I'm glad you finally got some good use out of it. As I said in the last video, Zebra was never intended to be a traditional wavetable synth. Urs simply wanted to streamline the basic concept of _drawing_ wavetables, and while I've also definitely said to myself "Who's got time for that", I've found the new wave editing tools to be very useful and _time_ _saving._ Manually drawing waveforms in Zebra 2's editor was...well...less than ideal, LOL. I still recommend these video's on Urs Heckmann's personal TH-cam channel... "Zebra3 Spline Editor Part 1: Curve Editing" "Zebra3 Spline Editor Part 2: Morphing" "Zebra3 Spline Editor Part 3: Guides and advanced editing" "Zebra3 Spline Editor Part 4: Oscillator" 👆This is _much_ better documentation than what's in the manual imo.
Howdy, @SoundAuthor! Thanks for watching once again and for your tip about the fact that the osc editor could, in fact, import .wav files! I did, in fact, check out those videos before recording this one, though I've not messed with the "guides" yet (though that is a very clever feature). They were illuminating for sure. I'm sure that there are probably some helpful hints in the Zebra *2* docs that might help understanding Zerbralette 3 further. But the current Z3 manual (which I know is a total work-in-progress) doesn't make any mention of the fact that the drag & drop file functionality even exists. So this is truly a "hidden" feature (and quite a gem in my opinion). Aside: I enjoyed Urs's derpy/chatty demeanor in those videos and they sort of remind me of someone, but I can't quite recall who... 🤔
Can't reply to any of the below messages...weird. I cut WaveEdit out of the equation and just decimate to Modwave. easy peasy-done. ZL3 is interesting for taking in 33 slice Audioterm and let it coin toss what it gives as a Modwave result. ZL3 is now part of my toolkit, along with Audioterm and WaveEdit. mix-match-experiment-have fun!
6:57 For those who are just looking for the answer to "How to Import Waveforms to Zebralette 3 Beta". Somehow it cannot be dragged from Bitwig's file explorer. It can only load files dragged from windows explorer.
9:30. There are actually several morphing options, including simple cross fade - which is what you are looking for in this example, to preserve, the bell-like quality. User guide page 30 🙂
@@howardscarr9518 thanks for watching and for the tip! Note that the manual was VERRRY draft at the time of this recording! I haven’t revisited Zebralette in a while.
I have noticed that I can drag full wavetable in Serum format rendered in Audioterm into the oscillator and it gives me 10 different single cycle waveforms. Neat.
@SouthShoreSonics, it looks like perhaps I hadn't tried that! Yeah, it does seem that you can drag and drop "Serum style" wavetables into the editor window. The rules about what it will do with them seem weird. It seems that if the wavetable has a small number of frames (like 2 or 4) only 1 gets imported. If I drop an 8-frame wavetable into it, I see it render 8 or 10 points. Wavetables with 64 frames seem to turn into 10 (probably even-spaced) frames from the file. This bears a lot further scrutiny! I guess this is going to turn into a whole mini-series on Zebralette 3...
Looks to me like this feature (being able to drag and drop wav wavetable files) is still in process. BUT, when you drag and drop a wavetable with 2048 samples-per-frame into the OSC editor, the "Import" section displays with 3 options (Tolerance, Max Points, and Max Frames, which can go as high as 16). It doesn't look like it works entirely properly (for example, I created some wavetables with 10 frames and tried to drag and drop them to the window with 10 set as "Max Frames" and got 9 curves instead. Also, seems to crash the plugin sometimes. But clearly this is something that is being developed. Guess I should supply some feedback over in their KVRaudio forums dedicated to this thing. Also: It seems to pick the frames to be converted not in an evenly-spaced way, but perhaps from the start or end of the wavetable. But like I say good to see that this seems to be a work-in-progress feature!
@@kcrosley Also I can sometimes drag wavetables into the MSEG slot as well, but most of the time I lose the animation of the oscilator...but I am still experimenting.
