Which "FM" synths really are FM?

[Ingonator]

As already mentioned multiple times it depends on which waveform you use. For Sine waves the result of FM and PM is quite similar and for example the first DX7 only included Sine waves.

One sense of the development of a newer FM synth like e.g. Tone2 Nemesis was to use any kind of available waveform which with PM could lead to strange results opposing to real FM.

Nemesis maybe has fewer operators but the use of complex waveforms in several cases could substitute having a smaller amount of operators (while it is still possible to use feedback for the carrier).
For example instead of using FM/PM to create a sawtooth from Sines you could just use a sawtooth waveform instead. You could even import single cycles that are based on samples of other FM/PM synths.

[BlackWinny]

And these two books are so simple to read (they introduce first the theory but once it is done they show how to use the theory in applications).

I think that everybody here went to school or have had learning courses for their profession. Do we learn how to apply a theory without first learn that theory?

Especially when it is so simple to have access to the sources... and to read them!

Does a synth maker create features without having first learned how work these features?

I have the feeling that many users confuse the learning of how to use a feature and the learning of how works a feature.

It is totally vain to tell "This synth uses this technical feature" and "That synth uses that technical feature" without having first learned what are technically these two features. Period.

And they are so simple to learn that it is incomprehensible to see so many people not wanting to begin by that: read the sources to be sure to not misunderstand the EXACT things.

After, to see which synth applies this technique and which synth applies that technique is far more clear! And without any doubt on what is what!

I'll not add anything. Read these f**king books.

And by the way, Chowning talks also about the phase modulation in the chapters about the frequency modulation, to explain what is common to them (because yes, there are some common things) and what differs between them.

I'll not add anything. Read these f**king books.

[Ingonator]

And these two books are so simple to read (they introduce first the theory but once it is done they show how to use the theory in applications).

I think that everybody here went to school or have had learning courses for their profession. Do we learn how to apply a theory without first learn that theory?

Especially when it is so simple to have access to the sources... and to read them!

Does a synth maker create features without having first learned how work these features?

I have the feeling that many users confuse the learning of how to use a feature and the learning of how works a feature.

It is totally vain to tell "This synth uses this technical feature" and "That synth uses that technical feature" without having first learned what are technically these two features. Period.

And they are so simple to learn that it is incomprehensible to see so many people not wanting to begin by that: read the sources to be sure to not misunderstand the EXACT things.

After, to see which synth applies this technique and which synth applies that technique is far more clear! And without any doubt on what is what!

I'll not add anything. Read these f**king books.

And by the way, Chowning talks also about the phase modulation in the chapters about the frequency modulation, to explain what is common to them (because yes, there are some common things) and what differs between them.

I'll not add anything. Read these f**king books.

I don't need a book for that. I know that Tone2 Nemesis includes both real FM and PM modes and i know that with complex waveforms (not a simple Sine) there is a difference. This is not just theory...

In Nemesis you could already notice the difference using the Init patch (from "ini all" at the File menu) where the Modulator is a Square wave and the carrier a Saw wave. If you start cranking up the "Neo FM" knob (= FM amount) and switch between the "NeoFM" (= real FM) and "FM/PM" (= phase modulation) modes depening on the FM amount the differncen in sound could be big.
With more complex waveforms the difference could be even bigger.


The points is that just because a synth is named a FM synth it does not mean it also actually uses real FM.

[wagtunes]

And these two books are so simple to read (they introduce first the theory but once it is done they show how to use the theory in applications).

I think that everybody here went to school or have had learning courses for their profession. Do we learn how to apply a theory without first learn that theory?

Especially when it is so simple to have access to the sources... and to read them!

Does a synth maker create features without having first learned how work these features?

I have the feeling that many users confuse the learning of how to use a feature and the learning of how works a feature.

It is totally vain to tell "This synth uses this technical feature" and "That synth uses that technical feature" without having first learned what are technically these two features. Period.

And they are so simple to learn that it is incomprehensible to see so many people not wanting to begin by that: read the sources to be sure to not misunderstand the EXACT things.

After, to see which synth applies this technique and which synth applies that technique is far more clear! And without any doubt on what is what!

I'll not add anything. Read these f**king books.

And by the way, Chowning talks also about the phase modulation in the chapters about the frequency modulation, to explain what is common to them (because yes, there are some common things) and what differs between them.

I'll not add anything. Read these f**king books.

Except I don't care.

Let's see, does my synth have operators like the DX 7? Can I manipulate them like the DX 7? That's enough FM for me even if the DX 7 wasn't an FM synth.

Don't need to do no reading to know what I want.

[Tricky-Loops]

And these two books are so simple to read (they introduce first the theory but once it is done they show how to use the theory in applications).

The question is "Which 'FM' synths really are FM?" and NOT "Which are the original books to describe FM and PM mathematically?"

