Wavetable Synthesizers

[generaldiomedes]

Personally I would only call a synth a ‘wavetable synth’ these days if It could handle 256 waveforms per wavetable and cycle through them. Serum, Avenger, Rapid, Wavetable, etc.

[fmr]

Personally I would only call a synth a ‘wavetable synth’ these days if It could handle 256 waveforms per wavetable and cycle through them. Serum, Avenger, Rapid, Wavetable, etc.

Yeah, I usually think of something like those when I see the word too, but "personally" is not a good way of defining things. Technically, those are "dynamic wavetable scanning" synths. There are other "wavetable" synths that look at a single cycle wave at a time, but are still "wavetable" synths, in this case, "wavetable lookup" synths.

That's why, when someone uses the word, should specify exactly what is meant with it.

[Dasheesh]

Synthmaster.

There used to be a defined difference between wavetabling and wave morphing. Not so sure it applies anymore, except an advanced wave morph like ZETA can create new waves where non existed by reshaping existing waves. Old style wave tabling was a bit rougher and more grungy. Pc’s let wavetabling become fixed wavemorphing like Arcsyn or spire. You guys can argue over the differences.

[fmr]

Synthmaster.

There used to be a defined difference between wavetabling and wave morphing. Not so sure it applies anymore, except an advanced wave morph like ZETA can create new waves where non existed by reshaping existing waves. Old style wave tabling was a bit rougher and more grungy. Pc’s let wavetabling become fixed wavemorphing like Arcsyn or spire. You guys can argue over the differences.

In my POV, wave morphing is limited to the morphing algorithms, while in a wavetable, theoretically, you can have a completely different wave in each position. Whether this would sound or would be interesting or even musically appealing, that's another story.

For example, what if you synthesize a musical phrase that has some bell like sounds periodically, over a pad like atmosphere? The wavetable can give you back the exact same effect, if scanned at a proper speed. Granular synthesis can also give you something similar. Using wavemorphing, I can't see how you would obtain that effect.

I'm, not saying morphing synthesis is not interesting, I'm saying that it's not a replacement, except for some specific results.

[Urs]

I don't like the term "morphing" for any kind of audio because it's utterly overused for all sorts of unrelated things. In imaging (e.g. video production), it refers to seamlessly transforming the outline of a two-dimensional shape from one form to another, while crossfading the embedded structure. It's inherently only spectacular in a shape where each coordinate on the outline can freely move in X and Y direction. This requires that multiple points can occupy a location on the X-axis. A waveform can't do latter. Hence only few can morph (e.g. PWM), most can just crossfade. A similar problem occurs in the time domain where the most interesting part of the spectrum is so sparse, real morphing would require inharmonic movements (i.e. not suited for most synths).

[fmr]

I don't like the term "morphing" for any kind of audio because it's utterly overused for all sorts of unrelated things. In imaging (e.g. video production), it refers to seamlessly transforming the outline of a two-dimensional shape from one form to another, while crossfading the embedded structure. It's inherently only spectacular in a shape where each coordinate on the outline can freely move in X and Y direction. This requires that multiple points can occupy a location on the X-axis. A waveform can't do latter. Hence only few can morph (e.g. PWM), most can just crossfade. A similar problem occurs in the time domain where the most interesting part of the spectrum is so sparse, real morphing would require inharmonic movements (i.e. not suited for most synths).

Yeah, I agree. When I see morphing, I tend to think in terms of crossfading (although, for the Kyma people, they somehow do the morphing in the spectrum domain, I think, which gives some different/interesting results).

[liv]

Any of you guys know any good Wavetable Synths, paid or free, that I should look into? Any help would be greatly appreciated.

Massive all the way.

[fluffy_little_something]

I can imagine how morphing between simple waveforms works, like from square to sine (which can be created by lots of squares to begin with, so I suppose one would sweep through the number of squares), but how would one morph between a sine and a complex, irregular waveform?

[Apratim]

Mpowersynth
Can hold 256 waveform states and can also "morph" seamlessly.but boy are the reverbs in this thing are a beauty of its own...just my thoughts

DS Thorn
Can draw only 16 frames but I love it more for its spectral filter than its wavetable.and this thing could be Harsh sometimes due to its spectral filter

MUX and reaktor
Modular environment you know it can do anything.... almost

[chroma]

I can imagine how morphing between simple waveforms works, like from square to sine (which can be created by lots of squares to begin with, so I suppose one would sweep through the number of squares), but how would one morph between a sine and a complex, irregular waveform?

the answer to that question is where the magic is...

[Dasheesh]

If you guys want to start drawing lines in the sand there is probably a difference between fixed wave morphing, and complex wave morphing (which is really rather FM like TBH). Urs makes a good distiction about the resolution of such synthesis. Encouraging the inharmonics. He might have enough vision to see that the whole model is going through a paradigm shift and have an idea about how to deal with it (wouldn’t surprise me).

[fluffy_little_something]

I can imagine how morphing between simple waveforms works, like from square to sine (which can be created by lots of squares to begin with, so I suppose one would sweep through the number of squares), but how would one morph between a sine and a complex, irregular waveform?

the answer to that question is where the magic is...

I think I wrote it the wrong way round, the square can be made from sines, right?

[Dasheesh]

I can imagine how morphing between simple waveforms works, like from square to sine (which can be created by lots of squares to begin with, so I suppose one would sweep through the number of squares), but how would one morph between a sine and a complex, irregular waveform?

the answer to that question is where the magic is...

I think I wrote it the wrong way round, the square can be made from sines, right?

Everything is made from sines if you are calculating per cycle. Most of the newer stuff doesn’t do it like that anymore. They are mapping waves and copying them across the pitch spectrum.

[jens]

technically i believe most synthesizers are wavetable, unless they are modeling waveforms mathematically in realtime. i am curious about the advantage of using wavetable (in the sense you mean of course) synths today when there are lots of ways to modulate the signal that were not possible when PPG was invented.

Like I said. The OP needs to differentiate between "wavetable lookup" synths, and "dynamic scanning wavetable lookup" synths.

Obviously that distinction went way overhishead...

[chroma]

I can imagine how morphing between simple waveforms works, like from square to sine (which can be created by lots of squares to begin with, so I suppose one would sweep through the number of squares), but how would one morph between a sine and a complex, irregular waveform?

the answer to that question is where the magic is...

I think I wrote it the wrong way round, the square can be made from sines, right?

It doesn't matter... the point is: there is no right or wrong way to morph between a sine and a square, or anything else for that matter.. The most trivial way is simple crossfading... basically, you just mix them; at the start, you have 100% sine and 0% square, then 90% sine and 10% square, all the way through. This is one series.

Another possible series is to have a set of steps, where each step you add the next harmonic in the right amplitube. If this sounds weird, consider that this is pretty much what happens when you open a low-pass filter - a completely filtered square is basically a sine, and it becomes more of a square as you open the filter, until it's fully open and not modifying the square at all (theoretically; real synths usually aren't quite a sine on the low end, and are often not quite the pure waveform on the high end either).

But you could just as easily add all the odd harmonics, then the even ones, or start at the highest and come down (which would basically be something like a highpass sweep + harmonic?), or something completely different.

There's no reason it even has to be a straight line... you could go from square, to triangle, to saw, to noise, to sine, and that's just as valid a path to morph from a square to a sine.

That's basically what makes a scanning wavetable interesting... a regular synth defines one or two paths between its waveforms (usually trivial mixing, which should give the same results as doing it spectrally AFAIK), a scanning wavetable synth lets you define any path you want, and then travel it. However, actually defining an interesting spectral path is not easy...