Also note that Zebralette 3 brings in only about half the morph into WaveEdit as it expects 64 slices and has no way of decimating besides OPEN TO FIT, which really mangled the sound. Modwave as you say, decimates in a much smarter way.
Missed this comment or forgot to respond to it: Yeah, banks in WaveEdit are strictly 64 frames. You could import individual waveforms though. A script to convert 101-frame ZBL wavetables to 64 frames would be trivial, BTW.
Hey @genuinefreewilly5706! Yeah, I strongly suspect that the Zebralette oscillator tosses high-harmonic content over to a shaped noise source or something. It's... an interesting choice. You can definitely hear in this video the difference between Z3 and the crystal-clear rendering of Korg modwave after we export a wavetable from Z3 and use it in modwave. It's pretty striking for those with ears to hear it.
@@kcrosley Its still in beta, and I am still figuring this thing out and other U-He vsts. When I render a track to audio in high res some of the artifact disappears somewhat. Also if I apply ''Crammit' I can sift out some more musical harmonics. (I'll try anything) At any rate I am not unhappy with Z3 :)
A lot of interesting points, although the bell like sounds aren't sounding like that, just because of the waveforms, but also because of the volume envelope. This makes me curious what pad, lead and string sounds will sound like. 😄
Hey @Hyper5nic, thanks for watching and listening! I'm referring to the timbre, not the amp on the patch. To hear what the polygonal series waveforms sound like, played back with high-fidelity in modwave, check out th-cam.com/video/lYjwlKiUjK0/w-d-xo.htmlsi=_N66vGVq5zWvIz8K&t=264 There's something a lil' funky going on with Zebralette's rendering of those types of sounds. (I think I made this clear in my video, but if not: The polygonal series waveforms have harmonics at very specific intervals. For example, the 6-sided polygonal waveform has harmonics at 1 (the fundamental), 6, 15, 28, 45, etc. in a 1/n relationship. And they have zero amplitude at any other harmonics. It's this particular and peculiar harmonic relationship that gives them their sound. Modifying those waveforms in the time domain - even slightly - changes those harmonic relationships and makes them lose their "magic".) What I'm hearing in Zerbralette's rendition of those (imported) waveforms is a lack of the highest harmonics and a bit of harmonic "fuzz". This may be due to inaccuracies in how the exact time-domain shape of the waveform is being approximated by Zebralette's spline tracing, or by the nature of the oscillator itself. I suspect it's that noise is a component of the spectral oscillator, but only U-He could tell us for sure! 😉
Yes, there is, but I’m not sure you could get exactly the same effect as a time-linear morph. My point was more pointing out the distinctive sound this sort of shape morphing gives, vs time-linear crossfading. (And it’s that idea that lead me to create the Bezier wavetables, which do have a unique sound to them.)
Hey @AnkitWP, thanks for watching! I've sort of thought the same thing (about data size for wavetables), though the interpolation approach of ZBL 3's spline oscillator gives very different results from time-linear interpolation/crossfading. I've mused in other videos of mine about other potential oscillator models such as having a set of onboard VAE models and then the wavetable is defined by a list of points or a control curve through that VAE's latent space. In my own wavetable collection, none of the tables has more than 64 frames (because of how modwave works), which helps keep individual file sizes manageable (though the entire collection is many gigabytes in size).
