Zebra GeoBlend waveform point count

[major_z]

Using oscillator's GeoBlend mode is quite powerful for importing custom waveforms in Zebra. However during my experimentation I find that the point count which is set to 128 is insufficient in some cases with more complex waveforms. Resolution of 512 or even higher would be much more powerful. I guess the point count is set to 128 because of the editing feature in the UI, but I see higher resolution as much more important. Your comments are welcome.

[mcnoone]

To the right of the Osc waveform is a knob called "Resolution"
Turn it to the right to increase resolution, and to the left to decrease it.
Turning to the right will increase memory usage also.
Hope that helps.

[major_z]

I do not think the 'Resolution' knob controls the number of points in the waveform. It has something to do with the wave morphing between waveforms.

[xh3rv]

Using oscillator's GeoBlend mode is quite powerful for importing custom waveforms in Zebra. However during my experimentation I find that the point count which is set to 128 is insufficient in some cases with more complex waveforms.

I think the count of points in GeoBlend is related to the number of harmonics in SpectroBlend, for reasons of implementation. What kinds of sound are you trying to import? Would it be possible to spread it across multiple frames of the OSC module? While a designed/synthetic waveform might lend itself to being spread across multiple GeoBlend frames, for something a little more irregular, SpectroBlend might make more sense (?).

[pdxindy]

Using oscillator's GeoBlend mode is quite powerful for importing custom waveforms in Zebra. However during my experimentation I find that the point count which is set to 128 is insufficient in some cases with more complex waveforms. Resolution of 512 or even higher would be much more powerful. I guess the point count is set to 128 because of the editing feature in the UI, but I see higher resolution as much more important. Your comments are welcome.

what sort of sound are you trying to import?

[major_z]

I am providing an example of one of the waveforms I experimented with, which is actually not so complex, but the artifacts are already visble. Here is a screenshot showing it in the original form:



And this is what I get when I try to import it in Zebra. I use a small script that I wrote, which does the conversion, similar to the other tools available.



Spreading across multiple GeoBlend frames could do the job if I split the data in different parts. Is that possible and how may I do that?

And I am not sure how may I convert my waveform data to SpectroBlend mode. Is there some way of doing this? That could prove to be quite interesting.

[xh3rv]

Could you just slice that waveform in a wave file editor? After that the trick is modulating the 'wave' parameter on the OSC module to sweep through the frames - this isn't sample accurate, but this is where increasing the 'resolution' can help I think.

Also, with the OSC module, a single cycle on a single frame will sort of be the limit to the accuracy of reconstructing the wave. I don't think it the OSC module was really designed to do really accurate sampling, but with that waveform it looks like something that should at least sound close (and good). Certainly people have been able to do some things along these lines, with that sort of waveform.

To import into spectrograph, if you can find any tool that spits out the spectral image of a waveform (FFT), one could do that per cycle I'd imagine. A while ago Billstei was developing a tool to do this specifically for Zebra (blueberry thing), and it was pretty interesting.

[major_z]

Could you just slice that waveform in a wave file editor? After that the trick is modulating the 'wave' parameter on the OSC module to sweep through the frames - this isn't sample accurate, but this is where increasing the 'resolution' can help I think.

Also, with the OSC module, a single cycle on a single frame will sort of be the limit to the accuracy of reconstructing the wave. I don't think it the OSC module was really designed to do really accurate sampling, but with that waveform it looks like something that should at least sound close (and good). Certainly people have been able to do some things along these lines, with that sort of waveform.

To import into spectrograph, if you can find any tool that spits out the spectral image of a waveform (FFT), one could do that per cycle I'd imagine. A while ago Billstei was developing a tool to do this specifically for Zebra (blueberry thing), and it was pretty interesting.

Thanks for the valuable information. I am going to try your ideas and let you all know the results of my findings. Thanks again!