There are people here who aren't mathematicians or programmers (and most of them aren't scientists, either). They want to have a short article about the differences, not hundreds of pages. (So I posted a link on page 1 here to a short article about the differences.) It's not about laziness, it's about time management - you have to set priorities, you cannot do everything just because you like to.

When I have some time, I read a lot of books about synths & effects, music theory, mixing, mastering etc. but I don't have the time to read all original books of all kind of synthesis techniques. I still have to go to work whether I like it or not, I don't have enough time to read books or watch waveforms in oscilloscopes the whole day...

[Ingonator]

I just did a Test based on the Init patch of Tone2 Nemesis using the "NeoFM" (= real FM) and "FM/PM" ( = phase modulation) modes.
Here is a screenshot of the patch (only Osc/Layer 1 is used and only the synthesis mode will be changed for the tests):

https://dl.dropboxusercontent.com/u/532 ... t%2001.png


The setting of the "NeoFM" knob in Osc/layer 1 represents the FM amount.


This is an audio demo using the "NeoFM" mode (= real FM):
Tone2 Nemesis - NeoFM vs PM Test 01

This is an audio demo using the "FM/PM" mode (= phase modulation):
Tone2 Nemesis - NeoFM vs PM Test 02


As both modes could have their uses you could combine them in one patch by e.g. using NeoFM in Osc1 and FM/PM in Osc2.
There is also a "FM/PM Vintage DX" mode which is more or less similar to "FM/PM" but is a bit more "noisy" like in a real DX7.

[Syntilla]

It is totally vain to tell "This synth uses this technical feature" and "That synth uses that technical feature" without having first learned what are technically these two features. Period.

And they are so simple to learn that it is incomprehensible to see so many people not wanting to begin by that: read the sources to be sure to not misunderstand the EXACT things.

Precisely.
And all those people who made great music with the DX7 thinking it used FM are complete f*ckwits.
Difference between science (or at least engineering) and art, I guess.
Never the twain ...

[MadBrain]

ALL PRACTICAL "FM" SYNTHS ARE PHASE MODULATION.

Everything by Yamaha. (including dx7, sy-99, fs1r)

Everything that emulates Yamaha (FM8, Dexed, Sytrus).

The reason for this isn't that phase modulation is easier to implement. The reason is that when using frequency modulation, if the input waveform isn't perfectly balanced (ie has any DC whatsoever), the carrier oscillator goes out of tune and it quickly becomes impossible to get any musical result. This makes feedback unusable, and any stack higher than 2 operators turns into atonal metallic bwongs as well. So basically PM is pretty much always more usable than frequency modulation.

Also, phase modulation with a waveform is equivalent to frequency modulation with the derivative of this wave. PM triangle = FM square. PM abs(sine) = roughly FM saw. PM sine = FM cosine. So both methods are the same, you just need to use softer modulator waveforms in PM.

As for Casio PD synths, the original CZ are basically a type of wave morphing, and are mathematically equivalent to PM with feedback all the way up and a triangle wave modulator. Also, the "Synth" portion of Roland LA synths is also a variant of phase distortion (designed to simulate an oscillator+filter). The Casio VZ-1 is different, and is basically a PM synth except that carrier frequency is forced to 0 (to get around the Yamaha patent), and with various hybrids between sine and saw as waveforms (to put back some even harmonics and sorta do an equivalent to feedback).

[Naenyn]

Well, it's official. This is the most pointless thread at KVR. And that's saying a lot with that CA fiasco going on for over 275 pages now.

Really, who cares?

I care! I asked the original question.

I was curious about the difference and which synths were one or the other. Considering new synths come out fairly regularly claiming to use some new form of synthesis, I find it interesting that there are synths out there that say they use one thing but actually use another. I'm also in the process of learning how synths work, so this discussion is very relevant to me at the moment.

This is more of an academic discussion than a discussion about which sounds better. If you aren't interested in the differences between FM and PM synthesis, you don't have to follow the thread.

[Naenyn]

You can't have negative frequency. You can have negative phase.

Yes you can. Frequency is change in phase over time, which can be negative. What you can't have is *pitch* below 0Hz (because pitch is log of a fundamental frequency).

From a mathematical perspective, doesn't a change of phase of a waveform eventually produce a resulting waveform that is identical to the original? Whereas frequency modulation will just continue to increase (or decrease, depending on direction) the frequency over time?

Maybe I'm misunderstanding the math...

[Ingonator]

Well, it's official. This is the most pointless thread at KVR. And that's saying a lot with that CA fiasco going on for over 275 pages now.

Really, who cares?

I care! I asked the original question.

I was curious about the difference and which synths were one or the other. Considering new synths come out fairly regularly claiming to use some new form of synthesis, I find it interesting that there are synths out there that say they use one thing but actually use another. I'm also in the process of learning how synths work, so this discussion is very relevant to me at the moment.

This is more of an academic discussion than a discussion about which sounds better. If you aren't interested in the differences between FM and PM synthesis, you don't have to follow the thread.

