Improvements!

[pdxindy]

I wish there was a voices option of 24

16 at a voice stack of 3 only gives a polyphony of 5... that is not enough and it is easy to hear the voice stealing

24 would provide 8 voices with stack of 3 and that would be much better... Right now I have to do some clumsy workaround with 2 instances. Why create such an arbitrary limitation???

[Gonga]

It's enough for polyphonic lead work, which is a lot more voices than I ever had on my monophonic lead analogs, so maybe it was a well-thought-out compromise for v 1.0

[trance_lucent]

I have managed to read whole thread!
Here is my propositions. It's not that it all of paramount importance, Zebra as it is now is already best software synthesizer, but I hope Urs like something.

1. Good distortion models (may be as separate modules). I see many people here want it.

2. Virtual analog oscillators (as separate modules or as option for current wavetable oscs). In example, Band Limited Impulse Train integration

Yes, I know, it differs from perfect current saw (for example, saw) only by color, but as I can see people often claim that plain Zebra oscs are too uncolored and weak.
Once I experimented in csound, trying to achieve Virus-like waveforms using blit and different filters and EQs. And a learned that sometimes such waveforms sound more tasty than plain perfect waveforms. Also it is interesting fact that usually such blit-integrated osc has different waveform (and coloration) for different tone pitch.

Of course, right now Zebra allows use such waveforms using impulse wavetable with filters. For example, very quick sketch for illustration. 303-like oscillator. I designed it quickly without referencing to original sound. Of course, It is possible to get original tb-303 or at least its samples and design a better patch using oscilloscope and spectrum analyzer.

Patch http://dl.dropbox.com/u/57065234/303VA.h2p (switch "invert" on for pulse waveform)
Demo http://dl.dropbox.com/u/57065234/303va.mp3 (again, better distortion would be good for Zebra)

But builtin VA-oscillators (maybe with some different models, vintage and modern like virus) can be useful.

3. I know, it is frequently requested. Filter patches by text searching in names and descriptions.

4. I know, it is frequently requested. Add "ignore osc settings" option to oscillator presets navigation.

5. Several "Random" sources for modulation, random value triggered by note start. I know, it is possible to use S&H LFO with zero speed, but I don't want to spend LFO module for that.

6. An comb-filter alternative. Formant-filter module with various acoustic instruments frequency curves like in Logic Sculpture synthesizer.

7. Possibility to perform FM and/or phase-distortion by modulating any waveform by any waveform. For example, add side-chain fm-input for osc module or at least add some brighter waveforms for FM-oscillators. I know, there will be aliasing, but may be some oversampling can do the job? Or just leave aliasing problem on user's responsibility (like in dx7 AFAIK).
Also would be good option to switch off phase initialization for FMO.

P. S. Sorry for my barely understandable English.

[grandmasterbird]

I'm not sure if it's been mentioned before but the waveshaper section in Zebra in my opinion sounds weak, I think it needs more bite to it.

Also I dont have Zebra to hand but is it the spectromorph option in the oscillator waveform section? Where you can draw in your own spectral waveforms with the mouse? I think that lacks guts and needs a bit more bite as well.

Just my 2c, not knocking Zebra as I think it's awesome, just think they are aspects that make me go to other synths whereas I would like to stick with Zebra.

[Howard]

...is it the spectromorph option in the oscillator waveform section? Where you can draw in your own spectral waveforms with the mouse? I think that lacks guts and needs a bit more bite as well.

I don't know what you mean by "spectral waveform" (or "bite", for that matter).

[grandmasterbird]

...is it the spectromorph option in the oscillator waveform section? Where you can draw in your own spectral waveforms with the mouse? I think that lacks guts and needs a bit more bite as well.

I don't know what you mean by "spectral waveform" (or "bite", for that matter).

Hi Howard,

that's probably because in all honesty I dont really know a great deal about what I'm trying to talk about, although I use Zebra in my music I'm not a sound designer by any stretch of the imagination.

Ok I just read the manual and what I meant was the geoblend and spectroblend.

And I think bite was the wrong way to express what I meant, what I meant is that I wish the geoblend & spectroblend oscillators sounded a bit fuller, or meatier, less thin?

This is of course my own perception/description of the sound and my opinion is based on what I have experienced so far. I do realise however that given my level of understanding of synthesis I may be missing a trick and dont know how to get the sound I'm after.

If this is the case then U-he will know to disregard my suggestion!

[trance_lucent]

And I think bite was the wrong way to express what I meant, what I meant is that I wish the geoblend & spectroblend oscillators sounded a bit fuller, or meatier, less thin?

grandmasterbird, The fact is Zebra oscillators are perfect wavetable oscillators. Its function is perfect playback of current waveform from table without any coloration (analog or digital). And this is not a demerit, this is the merit, the base property of a sound design station - you get exactly that you want. Needed coloration you can add manually by filter or shaper etc.

But it would be good if Zebra had an "virtual analog" oscillators, as I mentioned in earlier message. It is emulation of basic waveforms as they synthesized on hardware analog synthesizers - with coloration and interesting behavior (for example various coloration on various pitch). But this is not very important thing - you can draw "colored" waveform in wavetable or even emulate it with impulse waveform and filters (I attached example in my earlier message; If somebody wants Virus waveform emulation, I can send my experimental patches).

