Zebra2 Latest Version: 2.1 - Jan 19 2007

[Urs]

I think we could reproduce that with a zebra I'll try prove this tomorrow !

No need to prove it, I've already done that - I think it's in one of the 20 or so demo songs I've made so far... maybe I shouzld save the preset then...

Urs

[fgimian]

I think we could reproduce that with a zebra I'll try prove this tomorrow !

No need to prove it, I've already done that - I think it's in one of the 20 or so demo songs I've made so far... maybe I shouzld save the preset then...

Urs

[pdxindy]

I think the most powerful feature of Zebra that I don't think i've seen in many other synths (except Absynth but WAY less powerful) is the comb filtering. Those things can work magic for ambient stuff !

The comb modules are like a secret box of sound! I have never heard any synth as good for ambient soundscape stuff as Zebra2. Especially generative non-repeating soundscapes.

But then there's the Oscfx. The Oscfx are a whole field of exploration. The filter has such a liquid quality.

1 Osc and 1 Comb can create such complex timbres, with lovely dissonances/resonances, that evolve, take unexpected turns. Zebra2 so easily has rich, organic sounds that for me readily connect to inner feelings.

I've also come to appreciate the stereo path and how open and dimensional it makes the sound. Turn off all effects and one can still get a 'living' sound

Okay... I like Zebra2

[gbles]

Urs, I have something odd going on with user LFOs set to Lines in the latest build. I didn't notice it in the last build.
Listen to this preset:
http://dieterworld.com/files/LFOWeirdness.zip

Take a look at LFO4. For some reason, when it hits the second step there is a sudden jump in the filter cutoff (to which it is assigned). If you switch to Steps mode it does not do this.

I included another preset in that zip that is causing some problems. Its the OneOSCOnly patch. For some reason, it crackles like crazy. This is NOT unique to this build. I'm just doing some fast LFO modulation of the OSC wave - its a very basic patch.

[pdxindy]

Is the lowpass on the delay working correctly?

If I set up a stereo delay with feedback full down (no x-back), there is one delay iteration. Set mix all the way up to hear just the delay, Now turn the lowpass all the way down. It is not possible to cut out the high frequencies very much on that one iteration.

Turn feedback all up (use the highpass at about half to prevent boomy feedback on bass frequencies) and you still get 8-10 audible delays with lowpass all down.

I would think the lowpass should have more effect



also, interesting behavior... same setup - set lowpass full up, feedback full up, and highpass full up, hit a note and you will get a few audible iterations before fading out, then it will slowly return (VHF) getting louder and louder

[pdxindy]

Another question

Take a default osc, turn normalize up and use oscfx filter. Then turn it down... at -98 it has plenty of sound, at -100 it shuts off... even on a high note

So is this because the oscfx filter has some key follow built into it? So that no matter the note it will not filter lower than its fundamental frequency?

[luCiPHer]

Urs, I have something odd going on with user LFOs set to Lines in the latest build. I didn't notice it in the last build.
Listen to this preset:
http://dieterworld.com/files/LFOWeirdness.zip

Take a look at LFO4. For some reason, when it hits the second step there is a sudden jump in the filter cutoff (to which it is assigned). If you switch to Steps mode it does not do this.

I included another preset in that zip that is causing some problems. Its the OneOSCOnly patch. For some reason, it crackles like crazy. This is NOT unique to this build. I'm just doing some fast LFO modulation of the OSC wave - its a very basic patch.

the user lfo behaves like random glide before urs fixed that, so i guess he'll be able to fix that soon


if you reduce the resolution the crackling on the other patch goes away, but you can turn the lfo-speed up to 1/64, and no crackling will happen, so it can't just be the speed of the modulation that causes the crackles, can it?

[fab m.]

Hi Urs, I sent you an email message twice but you're probably quite busy with the final decorations on your masterpiece
So I'll express here my congratulations for the work you've done, I wish you the best!

[Urs]

Urs, I have something odd going on with user LFOs set to Lines in the latest build. I didn't notice it in the last build.
Listen to this preset:
http://dieterworld.com/files/LFOWeirdness.zip

Take a look at LFO4. For some reason, when it hits the second step there is a sudden jump in the filter cutoff (to which it is assigned). If you switch to Steps mode it does not do this.

