Tone2 will release Icarus - 3D WaveTable Synthesizer

[recursive one]

Well, since the pic is already "leaked" - what does the word "HyperStereo8" in the oscillaltor section mean? It is the name of the wavetable? Or an unison algorithm?

I did bot want to post details yet but as this could be seen at the screenshot i could tell this:
It's comparable to a Virus TI Hypersaw with 8 voices and stereo spread (with adjustable detune) but works with ALL waveforms, not just a Sawtooth. There aae also other modes besides that shown in the screenshot.

The mode could be set differently for each of the oscillators.

Thanks! Now I'm really interested, can't wait to test the thing.

[cytospur]

I guess, this may be related to the direction of wavetable scanning. E.g., in Dune2 you may change the wavetable position, which means that you morph the waveform A to B, then to C, then to D etc, and if you move back, you can only do it like D->C->B->A . Smoothly going from A to D, then to B, then to H is not possible in Dune but probably may be possbile in more complex wavetable implementations.

It is possible in D2

[recursive one]

How?

I mean, morphing from A directly to D skipping B and C

[AnX]

How?

I mean, morphing from A directly to D skipping B and C

Using the MSEG's maybe?

[cytospur]

How?

I mean, morphing from A directly to D skipping B and C

Using the MSEG's maybe?

This and/or the arp.

[recursive one]

How?

I mean, morphing from A directly to D skipping B and C

Using the MSEG's maybe?

Still don't see how. You can draw a vertical line in the MSEG so that the WT will abruptly go from one position to another, but in fact it will still scan thorugh all the wavefroms separating these positions in the wavetable just doing it as quickly as the selected modulation speed allows ( or do I misunderstand the nature of Dune's wavetables?)

And what if you want to morph from A to D slowly (still skipping C and B) and then slowly from D to B skipping C again?

Not sure if any of the wavetable synth can do that and if it would be musically useful, just trying to wrap my head around all this "nD wavetable" thing.

[wagtunes]

How?

I mean, morphing from A directly to D skipping B and C

Using the MSEG's maybe?

Still don't see how. You can draw a vertical line in the MSEG so that the WT will abruptly go from one position to another, but in fact it will still scan thorugh all the wavefroms separating these positions in the wavetable just doing it as quickly as the selected modulation speed allows ( or do I misunderstand the nature of Dune's wavetables?)

And what if you want to morph from A to D slowly (still skipping C and B) and then slowly from D to B skipping C again?

Not sure if any of the wavetable synth can do that and if it would be musically useful, just trying to wrap my head around all this "nD wavetable" thing.

I'm not aware of any wavetable synth that can "skip" individual waves in a table but I could be wrong. I just know that I don't own any.

[cytospur]

How?

I mean, morphing from A directly to D skipping B and C

Using the MSEG's maybe?

Still don't see how. You can draw a vertical line in the MSEG so that the WT will abruptly go from one position to another, but in fact it will still scan thorugh all the wavefroms separating these positions in the wavetable just doing it as quickly as the selected modulation speed allows ( or do I misunderstand the nature of Dune's wavetables?)

And what if you want to morph from A to D slowly (still skipping C and B) and then slowly from D to B skipping C again?

Not sure if any of the wavetable synth can do that and if it would be musically useful, just trying to wrap my head around all this "nD wavetable" thing.

Use MSEGs or arp to do the skipping. Use two voices for this. Use another MSEG to cross fade between the two voices. Do the skipping when the voice is silent.

[zerocrossing]

meh, i'm gonna wait for the 4D version coming out next month

I was just going to type, "I don't get out of bed for less than 4 dimensions!"

[recursive one]

Thanks, now I see. Cumbersome but technically possible

What are your thoughts, Mark, what is "3D wavetable" and how it is different from Dune's wavetable architecture?

[blacktomcat666]

Not sure if any of the wavetable synth can do that and if it would be musically useful, just trying to wrap my head around all this "nD wavetable" thing.

Indeed it can be useful - especially if you imagine other content than basic wave shapes (resynthesized speech, instruments...). Too bad that most of the WT synths have so little slots sizes like 412 - 4096 samples. With larger or arbitrary sizes one could use the WT features on very complex sounds and the multidimensional tables would be even more interesting... making an ever changing soundscape from a movie snippet or so.