As for manual coloration, Zebra indeed lacks of more shaping/distortion methods. While XMF or shaper module are good, it's coloration very soft and warm and sometimes I need harder saturation with more high harmonics.

But right now Zebra already capable to generate full and meaty and fat and warm and strong sound.

[Mighty Pea]

Urs, would it be possible to allow the mousewheel to scroll through the MSEG presets by hovering over it? The same way you can hover over (for instance) the MSEG's Timeunit selector and scrolling?

[Mike Thomson]

Hi,

i don't know if this has already been mentioned but it would be nice if the preset-browser of DIVA would have a category "Favourites" where only the presets, which are marked as favourite by the user, are listed.

Greetings from Germany,
Mike

[bmrzycki]

In Zebra's modmappers it would be nice if there was a "Manual" or "knob" source and when selected a knob would be added to the MMap. This knob could be midi learned, set to an XY parameter, or even modulated from the modmatrix.

The reason I'm asking for this is I wanted to create a non-linear detune shape with an MMap and then sweep the detune with an XY.

The only way I found to do this was to use an ENV as the source, set all knobs to zero but sustain, env shape to very concave. Then modulate the sustain knob in the XY. This means I use up an ENV just for this purpose and if I want to do a second shape (say a volume mix MMap) I use a second ENV. It's doable but a bit clunky.

Thanks for reading.

[jumpercable]

Hi Urs

- Ive a small request regarding Diva that I dont think many softsynth designers are doing -

U-he synths (the few I have anyways) have the brilliant option of setting different gliss lenghts for different voices for glissando. This makes every voice sound really alive. Brilliant for adding that organic "ensemble" sound to simple waveforms.

My question is this - why doesnt any synth designer seem to enable this type of option for the pitchbend wheel?

One of my pet hates of synths (maybe its only an issue in softsynths?) is that pitchbends using the pitchbend wheel are generally, well - unbelievable, unnatural sounding and cheesey. You might get the odd patch it sounds okay in, but in general I avoid using it. If the gliss option sounds so well and makes such a difference - would any synth designer be able to emulate that feature for pitchbend - so theres a slight (adjustable) delay in pitch bend data for each oscillator? It would dramatically improve the sound of pitchbend. Give whole new life to that control.

Im not sure would it sound better allowing this "delay" to happen to each oscillator or each voice of a chord. "Gliss" seems to be done per oscillator, and that works really well.

It seems so simple - Im either missing the option somewhere or theres a reason nobody does it! The uniform movement of the pitchwheel kills some really nice double stops and an organic slide up/slide down, and Im surprised nobody has really improved or developed on this situation yet.


Diva is great Urs - need to spend more time with it to really run it through its paces!

[xh3rv]

If the gliss option sounds so well and makes such a difference - would any synth designer be able to emulate that feature for pitchbend - so theres a slight (adjustable) delay in pitch bend data for each oscillator?

I might be confused what you mean by 'for each oscillator', however ...

I think the 'lag' in the modifications tab does what you want. Modulations sent through it are smoothed out over time, the knob controls the rate at which it heads towards the current (target) value. Pitchbend in the global tuning can be set to zero, and instead map PW onto 'lag', and map 'lag' to the tuning on the oscillators.

[lazerkind]

I think jumpercable wants to be able to set the lag for one of the osc's in the scillator section and not for the other, or maybe lag with different amounts per oscillator?
Right now you can only patch one source to the tuning of all osc's.
(The lag is very nice to have btw, I want that for Zebra )

[jumpercable]

If the gliss option sounds so well and makes such a difference - would any synth designer be able to emulate that feature for pitchbend - so theres a slight (adjustable) delay in pitch bend data for each oscillator?

I might be confused what you mean by 'for each oscillator', however ...

I meant each oscillator.

I think the 'lag' in the modifications tab does what you want. Modulations sent through it are smoothed out over time, the knob controls the rate at which it heads towards the current (target) value. Pitchbend in the global tuning can be set to zero, and instead map PW onto 'lag', and map 'lag' to the tuning on the oscillators.

Thanks for the suggestion - very clever - but unfortunately doesnt do what I want. With your method - the "lag" doesnt adjust per oscillator - theres only one set lag - then the control for lag under each oscillator adjusts the range of the pitchbend - so by moving the pitchwheel down, the note values of the oscillators end up having a radically different interval to what you started out with.

I just want a simple way to allow pitch bend to be non linear.

For example - with a pitch bend down - I want each oscillator to reach its defined pitch down note, but I want each oscillator to fall at a slightly different rate. Oscillator 1 could be the real value, and oscillator 2 could have its rate of descent changed.

(Course as another possibilty - it would be amazing to play a chord, use pitch bend and each midi "note" falls at a slightly different rate. This would be different than the above. Both options would be cool! ha. )

[jumpercable]

I think jumpercable wants to be able to set the lag for one of the osc's in the scillator section and not for the other, or maybe lag with different amounts per oscillator?
Right now you can only patch one source to the tuning of all osc's.
(The lag is very nice to have btw, I want that for Zebra )

If you can only patch one source of tuning I suppose this is a problem. I just want this option specific to midi pitch bend data. Its probably alot more complicated than I think.