I included another preset in that zip that is causing some problems. Its the OneOSCOnly patch. For some reason, it crackles like crazy. This is NOT unique to this build. I'm just doing some fast LFO modulation of the OSC wave - its a very basic patch.

The LFO thing is definately a bug. Looks like I forgot a line of code when fixing the user wave lines mode

The osc thing is weird. I have to analyse this a bit more...

Later,

Urs

[Urs]

Another question

Take a default osc, turn normalize up and use oscfx filter. Then turn it down... at -98 it has plenty of sound, at -100 it shuts off... even on a high note

So is this because the oscfx filter has some key follow built into it? So that no matter the note it will not filter lower than its fundamental frequency?

Yes, that's it.

The filter is not a normal filter, it's a spectral curve that's applied to the wave spectrum when the osc calculates its wavetable. The curve is stretched to the number of audible partials (between fundamental and 17kHz) while the shape is modified by the fx parameter (it's a steep one-octave fade, like a 100+ dB/octave filter, nothing you can get in normal filters). The lower 10% don't do anything but fade out the fundamental. Then, as long as there's *some* energy on any partial/fundamental, Normalize will push that to 0dB RMS level. If there's nothing left, Normalize does nothing and thus it suddenly drops.

Cheers,

Urs

[Urs]

I would think the lowpass should have more effect

Hmmm, yeah... gotta look into this... it's only a 6dB filter... maybe we need a switch for *more*...

also, interesting behavior... same setup - set lowpass full up, feedback full up, and highpass full up, hit a note and you will get a few audible iterations before fading out, then it will slowly return (VHF) getting louder and louder

Well... full on feedback is about +3dB or so, so it's actually more than 100%. Which is cool, because the delay has a nicely working limiter in the feedback path that inaudibly tames the delay in 99% of whatever audio material goes through it. But sometimes some stuff builds up with phase shifts and short spikes or something that tricks out the limiter, so it may do unexpected things...

Urs

[Urs]

Hi Urs, I sent you an email message twice but you're probably quite busy with the final decorations on your masterpiece
So I'll express here my congratulations for the work you've done, I wish you the best!

Hey fab,

Thanks man!

Sorry, I wanted to answer your email, but it came in at times where a lot of customer emails had to be answered first... this is indeed a time where a lot of things go over my head

Later,

Urs

[raikard233]

I've tried today the 03_08_06 beta and i'm because there isn't no more loop bug with Fx use!!! (Amd X2 3800 with FLStudio 6.0.8 and 6.3beta)

Thanx Urs!

[fgimian]

When I change locator position in my project, effect tails are cut off.

D'oh... oh...

Hmm, the whole logic was added for only one host (I think it was ACID, not sure). I'll check if I can make it so intelligent that it uses whatever playback start/stop detection suits the host best....

hey man, I should mention that when moving locator position I also get a CPU spike that hits around 40 % higher than what the synth is using. I am currently trying out Zebra2 in the context of a track, there are 7 Zebras so far and this problem unfortunately now causes my CPU to move into the red when I move locator position and thus I'm getting a huge amount of dropouts and crazy CPU shit going on

If you could disable this behaviour in a version, I can double check if the CPU spike is also removed too.

I'll show ya the track when I get a little more into it too!

Cheers
Fots

[pdxindy]

Another question

Take a default osc, turn normalize up and use oscfx filter. Then turn it down... at -98 it has plenty of sound, at -100 it shuts off... even on a high note

So is this because the oscfx filter has some key follow built into it? So that no matter the note it will not filter lower than its fundamental frequency?

Yes, that's it.

The filter is not a normal filter, it's a spectral curve that's applied to the wave spectrum when the osc calculates its wavetable. The curve is stretched to the number of audible partials (between fundamental and 17kHz) while the shape is modified by the fx parameter (it's a steep one-octave fade, like a 100+ dB/octave filter, nothing you can get in normal filters). The lower 10% don't do anything but fade out the fundamental. Then, as long as there's *some* energy on any partial/fundamental, Normalize will push that to 0dB RMS level. If there's nothing left, Normalize does nothing and thus it suddenly drops.

Cheers,

Urs

For a few months I was using Zebra2 and didn't pay much attention to the oscfx filter, just assuming it was a less capable version of the filter module. Then I discovered how unique and good it sounds. I get some really liquid sounds with it. The oscfx are like little hidden gems.