Yo Keith a few comments; 1) that first waveform you are showing is not a single cycle; it's dual. 2) you don't have to create an empty waveform, then copy from other waveform and past into empty waveform. This can all be done in one step; select waveform to copy from, right click into timeline and select duplicate. At some point (around 12 minutes in I believe) you speak of a dip creating resonances (moving a group of points in waveform down). That's not a resonance buddy! A resonance is a peaked spectrum (peak with neighboring frequencies amplified). Moving one or more points in waveforms sharply up/down creates stronger set of high(er) frequencies. 3) You _can_ import a long waveform that Z3 chops up into ever so many waveforms in the timeline. It does this by using some smarts to figure out what the relevant parts of that waveform are. Kinda cool but not always, I wish that process could be more deterministic. Lastly; on selling your waveforms; Waveforms can be grouped by just a few similar sounding categories. Different looking waveforms often do not sound much different (because their frequency content is not much different). I find that keeping a relatively small group of well understood waveforms on hand and doing most of the shaping with modulations, filters etc. creates more life-like sounds than endlessly ponder over a vast set of waveforms.
Thanks for watching and for your comments, @ItsOK-vu9gx. I feel like you're confused on a couple of points and also being a bit pedantic here. Responding to your comments: (1) Yes, that first waveform is one of my polygonal waveforms (I think it's pentagonal/5-sides) that has a base frequency of 2 and so in fact does "replicate" the base 1 waveform twice within this particular waveform. I computed the polygonal (and similar with base frequencies of 1, 2, and 3, different numbers of max harmonics, different offsets and then these are used in different wavetables. "Single-cycle waveform" in the context of the frames of a wavetable has absolutely nothing to do with the number of zero crossings, base frequency, or whether any portion of it is repeated or not. A single-cycle is what fits into the allotted waveform size. In modern wavetable synthesizers (and in the example shown here), the waveform is 2048 32-bit samples. It is, in fact a single-cycle waveform. (The fact that it's essentially the base 1 version pitched up an octave is perceived as a *timbral* difference, not so much as a pitch (though layered on top of the base 1 version would be a very similar sound to taking the base 1 and pitch shifting it up). For examples that make this very clear, see my video on Morlet wavelets - th-cam.com/video/MS8YQ4bX1fM/w-d-xo.html (2) We can't take a normalized waveform like this and push some points of its amplitude past the maximum value now, can we? No, we cannot. So I am pulling a section of the waveform "down" toward zero amplitude which creates two resonances -- one to the left and one to the right. And you can hear what I'm talking about. Sheesh. (3) Whatever, it doesn't do that well. What I'm talking about here is that we can take *de facto* standard 2048 sample waveforms (as would be found in the frames of a modern wavetable) and they import properly. ZBL3 is very very much a beta and it's evolving and I'm sure there will be more sensible waveform and wavetable import options in due time. Also, I generally don't like when people import a long sample as the source for a wavetable. Why not just use a freaking sampler in that case, right? I make wavetables with math. (4) I don't really understand your point about the "look" of waveforms versus their sound. In general, my waveforms and wavetables are made with mathematical processes in the *spectral* domain and so the time-series representation of the waveform isn't of much concern to me, except in the case of things like my Bezier wavetables (which were inspired by ZBL3's approach to waveform deformation -- learn about those in my original video about Zebralette here th-cam.com/video/N8G6TJDcoZo/w-d-xo.html) and are explicitly the result of a *time-domain* technique. I guess you're one of those people who thinks that you *can* have too many waveforms and wavetables. I'm of the opposite persuasion. Nobody's putting a gun to your head and forcing you to buy KRC Mathwaves. Only cool people buy, like and use KRC Mathwaves. It's not for everybody. It's OK to be in the uncool camp. (But, you are missing out. And I can feel your FOMO rising.)
@@kcrosley Well mr Keith sorry you felt "spoken to", develop a little robustness I'd say and a mind to learn something. Respecting your space, I won't clutter up your channel with comments again.
I'm glad you finally got some good use out of it. As I said in the last video, Zebra was never intended to be a traditional wavetable synth. Urs simply wanted to streamline the basic concept of _drawing_ wavetables, and while I've also definitely said to myself "Who's got time for that", I've found the new wave editing tools to be very useful and _time_ _saving._ Manually drawing waveforms in Zebra 2's editor was...well...less than ideal, LOL. I still recommend these video's on Urs Heckmann's personal TH-cam channel...