If you check my sound examples above (at this page) you could hear there is a difference. I think the problem is that many people have not used synths that could use actually both real FM and phase modulation or have not tried using other waveforms than Sine.

[AdmiralQuality]

You can't have negative frequency. You can have negative phase.

Yes you can. Frequency is change in phase over time, which can be negative. What you can't have is *pitch* below 0Hz (because pitch is log of a fundamental frequency).

Try it in Poly-Ana. When FM modulated, Poly-Ana's oscillators are designed to give the same musical pitch range positive and negative
around the center frequency the oscillator is set at.

Yes, that's because you're modulating pitch (in semitones) not frequency (in Hz, or expressed as a linear multiple of the root frequency).

Exactly.

This is the obvious setting for balanced vibrato and trill effects. The pitch goes down just as many semitones as it goes up. But you can't ever go negative, you just keep finding lower and lower octaves -- right? So when you apply this, even a little bit (where we don't approach 0 or negative frequency at all) the oscillator seems to go out of tune, at least when modulated at an audio rate. Perhaps there's another way of scaling the frequency modulation so this doesn't happen, but then the FM would be less useful for typical LFO applications (vibratos and trills).

Indeed there is. Linear scaling as described above. It's better for "FM" sounds, but not as good for LFO applications. So in the case of Cypher, the audio rate FM input is linear frequency scaling, the control rate LFO input is pitch scaling. Does it sound different to Phase Modulation? Arguably not by much, but I can't see how you would do PM on a pure VCO-style oscillator - it's almost a contradiction in terms.

You said it right there. "Style". Nothing to stop us from doing it here in the digital domain. I believe there have been pure analog PM synth implementations but I could be wrong on that.

I'll have to give linear FM a try sometime. Yes, more accurately the FM in Poly-Ana would be called "Pitch Mod" (which would give us a conflict of abbreviations against Phase Mod). And I can't do the audio-rate trick, as there is no conceptual difference between audio-rate and LFO on Poly-Ana. (The LFOs go well into audio-range and the oscillators can be driven to sub-sonic pitches.) Would the multiplicative ranges still be be recipricols of each other? For example 0.5 to 2.0? (Which is still the same interval as +/- 1 Oct, but a linear progression through frequency instead of pitch. And if so I still don't see how the frequency would ever go "negative".) Anyway, if it sounds substantially the same as the existing PM mode it probably wouldn't have been worth implementing, but I will try it sometime and judge with my ears.

But yes, PM was too powerful a feature to pass up. While effectively only having a couple of operators, you can get a LOT of PM sound out of Poly-Ana. And with a resolution that far surpasses anything the DX family could ever do. It's simultaneously familiar and unique sounding.

[Tricky-Loops]

If this is "true" FM in Nemesis, it sounds much more metallic with more higher harmonics than the PM example. But maybe they combined it with pitch modulation or formant synthesis? How can we be sure that it's really "true" FM?

[Ingonator]

If this is "true" FM in Nemesis, it sounds much more metallic with more higher harmonics than the PM example. But maybe they combined it with pitch modulation or formant synthesis? How can we be sure that it's really "true" FM?

Because this is one reason why Nemesis was developed...

The resulting sound could be quite different based on the setting of the FM amount knob. Sometimes small amounts of FM are enough.
Anyway if you like the result of PM you could also combine both in a single patch. At certain waveforms PM will become more and more useless, the Saw and Square waveforms used in the example still more or less "Standard" and you could already notice a difference.

A quote from the Nemesis website:

Nemesis offers an innovative approach called 'neoFM' synthesis which is genuine FM synthesis.
Most conventional FM synths such as the DX7, use phase modulation (PM) instead of genuine FM synthesis, yet such synths are branded as FM synthesizers for marketing reasons. However, the traditional PM approach is limited to using dark sounding waveforms like sinoids or triangles, which results in a dull or bell-like sound. Classic waveforms like a sawtooth do not work well with it because they result in an unpleasant, sharp and excessively metallic sound.
'Real analog FM' suffers from drift and lacks precision, resulting in a disharmonic sound, which is musically not very useful.
The innovative neoFM approach, exclusive to Nemesis, combines only the advantages of both types of synthesis without suffering from any of the weaknesses associated with PM or 'Real analog FM' synthesis.

Aside from NeoFM synthesis, Nemesis features an astounding selection of no fewer then 22 combinable synthesis types. From Waveshaping, Phase Distortion, Sync, Wavetable, Ringmod, Vintage FM, up to the Reso and Advanced Formant Synthesis exclusive to Nemesis.

Of course also real FM could sound metallic and/or piercing but this also depends on the FM amount and the used waveforms.

[Kaboom75]

I thought the inventor of FM sold his patent to Yamaha to make the DX line of synths. Looks like I got some reading to do. When someone says FM I always think of DX7, Sega Megadrive and u-he Bazille.