zebra xt

[steff3]

hello,

so I download the classical waveforms and plenty of waveforms -> I think those belong into the Oscillator's directory? And are loaded as oscialltors .... right?

But Zebra just dies and Logic with it .....

a) these are Oscillators for getting load into oscs? Or are they presets (loaded with the preset browser)?

b) Do I need to install the latest beta to use these?

Thanks

best

[justin3am]

I can't say weather or not it's the reason why Logic is crashing but I'm using Live in OSX and I needed to use Z2b1 to be able to use these.

[jupiter8]

hello,

so I download the classical waveforms and plenty of waveforms -> I think those belong into the Oscillator's directory? And are loaded as oscialltors .... right?

But Zebra just dies and Logic with it .....

a) these are Oscillators for getting load into oscs? Or are they presets (loaded with the preset browser)?

b) Do I need to install the latest beta to use these?

Thanks

best

Download the latest beta and you'll be fine. The waves are too big for the old Zebra2. That has been fixed in the latest beta.
Discovered by yours truly.I rule.

EDIT: And yes you load them with the oscillator browser.Though i believe you could load them as presets too but i have'nt tried that so i'm not sure. But in theory that should be possible.But it's easier anyways to load them from the osc. browser as you don't have to look for them.They're right there.

[suthnear]

@ Vectorman re: VS waves - they're in the same format as the xt waves so if you can load the one lot, you should be able to load the other. Since Howard has raised issue with them (but without providing me the promised examples ) I know that they work for at least someone other than me. So I'm not sure why you can't load them, sorry.

@ Howard - well?

@ Invect re: missing wavetables - I built the wavetables from the xt's sysex data. The xt I used for this only returned sysex data for the wavetables that were attached here. I tried to get the tables in question but no data was returned. Hence no tables.

[Vectorman]

Thanks for the reply, suthnear. Hopefully the issue I had with loading the waves from the originally posted ZIPs will resolve itself for me when the next major Zebra update comes out and/or when I feel that Mackie Tracktion 3 has become bug-free enough to update to (in the slim chance that something in T2 is causing the problem). In the meantime, I do have the VS waveforms available to me in the Korg Wavestation VSTi, so it's all good.

Do you think using the Wavestation VSTi is "cheating," since my Wavestation hardware synth is what I "retired" to get the Zebra discount? Hehehe...

[Howard]

@ Howard - well?

Yes, very well thank you... ah, you meant something else Those wavessets. All the waves that look like they should be sines are buzzy, and I hear a similar buzz from other waves (i.e. all the "purer" ones). Do they really sound like that straight out of the VS? In comparison, the Augur versions are very clean... it would be nice to have those in Z2 .

Example MP3 here: http://www.box.net/shared/7y543m7afc
Although the MP3 encoding adds a bit of sizzle, the third note is obviously a whole lot buzzier than the other two!

BTW you can download the free Augur DLL here (its tiny): http://users.tkk.fi/~ajhuovil/Augur.dll. A good but very unfinished Prophet VS clone by Antti of Asynth and John Bowen "Solaris" fame, among others. Both Asynth and Augur show that he knows his onions, so I was assuming that his wavetables are 1:1 copies of the VS tables.

[suthnear]

Howard,

I've spotted a few zero crossing errors, the source of the slight buzziness in some waves. I'm not sure what's causing it, though and I'll have toi have another look at the data. It's almost certainly either a boundary condition violation or an out by one error. I'll definitely fix this and repost the waves hopefully later tonight (but, depending on the difficult of finding and fixing the error, perhaps only on the weekend).

This notwithstanding, I have compared the z2 vs waveshapes with the evolver's vs waves and they're identical in appearance. Aside from being in anti-phase, Augur's waveshapes also look the same. They definitely sound different, though. This is almost certainly down to z2's sound engine as there is definitely a sonic similarity between the xt and vs waves when played through the geoblend.

