Zebralette 3 Wavetable Engine Discussion

[hakey]

Ah, okay. No great loss. If I hadn't seen it in Hive, I wouldn't have thought of it.

[C.Ray]

Hello, I think the snap to grid moves too easily compared to other editor grids IMO.. Thanks!

[jooster]

Couple of .wav import questions;
a. What are the .wave settings for best result? (sample freq, mono/st, #bits, sampled at which key)
b. Importing a long(er) .wav results in a bunch of curves filled with consecutive sections of the .wave file - nice! What is the length of each curve, 1024 samples? I didn't find this in the manual.
c. The resolution parameter has no effect on the import, just playback - correct?
d. The import toolbox shows two options right next to close button that appear to have little/no effect - are these meant to indicate straight line or bowed line connections between points?
e. When importing, would be nice if the top display could indicate a progress bar or pct done.

W.r.t. "pre listen"; currently it turns on/off the morphing but leaves OSC FX1/2 effects in place. I'd like to have a pre-listen button that could cycle through normal/morph off/morph and FX off.

[Urs]

Couple of .wav import questions;
a. What are the .wave settings for best result? (sample freq, mono/st, #bits, sampled at which key)
b. Importing a long(er) .wav results in a bunch of curves filled with consecutive sections of the .wave file - nice! What is the length of each curve, 1024 samples? I didn't find this in the manual.
c. The resolution parameter has no effect on the import, just playback - correct?
d. The import toolbox shows two options right next to close button that appear to have little/no effect - are these meant to indicate straight line or bowed line connections between points?
e. When importing, would be nice if the top display could indicate a progress bar or pct done.

W.r.t. "pre listen"; currently it turns on/off the morphing but leaves OSC FX1/2 effects in place. I'd like to have a pre-listen button that could cycle through normal/morph off/morph and FX off.

a+b: The importer detects sample pitch and slices the sample into as many cycles as there are, then selects those that seem most different from each other. Samplerate, bit depth and such should not matter. Low pitched samples have more detail than high pitched ones, naturally.

c: correct

d: Once imported you can apply the settings again and again, until the result is best. That is where those settings matter, and with tolerance they can cause the engine to use a lot of time.

e: Yes, absolutely.

Re Pre-Listen: We won't bypass the OscFX, but we have planned to add bypass switches to the OscFX themselves.

[sveinbjorn]

Yes, if you only wanted it to use Hive or Serum wavetables, then yes, it is indeed useless to you.

I hear Hive is good for that.

(sorry)

[jooster]

a+b: The importer detects sample pitch and slices the sample into as many cycles as there are, then selects those that seem most different from each other. Samplerate, bit depth and such should not matter. Low pitched samples have more detail than high pitched ones, naturally.

Ah, that explains what I am seeing - the process is less predictable than what I thought, that's fine. So if I am importing a sample with a specific waveform development I want to capture, is it best to chop it up in parts so I can designate it to whatever curve block I see fit? If so, what is the appropriate sample length for that?

d: Once imported you can apply the settings again and again, until the result is best. That is where those settings matter, and with tolerance they can cause the engine to use a lot of time.

I think I understood that but what I don't understand is what these two options do - I see little effect in the samples I imported (which may just be an artifact of my specific data).

Re Pre-Listen: We won't bypass the OscFX, but we have planned to add bypass switches to the OscFX themselves.

Great, those will be handy.

[Urs]

The idea is that you do not need to care about the length of the sample or the frequency or the cycles... it'll figure it out.

The main problem is atm that some algorithms either take a lot of memory or a lot of time or both. I have some ideas to bring that down, but once I did, it should be fine. Only, the parameters for the import will change a bit so we don't want to waste time now to document it too deeply.

[Shabdahbriah]

for me a conversion concept of wavetable files in a easy manner would be great. I did some experimenting with audacity raw file import with succes. And there is converting with moss app witch i dit not try yet. And synthesis technologie wave edit.

All these containers, riff's can there be something in zebralette 3 that can handle .wt / .wtrs /.wav / and so on
More acces to metadata.

Is there a synth other than the Waldorf NW1 (which I have) that uses .wtrs?

[Urs]

Well, .wav is the most supported, actually documented and readily available format for audio files. IMHO it's the de facto standard for wavetables. There are probably a couple of wavetable synths that don't support it, or some maybe have obscure requirements. Export for such is however beyond the scope of our work. Wavetable export in Zebralette 3 is a small bonus, albeit not an entirely altruistic one: We do it mainly because we hope to give Hive a little boost. Hive imports them easily.

But then, .wav import is not exactly rocket science either. Let those devs know that you'd love to import wavetables in Hive/Serum format into their products, and if they're reasonable, they will do it!

[kevinkvr]

[/quote]

Is there a synth other than the Waldorf NW1 (which I have) that uses .wtrs?
[/quote]

I contacted Waldorf and asked them if any of there products, hard and software were compatible.
They said no! So all there stuff is not compatible. But again Try RAW import with Audacity

P.s. LOOKING forward to al the new U-He stuff!

[kevinkvr]

@Urs
Wil there be some kind of spectral waveform sampling? And maybe a Freeze funktion? There are now some eurorack modules that can do that and it sound's great! The expert sleepers disting EX. And i also found a M4L device..spectral freeze pro
https://newfloremusic.gumroad.com/l/disWr

[Urs]

@Urs
Wil there be some kind of spectral waveform sampling? And maybe a Freeze funktion? There are now some eurorack modules that can do that and it sound's great! The expert sleepers disting EX. And i also found a M4L device..spectral freeze pro
https://newfloremusic.gumroad.com/l/disWr

We're considering a granular texture generator for Zebra 3, but nothing like it for Zebralette since it's basically a wavetable oscillators and the scope is waveform and wavetable generation.

[robisme]

Would it be possible to change a preset without loosing the sound of presently pressed notes ? I mean, a way to anticipate the preset of the next notes with a seamless transition.
Now that Zebralette comes with a morph feature, I wonder if a -limited of course- kind of morphing between presets would be an idea ?

[buing]

I think unmorph curve option would be nice.
I don't know 'unmorph' is exact word, I'm beginner for English, sorry.
When I add or edit new curve, it's always interpolated smoothly and automatically, as you know it's named Curve Morph in Zebralette.

But in terms of history of wavetable, there was no smoothed morph in very first version of wavetable synthesizers.
I guess many people imagines naturally that smoothed morphing is friend with wavetable.
However unmorphed tables still valid and useful option for these days, I also making unmorphed wavetable for my sound design still.
And still, some synthesizers such as Phase Plant or Serum offers a option for unmorph wavetable.

[yellowmix]

Export feature request (not necessarily for Zebralette 3). Flexible export parameters, all WAV since most wavetable synths can accept or convert from it. Since the engine is partial/vector based I'd like to skip using Audioterm since it decomposes it to partials then back into a wavetable, and possibly interpolating. It'll be faster and cleaner if we can get it direct from the Zebra3 wavetable engine, and it'd be great to use it as a primary and complete wavetable creation tool.

Basically:

• Number of frames,
• Sample length per frame,
• Bit-depth, and
• Sample rate (Edit: redundant with sample length, but would be part of WAV metadata).