[recursive one]

Not sure if any of the wavetable synth can do that and if it would be musically useful, just trying to wrap my head around all this "nD wavetable" thing.

Indeed it can be useful - especially if you imagine other content than basic wave shapes (resynthesized speech, instruments...). Too bad that most of the WT synths have so little slots sizes like 412 - 4096 samples. With larger or arbitrary sizes one could use the WT features on very complex sounds and the multidimensional tables would be even more interesting... making an ever changing soundscape from a movie snippet or so.

Or another possible application would be "amination" of digital osillators by making a wavetable from a bunch of slighlty different waveforms and setting a moldulation control so that these waveforms would randomly morph into each other when a note is held.

[Ingonator]

nvm

[aMUSEd]

Wish they would take that shop button off their plugins. I don't want a shop inside a plugin.

[chroma]

i've often thought a 3-D wavetable synth would be interesting (or even N-dimensional). in practice though, no one really would use it much, due to setup time.

in general, though, you can think of an N-dimensional wavetable as a sampled version of their z-plane filters. All of it is basically being able to morph. Morph just means go from one sound to another - it's not a fixed process. E.G. if you go from a square wave to a saw wave, what does that actually mean? you could do a power crossfade, you could remove harmonics one by one until sine then put them back, there are an infinite number of possible paths between saw and square. z-plane filters are similar - you could xfade between an LPF and an HPF, but z-planes let you with a single parameter walk an arbitrary set of changes between those two extremes.

there are two reasons people use wavetables... one of them is to deal with the above. it's probably no accident that ensoniq created both transwaves and z-plane filters. it's also no accident that the ASR-10 has both transwaves and non-resonant filters, as that's one of the things they expected a wavetable to do.
i don't know what the thinking really was behind the originals, but i think that was at least part of the motivation of transwaves.

imagine a filter sweep with a sawtooth. you probably know that good filters eat CPU for breakfast, but playing samples is pretty easy. imagine that instead of creating a saw wave and putting through a filter (which sweeps through 128 settings), you sample it at each of the 128 filter settings in small loops. now, you have a sample with the sawtooth through the filter completely closed, with it a little open, etc. instead of actually doing a filter, you just play back the proper loop as you sweep the filter. that's a 1D wavetable.

now, imagine that you also want to be able to modify resonance. now, we have 128 samples, one for each of the 128 filter settings. add another dimension - for each of these samples, create 128 samples, one for each resonance setting. we now have 128 x 128 samples - one for filter closed, no resonance; another for filter closed, a little resonance; another filter closed, max resonance, another for filter half open no resonance, another for filter half open, max resonance, etc. that's a 2D filter.

of course, doing this is not like a filter - you're baking the cutoff frequency into the waveform - as you pitch this up and down the keyboard, the real filter cutoff will change; you can kind of simulate this by keytracking->waveoffset. and waveforms change on single loop boundaries, so however long the loop is, is how often you can update the settings. Also, this is only a single waveform - if you wanted to do square instead with this filter, it's a whole separate 128x128 wavetable, with all new waveforms. On the other hand, it's like a sample. It's similar to how we do e.g. strings today - lots of samples, and when you do expression you're crossfading between sometimes many samples to get it to sound real. It's not exactly right, but it's closer than synthesis, so we live with it.
In doing this, you can get the complex timbral changes that hard vs. soft bowing, getting close to overblow in horns, etc. that aren't anything close to simple filtering.

In practice, when the ASR-10 came out, CPU power was a a premium. Memory was too, but with small enough waveforms, you could make this work, and so a sampled resonant sweep was 'cheaper' than implementing a filter per voice. That's all changed, of course, and while ZDF filters etc. are still very CPU-hungry, putting a simple resonant filter on every voice is not difficult anymore. So there aren't many things left really that would be easier to do as a wavetable. And like crossfading through string samples, wavetables can do any arbitrary timbral series you want, but it takes a lot of work to set all that up.

A lot of what people like about wavetables isn't that anyways... the artifacts of the wave sequencing process are often what people are after as much as the wave sequencing itself (the second reason for wavetables). the crunchy sound of the low-fidelity sample engines etc. often created magic. assuming this will have the ability to reproduce some of those as well.