I also found a bank of vs waves that had been converted to the microwave and they're somewhat different again. I'll also post those later for the completionists

[Howard]

I've spotted a few zero crossing errors, the source of the slight buzziness in some waves.I'm not sure what's causing it, though and I'll have toi have another look at the data. It's almost certainly either a boundary condition violation or an out by one error. I'll definitely fix this...

Yeah! .

I have compared the z2 vs waveshapes with the evolver's vs waves and they're identical in appearance... Augur's waveshapes also look the same. They definitely sound different, though.

Just one practically invisible "vertical line" can make quite a buzz, so I don't think we should trust appearances.

This is almost certainly down to z2's sound engine as there is definitely a sonic similarity between the xt and vs waves when played through the geoblend.

Both have also been loaded into WavToZ2, so personally I'd tend toward "definately maybe".

Thanks for the great work, jupiter8 and suthnear. Very much appreciated...

[suthnear]

Agreed about appearances, but visual similarity is a sanity check as to the essential correctness of the approach. I've found the error - the very last entry in some waves is not the same level as the very first entry and this creates the slight buzzing. However, it's not as simple as just ensuring that these two entries are equal. I need to see why they're not lining up in the first place. So I'll have to take a very careful look at the data generation process. Hopefully I'll have time tonight

[suthnear]

Ok, I've now spent a few hours studying the data and the conversion is being performed correctly. I also examined a few of the waves that were exhibiting the buzzing and it is definitely being caused by the beginning/end of the table. If I manually change the beginning and end value to be the same the buzz is either eradicated or substantially lessened.

I've picked out the data for a few of the tables that exhibit the buzz (all from vs_waves 01):

Wave[0] = 0.0;
Wave[1] = 0.25244140625;
Wave[2] = 0.34765625;
Wave[3] = 0.4775390625;
...
Wave[125] = -0.4775390625;
Wave[126] = -0.34765625;
Wave[127] = -0.25244140625;
Selected.WaveTable.set(4, Wave );

Wave[0] = 0.0;
Wave[1] = 0.62890625;
Wave[2] = 0.47509765625;
Wave[3] = 0.62158203125;
...
Wave[125] = -0.62158203125;
Wave[126] = -0.47509765625;
Wave[127] = -0.62890625;
Selected.WaveTable.set(5, Wave );

Wave[0] = 0.99951171875;
Wave[1] = 0.94580078125;
Wave[2] = 0.8125;
Wave[3] = 0.6630859375;
...
Wave[125] = 0.6630859375;
Wave[126] = 0.8125;
Wave[127] = 0.94580078125;
Selected.WaveTable.set(6, Wave );

In each case you can see that wave[0] and wave[127] have differing values. When I set these to the same value, the buzzing mostly goes away. However, the difference between wave[127] and wave[0] is, in each case, also the same as the difference between wave[0] and wave[1]. So it stands to reason that this difference should make no difference.

I therefore think that there might be a boundary condition error in the code that deals with looping...

[Howard]

Thanks Suthnear.
Yep - Urs better have a look at that before too many people use this osc mode...

[suthnear]

I thought about writing a smoothing routine but quite aside from the practical impossibilities (i.e. what works for some waves messes up others), I didn't convert these to get perfect reproductions. I was mainly interested in splicing a little of the genetic data from a synth I loved and used years ago into a synth that I love and use now. A mutation rather than a clone

[Howard]

I didn't convert these to get perfect reproductions...

That's the "glass half full" way of looking at it. As a "glass half empty" person, I'd much prefer it to work the way it "should"!

[Urs]

Hmmm... weren't the Microwave wavtables somehow "mirrored"? Like, a wavetable represents only half the wave while the second half is identical but reverse and upside down?

Could that be part of the problem?

[suthnear]

Yes, that's correct, but I compensated for that with the microwave waves. The problem is with the prophet vs waves, which don't share this very waldorfian idosyncracy...