would not need an extra sequencer: 4 Mappers + 4 Pitches would do the trick. The reason they don't is that a proper sequencer not only deals with pitch but also with gates.

[exmatproton]

So, sequencing has something to do with Pitches and Gates. Commonly, there's one pitch and one gate per voice. Gate goes high, voice starts, pitch is applied and when gate goes low, the voice fades out.

So just focussing on the Pitch modules falls short of integrating gates. No matter how one changes perspective, there will always need to be some kind of gate or event that starts a voice, and ends it.

I am not sure what the issue is here. Since the 4 different pitches already start with 1 gate event, this gate event can be used to start a arp/seq. Am i missing something?

hehehe, you can't just put arpeggiators/sequencers into the pitch modules as the pitch modules are part of the voice, and the voice needs to be started by the gate from the sequencer. So if the sequencer is inside the voice, it can't trigger the voice. The sequencer needs to be outside, before the voices.

Otherwise we would not need an extra sequencer: 4 Mappers + 4 Pitches would do the trick. The reason they don't is that a proper sequencer not only deals with pitch but also with gates.

i thought of something like; gate -> arp/seq (build in gates) (when an arp/seq is disabled, the main gate circumvents the arp/seq) -> pitch
right?
I only named the pitches tab because for ux reasons. I believe this would be a good place to put those (with buttons to extend a detailed editor)

[Chalisque]

With regards to control ABCD/macros, given that changing the parameter order breaks automation, perhaps it's worth pre-allocating 'more than you need', e.g. 16 or even 32 controllers at the start of the parameter list, all but 4 being 'macro reserved' that do nothing. So that you can then decide to increase the number of controllers later without breaking anything (or not), but crucially you're not committing to the long-term number of controllers at this early stage. The other thing about having e.g. 32, is that it is then clear to sound designers that they're not expected to populate the entire macro list, as happens with e.g. 4 or 8 macros. A separate idea to ponder is adding 'virtual macros' which combine controllers using expressions, e.g. controller V1 = 2*A+0.5C^0.5 or something. But the important thing at this early stage when it's ok to break things, is not to commit yourself to the absence of such things at such an early stage.

[Funky40]

With regards to control ABCD/macros, given that changing the parameter order breaks automation, perhaps it's worth pre-allocating 'more than you need', e.g. 16 or even 32 controllers at the start of the parameter list, all but 4 being 'macro reserved' that do nothing. So that you can then decide to increase the number of controllers later without breaking anything (or not), but crucially you're not committing to the long-term number of controllers at this early stage.

Yes,

please !

[JoeLowery215]

Thanks Urs. Is it possible to add a perfect sine shape to the toolbox?

Not mathematically possible. Our sine accuracy exceeds that of any paper I found about reproducing a sine with cubic Bezier splines.

You can theoretically get closer by using more than 2 curve segments. But I did not bother, because 80+ dB is pretty much inaudible already, and beyond CD quality.

Oh that's fine I don't care about the extra harmonics...I just meant the shape you currently have is slightly bent so it would be nice to just have a perfect curve (or whatever you would call it) to draw in Sine shapes

What do you mean by "slightly bent"? A triangle shape can turned into a sine that isn't bent at all.

Like this!

[Urs]

Hehehe, that Shape is a raised Cosine. It's like a perfect sine when you draw just one of those across the entire waveform. It's just offset by a quarter.

However, in the upper row of shapes, the second one is a quarter sine. 4 of those make for a pretty good sine, no bends required.

[Sound Author]

As someone who's tried to create an elegant, lightweight sine function within Reaktor Core, I'm actually very impressed that this is even possible just with bezier splines. Sure, there' a few lower-order harmonics clunking around at (checks notes) -80dB? That's actually pretty remarkable.



You'd think a sine function would be simple, right? Well, mathematically, it is! ...Until you enter into the digital domain where oscillators advance phase in discrete steps, which is why lookup tables are commonly used to get as close as possible.

[Urs]

As I said... we had some reference stuff about sines made from Bézier curves, which had a lot less headroom... I think we found the best possible curve parameters...

[sunthief01]

hello all, just checking in about status of the linux build for the most current revision. excited to get it when it's ready, any news? Zebra 3 (and Uhbiks) + Bitwig 6 on Fedora Kinoite has been amazing thus far.

[JoeLowery215]

Hehehe, that Shape is a raised Cosine. It's like a perfect sine when you draw just one of those across the entire waveform. It's just offset by a quarter.

However, in the upper row of shapes, the second one is a quarter sine. 4 of those make for a pretty good sine, no bends required.

Ah there it is! Thanks that works beautifully.