"Zebra3 Spline Editor Part 1: Curve Editing"
"Zebra3 Spline Editor Part 2: Morphing"
"Zebra3 Spline Editor Part 3: Guides and advanced editing"
"Zebra3 Spline Editor Part 4: Oscillator"
👆This is _much_ better documentation than what's in the manual imo.
Howdy, @SoundAuthor! Thanks for watching once again and for your tip about the fact that the osc editor could, in fact, import .wav files! I did, in fact, check out those videos before recording this one, though I've not messed with the "guides" yet (though that is a very clever feature). They were illuminating for sure.
I'm sure that there are probably some helpful hints in the Zebra *2* docs that might help understanding Zerbralette 3 further. But the current Z3 manual (which I know is a total work-in-progress) doesn't make any mention of the fact that the drag & drop file functionality even exists. So this is truly a "hidden" feature (and quite a gem in my opinion).
Aside: I enjoyed Urs's derpy/chatty demeanor in those videos and they sort of remind me of someone, but I can't quite recall who... 🤔
Can't reply to any of the below messages...weird.
I cut WaveEdit out of the equation and just decimate to Modwave. easy peasy-done.
ZL3 is interesting for taking in 33 slice Audioterm and let it coin toss what it gives as a Modwave result.
ZL3 is now part of my toolkit, along with Audioterm and WaveEdit.
mix-match-experiment-have fun!
Heyy. Thats the 2nd time this week you break down one of the more unapproachable vsts in my arsenal. Thanks!
Awesome! Glad I could help!
6:57
For those who are just looking for the answer to "How to Import Waveforms to Zebralette 3 Beta".
Somehow it cannot be dragged from Bitwig's file explorer. It can only load files dragged from windows explorer.
I have a supernatural gift for searching for I for on a vst, and finding the one Bitwig user who uploaded about it.
Thanks for watching!
9:30. There are actually several morphing options, including simple cross fade - which is what you are looking for in this example, to preserve, the bell-like quality. User guide page 30 🙂
@@howardscarr9518 thanks for watching and for the tip! Note that the manual was VERRRY draft at the time of this recording! I haven’t revisited Zebralette in a while.
I have noticed that I can drag full wavetable in Serum format rendered in Audioterm into the oscillator and it gives me 10 different single cycle waveforms. Neat.
@SouthShoreSonics, it looks like perhaps I hadn't tried that! Yeah, it does seem that you can drag and drop "Serum style" wavetables into the editor window. The rules about what it will do with them seem weird. It seems that if the wavetable has a small number of frames (like 2 or 4) only 1 gets imported. If I drop an 8-frame wavetable into it, I see it render 8 or 10 points. Wavetables with 64 frames seem to turn into 10 (probably even-spaced) frames from the file. This bears a lot further scrutiny! I guess this is going to turn into a whole mini-series on Zebralette 3...
Looks to me like this feature (being able to drag and drop wav wavetable files) is still in process. BUT, when you drag and drop a wavetable with 2048 samples-per-frame into the OSC editor, the "Import" section displays with 3 options (Tolerance, Max Points, and Max Frames, which can go as high as 16). It doesn't look like it works entirely properly (for example, I created some wavetables with 10 frames and tried to drag and drop them to the window with 10 set as "Max Frames" and got 9 curves instead. Also, seems to crash the plugin sometimes. But clearly this is something that is being developed. Guess I should supply some feedback over in their KVRaudio forums dedicated to this thing.
Also: It seems to pick the frames to be converted not in an evenly-spaced way, but perhaps from the start or end of the wavetable. But like I say good to see that this seems to be a work-in-progress feature!
@@kcrosley Also I can sometimes drag wavetables into the MSEG slot as well, but most of the time I lose the animation of the oscilator...but I am still experimenting.
