FM synthesis / Analog synthesis sound/technology difference?

[Sound Mechanics]

I am curious about the type of sound that FM produces compared to analog. (DX7 v Minimoog)
I have compared my FM7 to my Korg Legacy MS-20 with obvious differences but wonder what is actually going on 'under the bonnet'.
I am not a programmer but I have read a little on the differences in the type of synthesis. But they both essentially use oscillators.
I presume a DX7 had some sort of digital oscillators, like VA synths? Can you have FM synthesis with minimoog oscillators and combine the rest with FM type synthesis

I am probably getting this totally wrong so I thought this would be the place to learn

[VicDiesel]

But they both essentially use oscillators.

Kinda.

I presume a DX7 had some sort of digital oscillators, like VA synths? Can you have FM synthesis with minimoog oscillators and combine the rest with FM type synthesis

You're quite on the wrong track.

Analog synthesis uses oscillators with a rich sound: apart from the fundamental they have lots of overtones. You then use a filter to lose some of those and thereby determine the tone.

In FM synthesis you start with a sine wave, which has no overtones, and then modulate it to get overtones. These overtones are very different from what you get from filtering a sawtooth et cetera in analog synthesis. In fact, apart from overtones you get undertones, making the whole concept of a fundamental a bit hazy.

So there are big differences. FM is much harder to get your head around.

Victor.

[Sound Mechanics]

Thanks Victor for the reply.

What do you mean by overtones?
What about an analog sine wave, what's the difference to the FM sine wave?

cheers

[aallvor]

What about an analog sine wave, what's the difference to the FM sine wave?

cheers

Nothing I should think. A sines a sine

[fmr]

I presume a DX7 had some sort of digital oscillators, like VA synths? Can you have FM synthesis with minimoog oscillators and combine the rest with FM type synthesis

You're quite on the wrong track.

Analog synthesis uses oscillators with a rich sound: apart from the fundamental they have lots of overtones. You then use a filter to lose some of those and thereby determine the tone.

In FM synthesis you start with a sine wave, which has no overtones, and then modulate it to get overtones. These overtones are very different from what you get from filtering a sawtooth et cetera in analog synthesis. In fact, apart from overtones you get undertones, making the whole concept of a fundamental a bit hazy.

So there are big differences. FM is much harder to get your head around.

Victor.[/quote]

Well, he isn't completely wrong in the first place, and you are not completely right.

1. There is not any obligation for FM to use sine waves. You can use anything, including samples (you can't go more complex than this).

2. You could make FM on a Minimoog, if you had the routing options, as you have e.g. on a Moog Modular.

3. FM is based on the creation of partials done through the modulation of a wave (ANY wave) by another wave, both in the audible spectrum. If you had an LFO and could raise it frequency until it reaches the audible spectrum, you would achieve FM. The more modulation intensity you put, the more partials you create - it's like a filter behaviour on the opposite. What kind of partials depends on the pitch of the modulator and of the modulated (which is called carrier, because it's the oscillator that's going to be audible). Since you need so many options on each oscillator, actually they are almost a simple synth by themselves, with LFO, pitch envelope, amplitude envelope, etc. That's why they are called operators, and not simply oscillators.

1. What is called "analog saynthesis" is in fact subtractive synthesis, because we start with complex waves, and then filter the content to achieve other timbres.

5. FM, to be done in a usable matter, requires more oscillators than usually analog synths have (four is more or less the minimum, but six or more is desirable).

6. FM synths could also perform simple additive synthesis, due to the number of oscillators they have (all oscillators are treated as carriers).

6. Today FM synths also have filters, which means that they are also subtractive, as well as FM (and you have six oscillators, instead of two or three).

You can find much more information on the web about FM (even the book written by John Chowning abd David Bristow about programming the DX7).

[John Vulich]

2. You could make FM on a Minimoog, if you had the routing options, as you have e.g. on a Moog Modular.

BTW, the Minimoog does have FM routing. Osc 3, which doubles as an LFO, can be used for FM in a low frequency and audio range mode.

[Sound Mechanics]

Thanks all - very interesting and I'm a little wiser now!

[Sound Mechanics]

With the "search" for digital creating that "analog" sound, really FM is an ideal form of synthesis for computers?
No analog v's digital argument here please!

[VicDiesel]

With the "search" for digital creating that "analog" sound, really FM is an ideal form of synthesis for computers?
No analog v's digital argument here please!

The analog vs digital argument is orthogonal to Subtractive/FM. You can do FM with analog hardware, and you can do traditional Oscillator-Filter stuff with digital software.

FM would be ideal if it could create every sound. Don't know if it can. I guess if you allow arbitrary waveforms, you don't modulate them, and you then put a filter on them, then an FM synth can do anything a subtractive one can.

Victor.

[Sound Mechanics]

With the "search" for digital creating that "analog" sound, really FM is an ideal form of synthesis for computers?
No analog v's digital argument here please!

The analog vs digital argument is orthogonal to Subtractive/FM. You can do FM with analog hardware, and you can do traditional Oscillator-Filter stuff with digital software.

FM would be ideal if it could create every sound. Don't know if it can. I guess if you allow arbitrary waveforms, you don't modulate them, and you then put a filter on them, then an FM synth can do anything a subtractive one can.

Victor.

I think my FM7 can do all of this!

[Radek]

FM synths have also a feedback loop what make things ever more complicated than usual substractive aproach. A small change in one parametr of one operators can results in a vastly different sound.

[Borogove]

FM synths have also a feedback loop what make things ever more complicated than usual substractive aproach. A small change in one parametr of one operators can results in a vastly different sound.

But, again, there's no reason you can't put a feedback loop into an analog modular FM setup, and most FM synths have a filter, so they're "subtractive" as well.

[Toxikator]

Not only that, but these days most of the upper-tier FM synths are really also equipped with capable subtractive features. Lok at z3ta+, rhino, Absynth, Sytrus; FM by design but with subtractive abilities too.

FM synthesis doesn't have to start with sines, either; if you think of PWM (which is a form of phase modulation which is nearly identical to FM synthesis), that's a form of "FM" practiced on non-sine oscillators.

[emdot_ambient]

Hemmick Reef, you probably should do some more reading on synthesis in general. http://www.kvraudio.com/forum/viewtopic.php?t=76293 is a good place to start.

Computer technolongy has really blurred a lot of the lines in traditional synthesis, so it's easy to get confused by we hair-splitting, pedantic and prone-to-obfuscation KVR members.

Read about the basics first, then come ask us to refine your knowledge to the n-th degree!

[Toxikator]

Indeed. Start with Wikipedia, it provides a good overview (along with technical jibbajabba). Read up on subtractive and FM/PM synthesis. If you like, you can also read up on Wavetable synthesis, modelling synthesis, and additive synthesis