Can someone explain wavetable morphing?

[Architeuthis]

Hey peoplez,

So, I've become interested in taking a handful of single-cycles of some instrument or voice samples and crossfading between them. So for example, take an oboe sample that is played from soft to loud, then choose, I dunno, 128 discrete cycles from 128 places along the soft-to-loud sample. Then use zebra to crossfade through the cycles to create what sounds like a resynthesized oboe crossfading between dynamics.

Can Zebra do this with its wavetable engine?

Now, let's say I do this at a bunch of different pitches. So we have 128 cycles from soft to loud from C3 to C4. Hmm, I wonder if I could simply create 13 instances of Zebra and load up the 128 cycles per pitch in each Zebra and then get a midi splitter to split each key to the corresponding instance of Zebra.

Anyway, basically I want to crossfade or somehow morph between cycles.

Edit: Zebra has 16 slots for cycles. I wonder if 16 slots would be enough to capture the dynamic range, and then Zebra would morph via highest resolution or something.

[xh3rv]

Just to toss a reply ... I think I'd prefer patching dynamics after setting how spectrum evolves across the range of pitches. I feel like I've been able to get relatively clean, precise things with dynamics through osc fx/filters, especially with fine-tuning expressive control of chiff and vibrato. In contrast getting the dynamics right for one tone, then trying to relate that across pitch range seems frustrating. To the extent one can say, this module does pitch range, that one does dynamics, another turns MIDI expression into phat oboe-esity, anyway.

Particularly, there's some distance between the physically located 'filters' created by an oboe's structure, and how that can be approximated by some configuration of filters/filter-like elements of Zebra. Trying to navigate this just on frequency metrics is (like, literally ...) convoluted. But I've been surprised that if just a subtle amount of spectral shaping is happening there, then utilizing drive and resonance to handle articulations of breath can be - not completely perfect, but much more straightforward.

'Course I probably end up with two saws and a ping-pong delay so ...

[Erkenfresh]

Well you have four OSCs each with 16 waves, so I count 64 frames you could use. With the MSEGs, you could adjust which OSC is playing at each instant of time, in effect chaining them together. Good luck!

[n0rd]

Check out "Zebra Tutorial 36 - Scanning"

[Sendy]

It honestly sounds like a sampler would suit your needs better in this particular case.

Wavetable morphing is based on the control points, they move around without ever changing order, so getting a certain sound is largely down to knowing how to build the wavetable with this in mind. IME it's more suited to geometric experimentation rather than recreating real sounds, though obviously with the rest of Zebra called into play you can get the realism if you're good.