Also note that Zebralette 3 brings in only about half the morph into WaveEdit as it expects 64 slices and has no way of decimating besides OPEN TO FIT, which really mangled the sound.
Modwave as you say, decimates in a much smarter way.
Missed this comment or forgot to respond to it: Yeah, banks in WaveEdit are strictly 64 frames. You could import individual waveforms though. A script to convert 101-frame ZBL wavetables to 64 frames would be trivial, BTW.
Zebralette3 manual, Page 31 explains the current state of waveform export/import.
Enjoying your videos on Zebralette 3 I think its this one where you mentioned the 'hoover' sound. I have also run into it, and it is annoying.
Hey @genuinefreewilly5706! Yeah, I strongly suspect that the Zebralette oscillator tosses high-harmonic content over to a shaped noise source or something. It's... an interesting choice. You can definitely hear in this video the difference between Z3 and the crystal-clear rendering of Korg modwave after we export a wavetable from Z3 and use it in modwave. It's pretty striking for those with ears to hear it.
@@kcrosley Its still in beta, and I am still figuring this thing out and other U-He vsts. When I render a track to audio in high res some of the artifact disappears somewhat. Also if I apply ''Crammit' I can sift out some more musical harmonics. (I'll try anything) At any rate I am not unhappy with Z3 :)
A lot of interesting points, although the bell like sounds aren't sounding like that, just because of the waveforms, but also because of the volume envelope. This makes me curious what pad, lead and string sounds will sound like. 😄
Hey @Hyper5nic, thanks for watching and listening! I'm referring to the timbre, not the amp on the patch. To hear what the polygonal series waveforms sound like, played back with high-fidelity in modwave, check out th-cam.com/video/lYjwlKiUjK0/w-d-xo.htmlsi=_N66vGVq5zWvIz8K&t=264
There's something a lil' funky going on with Zebralette's rendering of those types of sounds. (I think I made this clear in my video, but if not: The polygonal series waveforms have harmonics at very specific intervals. For example, the 6-sided polygonal waveform has harmonics at 1 (the fundamental), 6, 15, 28, 45, etc. in a 1/n relationship. And they have zero amplitude at any other harmonics. It's this particular and peculiar harmonic relationship that gives them their sound. Modifying those waveforms in the time domain - even slightly - changes those harmonic relationships and makes them lose their "magic".) What I'm hearing in Zerbralette's rendition of those (imported) waveforms is a lack of the highest harmonics and a bit of harmonic "fuzz". This may be due to inaccuracies in how the exact time-domain shape of the waveform is being approximated by Zebralette's spline tracing, or by the nature of the oscillator itself. I suspect it's that noise is a component of the spectral oscillator, but only U-He could tell us for sure! 😉
Isn’t there a way to change how the wavetables morph from point to point?
Yes, there is, but I’m not sure you could get exactly the same effect as a time-linear morph. My point was more pointing out the distinctive sound this sort of shape morphing gives, vs time-linear crossfading. (And it’s that idea that lead me to create the Bezier wavetables, which do have a unique sound to them.)
Zebralette is the best SoftSynth Plugin is Zebralette still free to use ??????
Zebralette 3 is, AFAIK, free forever. It’s a preview of the (at some point) forthcoming Zebra 3.
This is what wavetable synths should behave like. I hate having gigabytes of wavetables included in presets.
Hey @AnkitWP, thanks for watching! I've sort of thought the same thing (about data size for wavetables), though the interpolation approach of ZBL 3's spline oscillator gives very different results from time-linear interpolation/crossfading. I've mused in other videos of mine about other potential oscillator models such as having a set of onboard VAE models and then the wavetable is defined by a list of points or a control curve through that VAE's latent space. In my own wavetable collection, none of the tables has more than 64 frames (because of how modwave works), which helps keep individual file sizes manageable (though the entire collection is many gigabytes in size).
