Triangle Waveform Sounds Like a Square Wave?!

Try this:

Create an empty patch, turn on operator a and b.

Turn up the level on a and set to sine wave then play middle C.

Set b to modulate a and set b's ratio to 0.0020.

Set b's waveform to Triangle (3) and turn up b's level going to a.

make sure the envelopes for each operator are set to the gate preset.

Does it sound like b is a triangle wave? It sounds like a Square wave to me. Also, the Square waveform seems to have a really short pulse, more like a rectangle, and a long flat middle.

Is it just me?!

 

From other FM synths I know that at a certain modulation those synths reach a saturation point. So what you probably hear is that the upper and lower peaks of the triangle were cut off, so the result is almost a square wave...

 

If the modulator was saturated, then the sine wave would be saturated also and it's not, it sounds like a sine wave, but the triangle sounds like a square wave to me, and so does the square wave.

 

That effect also appears when using a sine wave but since the gradient of the sine isn't a linear function the difference isn't that noticeable...

 

So, how do I prevent the saturation and get a triangle wave as a modulator instead of a saturated triangle (square) wave? I turned the level coming out of modulator b down, it doesn't seem to be as audible.

The other waves sound a little strange also (square, soft square, sawtooth, etc.). They seem to stay at 0 for a long time and then shoot up positive or negative very briefly.

 

Well, the modulation depth is depending on the frequency of the modulator, as lower its frequency as lower the influence on the carrier. That way the modulation index (and that way the created overtones) stays the same all over the keyboard. You could try to modulate the carrier with more than one modulator, set them to key sync and their modulation level to around 40, but this would probably waste all the operators left and it seems to me that the result isn't that much better anyway.
What exactly do you want to archive, there might be another way to do the same thing.

 

I'm trying to learn FM synthesis, so I'm slowing down a modulator to a speed where I can actually perceive what it is doing to the carrier. I didn't realize that all modulators' waveforms will be saturated (or maybe clipped is a better word).

When setting the LFO to modulate the pitch of the carrier, the waveforms (of the LFO) sound correct. I'm just curious as to why I can't use one of the operators as a modulator to achieve the same result. In other words, why are they saturated/clipped?

 

Well, actually all the so called FM synths, for instance the legendary DX7, are using Phase Modulation and that's why they behave a bit different compared "real" Frequenzy Modulation or when using an LFO.

Check the following link for a more detailed description.

http://www.funet.fi/pub/sci/audio/misc/pm-intro

It's maybe a bit hard to understand with all those formulas, but I think essence is that PM is easier to implement and easier to use. I also found a page with links to the Yamaha US Patents that seems to proves the posting...

 

no saturation

Whatever you are hearing, it is not because of saturation of modulation amounts.
The modulation does not saturate (unless you route it though the X operator).
But Summa is right, in that the FM used in FM7, like in all the Yamaha synths, should correctly be called Phase Modulation.
But in essence, Phase Modulation and Frequency Modulation is the same thing, except that in PM the modulation is stronger at high modulation frequencies and lower at low modulation frequencies.
If you slow your modulator right down, the effect will get so small you don't notice it anymore.

ciao,

Michael

 

Hi Michael,

also things like cascading modulators withouth getting a DC offset and feedback wouldn't work on a real FM synth...

But you might be able to tell me why the FM7 (like every FM - synth I know) is starting to create additional harmonics, even when modulator and carrier are set to 1:1, that aren't "part" of the modulator when reaching a certain amount of modulation, and before that level the amount of harmonics are increasing almost equivalent to the modulation level.
My guess is, that at a certain point the carrier isn't able to "follow the influence" of the modulator so it's like working with a saturated/clipped modulator waveform.
I know that this description might be inaccurate, since I only describe what I recognise...

I hope I'm not too insisting or annoying

Regards!
Summa

 

hard maths

The generation of harmonics in FM is quite a complex thing if you try and analyse it. For a mathematical description you have to resort to Bessel functions.

In a musical sense, this means that for low FM amounts you have only few harmonics and at high FM amounts you get a lot high harmonics, certain low harmonics even disappear at some modulation amount.

This is just the way FM works, in DX like in FM7. But it can be a bit hard to predict intuitively, unlike the much easier interaction between an oscillator and a filter.

All you really need to know about FM is that you can get harmonics at any frequency which is a multiple of the difference between carrier and modulator frequency. For example, 1:3 has difference 2, so you get 1,3,5,7...
The actual amplitudes of the harmonics depend in a complex way on the modulation amount, but basically: the more modulation, the stronger the higher harmonics.

 

I don't have problems programming sounds for a FM synth, I'm for some years at a point were I find FM quite intuitive. But since half of my equipment consists of FM - Synths it's no wonder I feel that way...

Well, it's not only the amount of harmonics, it's also how the FS1R or VZ-1 reacts when modulating a saw with a sine at a low fixed frequency that let me think that there's probably some kind of clipping. Other than the FM7 both create phasing instaed of pitch modulation and at a certain amount of modulation the phasing stops for a little moment and that starts again like if the sine wave of the modulator has been clipped, that pause is increasing when I increase the modulation amount. I guess that's about the same thing Ted experiencing with his triangle waveform.