User waveforms

[raikard233]

Here's two presets made with the Howard OSC preset:

Raikard Bass 001
Raikard Chord 001

I did a demo tune of the two presets with two Zebra2 instances, one Microtonic and a 909ride cymbal sample:

ZebraSample01

10mb
about 11 min lenght...sorry, that's the overfun coming from my midibox!!!



Will post the presets also on the "presets exchange area".

Rik

[luCiPHer]

groovy stuff

[replicator]

Anyone have any tips for getting a nice source waveform with a 128 sample cycle period? I'm kinda doing it a bit hit and miss - aer there any particular frequencies or keys that would give a nice 128-sample waveform cycle period?

Have you tried Audobe Audition? You can generate clean tones, and import waveforms which you can time stretch to a 128 sample cycle.

Hmm, some of the imported waveforms are hardly visible in the spectroblend but the waveforms output some flavour:) Maybe a particular frequency does matter. DB altso?

Replicator

[Howard]

Anyone have any tips for getting a nice source waveform with a 128 sample cycle period? I'm kinda doing it a bit hit and miss - aer there any particular frequencies or keys that would give a nice 128-sample waveform cycle period?

In Cool Edit Pro (or Adobe Audition?), get a complete cycle from your wave, open the context menu (right click?), select "Wave Properties", select the tab "File Info" in which you will see the number of samples at the bottom. Get your calculator out, divide 128 by this number and multiply by 100 (for % stretch). Open the effect "Time/Pitch, Stretch", select Constant Stretch and set mode to "Resample - preserves neither". Enter your result (rounded off to about 2 decimal places) into the "Ratio" field. Basta

[Howard]

Hmm, some of the imported waveforms are hardly visible in the spectroblend

Normalize waves (and remove DC if necessary) before saving. I suppose you could maximize the wave in Zebra2 if you forgot to do that...

[Howard]

Does anyone have a good method (workaround?) for packing several such waves into a single table?
Or isn't this possible yet?

[replicator]

Remove DC and normalize waves and before saving. I suppose you could maximize the wave in Zebra2 if you forgot to do that...

Oki, thanks. Do you know which DB user level when normalizing that gives best results before saving? +- DB range:)?

I guess you can merge several waves together in a single table within Audition, seriell or parallel. But you can merge waves together in Zebra already:) Use 2 OScs with different waveforms, or use the wave shaper within 1 OSc.

Regards

Replicator

[Howard]

Do you know which DB user level when normalizing that gives best results before saving? +- DB range:)?

I assume you're talking about DC offset? Suck it and see...

I guess you can merge several waves together in a single table within Audition, seriell or parallel. But you can merge waves together in Zebra already:) Use 2 OScs with different waveforms, or use the wave shaper within 1 OSc.

I meant loading a single Zebra wavetable with different custom waves. OK, I've sussed it out already - copy each one over to where you want it, then load the next one (because they're always loaded into Wave 1 by default). Maybe I'm still missing something obvious...

[replicator]

Howard wrote:
Remove DC and normalize waves and before saving. I suppose you could maximize the wave in Zebra2 if you forgot to do that...

I figured it out Howard..

Thanks

Replicator

[beej]

Anyone have any tips for getting a nice source waveform with a 128 sample cycle period? I'm kinda doing it a bit hit and miss - aer there any particular frequencies or keys that would give a nice 128-sample waveform cycle period?

In Cool Edit Pro (or Adobe Audition?), get a complete cycle from your wave, open the context menu (right click?), select "Wave Properties", select the tab "File Info" in which you will see the number of samples at the bottom. Get your calculator out, divide 128 by this number and multiply by 100 (for % stretch). Open the effect "Time/Pitch, Stretch", select Constant Stretch and set mode to "Resample - preserves neither". Enter your result (rounded off to about 2 decimal places) into the "Ratio" field. Basta

Wicked stuff, thanks for that H, will give that a shot...

[beej]

Ok, I'm getting some odd behaviour with this. I've just done a bunch of waveforms to get a hang on the process and work out the best tools/workflow etc. The resultant waveforms end up in Z2 ok and look ok (ie as you select the different waveforms in the browser, the waveform display changes and looks right), but often the sound doesn't change while I select them, whereas this works ok with the regular provide init waveforms.

If I go back to some of the preset ones they work as well and change the sound, but these ones generated by the utility seem a little flaky.

If I come out of the waveform browser and toggle to spectroblend and then back to geoblend, then the waveform sounds as it should, so I'm guessing it's a minor Z2 niggle that Urs would need to take a looksee at, I think...

[beej]

I meant loading a single Zebra wavetable with different custom waves. OK, I've sussed it out already - copy each one over to where you want it, then load the next one (because they're always loaded into Wave 1 by default). Maybe I'm still missing something obvious...

Howard, at the bottom of each h2p file there is this line:

Selected.WaveTable.set( 1, Wave );

The 1 corresponds to slot 1 in the oscillator's wavetable.

So if you have 16 files, in a text editor just number them:
Selected.WaveTable.set( 1, Wave );
...
Selected.WaveTable.set( 16, Wave );

...and they will load into the appropriate slot when you select them in the osc preset browser.

[Howard]

...If I come out of the waveform browser and toggle to spectroblend and then back to geoblend, then the waveform sounds as it should, so I'm guessing it's a minor Z2 niggle that Urs would need to take a looksee at, I hink...

Confirmed - I've had that problem as well.

Howard, at the bottom of each h2p file there is this line:
Selected.WaveTable.set( 1, Wave ) ...

Thanks beej!

[beej]

Ok, here's my first test oscillator bank.
This consists of two sets of waveforms derived from impOSCar, with 14 waveforms in each set.

Triangle
Saw
Square
Pulse 50%
Pulse 65%
Pulse 80%
Pulse 99%
PWM 50%
PWM 80%
PWM 99%
Harpsi
Gritty
DoublePulse
StrongLead

In the first set, called "imp_nodrive.h2p", the oscillators are 'sampled' with the default amount of filter drive. The second set "imp_fulldrive.h2p" are 'sampled' with the filter drive on full, which is a bit fatter.

Don't expect the things like PWM to sound like impOSCar's excellent PWM, as the motion of the software is more sophisticated and can't be captured in a static single-cycle sample. You'll have to buy impOSCar to really sound like impOSCar...

Just treat these as extra waveforms and have fun. I've just been playing with these in a two dual-osc dual XFM patch, and they sound pretty darn good! I can't compare them with the impOSCar waveforms I hand tweaked the hard way, as I lost them in an HD crash though...

Grab them from: http://bj.angeltowns.com/files/z2_imposcillators.zip

Unzip them and put the two h2p files in your Z2 oscillators directory.

Thanks to FEV for the utility - if you could add directory recursion to process a whole directory full of samples into h2p files, that would be super-handy!

Oh, and for the resampling waveforms to 128 samples thing - my old copy of TC's Spark has an excellent feature. With a single cycle selected, go to timestretch, it tells you it's 142 sample long - you change it to 128, and it's done. And that 128 value stays selected, so you can batch process a whole bunch of waveforms into a 128 sample length, and normalise them at the same time. Very nice! And no calculator required!

[raikard233]

Thank you for the OSC supply Beej!
I will try them immediately!
As for the Howard OSCs if i do some presets i will share them with you all!

Rik