Yo Keith a few comments; 1) that first waveform you are showing is not a single cycle; it's dual. 2) you don't have to create an empty waveform, then copy from other waveform and past into empty waveform. This can all be done in one step; select waveform to copy from, right click into timeline and select duplicate. At some point (around 12 minutes in I believe) you speak of a dip creating resonances (moving a group of points in waveform down). That's not a resonance buddy! A resonance is a peaked spectrum (peak with neighboring frequencies amplified). Moving one or more points in waveforms sharply up/down creates stronger set of high(er) frequencies. 3) You _can_ import a long waveform that Z3 chops up into ever so many waveforms in the timeline. It does this by using some smarts to figure out what the relevant parts of that waveform are. Kinda cool but not always, I wish that process could be more deterministic. Lastly; on selling your waveforms; Waveforms can be grouped by just a few similar sounding categories. Different looking waveforms often do not sound much different (because their frequency content is not much different). I find that keeping a relatively small group of well understood waveforms on hand and doing most of the shaping with modulations, filters etc. creates more life-like sounds than endlessly ponder over a vast set of waveforms.
Thanks for watching and for your comments, @ItsOK-vu9gx. I feel like you're confused on a couple of points and also being a bit pedantic here. Responding to your comments:
(1) Yes, that first waveform is one of my polygonal waveforms (I think it's pentagonal/5-sides) that has a base frequency of 2 and so in fact does "replicate" the base 1 waveform twice within this particular waveform. I computed the polygonal (and similar with base frequencies of 1, 2, and 3, different numbers of max harmonics, different offsets and then these are used in different wavetables. "Single-cycle waveform" in the context of the frames of a wavetable has absolutely nothing to do with the number of zero crossings, base frequency, or whether any portion of it is repeated or not. A single-cycle is what fits into the allotted waveform size. In modern wavetable synthesizers (and in the example shown here), the waveform is 2048 32-bit samples. It is, in fact a single-cycle waveform. (The fact that it's essentially the base 1 version pitched up an octave is perceived as a *timbral* difference, not so much as a pitch (though layered on top of the base 1 version would be a very similar sound to taking the base 1 and pitch shifting it up). For examples that make this very clear, see my video on Morlet wavelets - th-cam.com/video/MS8YQ4bX1fM/w-d-xo.html
(2) We can't take a normalized waveform like this and push some points of its amplitude past the maximum value now, can we? No, we cannot. So I am pulling a section of the waveform "down" toward zero amplitude which creates two resonances -- one to the left and one to the right. And you can hear what I'm talking about. Sheesh.
(3) Whatever, it doesn't do that well. What I'm talking about here is that we can take *de facto* standard 2048 sample waveforms (as would be found in the frames of a modern wavetable) and they import properly. ZBL3 is very very much a beta and it's evolving and I'm sure there will be more sensible waveform and wavetable import options in due time. Also, I generally don't like when people import a long sample as the source for a wavetable. Why not just use a freaking sampler in that case, right? I make wavetables with math.
(4) I don't really understand your point about the "look" of waveforms versus their sound. In general, my waveforms and wavetables are made with mathematical processes in the *spectral* domain and so the time-series representation of the waveform isn't of much concern to me, except in the case of things like my Bezier wavetables (which were inspired by ZBL3's approach to waveform deformation -- learn about those in my original video about Zebralette here th-cam.com/video/N8G6TJDcoZo/w-d-xo.html) and are explicitly the result of a *time-domain* technique.
I guess you're one of those people who thinks that you *can* have too many waveforms and wavetables. I'm of the opposite persuasion. Nobody's putting a gun to your head and forcing you to buy KRC Mathwaves. Only cool people buy, like and use KRC Mathwaves. It's not for everybody. It's OK to be in the uncool camp. (But, you are missing out. And I can feel your FOMO rising.)
@@kcrosley Well mr Keith sorry you felt "spoken to", develop a little robustness I'd say and a mind to learn something. Respecting your space, I won't clutter up your channel with comments again.