Thanks for the question... just to clear up one point... FM did “take off in popularity”... like no single synth technology before it. That is the legend of the DX7. If you weren’t there, I guess what remains is the afterglow, I had a front row seat to that whole era (lol)... but up until the DX7 (1983) a wildly successful synthesizer would sell 1000 units in a year. You can figure this based on the original MiniMoog, 1971-1984, sold about 13,000 original units in its thirteen year lifespan.

The DX7, by contrast, sold that many in the USA in the first couple of shipments (the first ones arrived in June, and then a shipment in around October of that year, 1983). It’s really hard to imagine just how EVERYWHERE the DX7 was... I can recall how surprised even Yamaha was at the demand and its success. No one knew the synthesizer market was that big!

Earned the Status: Over-used
When you hear it was “over-used” (a popular thing to say looking back, because it was everywhere) But you must appreciate that this is a phrase that has to be earned. And earned it was, the DX7 and TX816, truly put synthesis in the hands of the common musician. And in virtually every recording studio, world-wide. The ‘bang for the buck’ ratio was extremely high. (Analog synths were running about $1000 per note of polyphony) and topped out at about 8 notes. (Yes two Prophet 5s made a Prophet 10, and two Matrix 6s made a Matrix 12, but that was note real polyphony allocated across all keys - it was 2 five note or 2 six note engines. The DX7 was 16 note poly and was a fraction of the price of the popular analog synths. Doubled the polyphony of the 8-note beasts, at a way lower price, 16 notes of polyphony and came with a sustain pedal, a sustain pedal, a breath controller, and a sweep pedal — synthesizers did not need sustain pedals before this (lol)

To understand the depth of FM popularity... based on the success of this (DX7) synthesis: you could study it in colleges and universities at this time. Schools started to expand the previously small niche synth room into full fledged music programs on synthesis and programming FM... it had teeth and credibility in the halls of academia.

But, in reality, it was not the math-heavy stuff that made the DX7 popular in the marketplace... it was the PRESETs... and the portability (over lugging a Rhodes).

This is why when you search online for 6-Operator FM sounds you windup with tens of thousands of sounds. People wouldn’t necessarily program them themselves, they would *collect* them (like baseball cards).

Those who did learn to program (a smaller subset of the whole) have kept FM alive all these many years later. If you are curious about FM — this website has what you need to understand it’s history and it’s evolution. Follow the featured FM article links: Synth Programming Links

Presets — particularly, the electric tine piano emulation, became a studio standard. It was Rhodes like, and always studio ready, in tune, and crystal clear. This put the synthesizer on not just jazz fusion, and rock recordings, but on so-called “legitimate” music recordings: pop and easy listening, adult contemporary, country, and western (even), genres that previously thought synthesizers were foreign objects from another planet.

Additive synthesis, is just one of the approaches to building sounds, and naturally can be implement with FM-X... In the “Learn” area of this site, Dr. Manny Fernandez takes you through building sounds by constructing different components of the sound, then adding them together. He actually tackles the creation of an acoustic piano using FM... there was a moment at a recent NAMM Show we got to introduce Dr John Chowning to the FM-X (MONTAGE) — he discovered FM as a method to create musical tones with electronic oscillators. Watching him listen to what now is possible, put a lot of things in perspective. It really does work on the same fundamental mathematics of how sounds, harmonics, and music are created in nature. And there truly are things in FM yet to be discovered.

To answer the specific question... FM can surely recreate the fat sounds from an analog synth... but as with any emulation, it does it in a different fashion than the actual instrument. Therefore some of the “behaviors” will be different. Don’t let that confuse you or the issue.

A keyboard Synthesizer can emulate a violin, but it does not use a bow to attack or turn ‘on’ the note, so anticipate the behavior to be different.
When a keyboard Synthesizer emulates a piano, it does not use a string, or hammers, or felt, or a mechanical pedal that shifts the dampers, so anticipate that some of the behaviors will be different.
When an FM synth emulates an analog synth, it does not send use voltage control to tune the pitch (1 volt per octave), so again anticipate that some of the behaviors will be different. Same as when emulating any musical tone or instrument.

That said... analog synth sounds are among the sounds you can emulate with FM-X.
But if some tells you they can always tell an FM synth from an analog synth, don’t argue with them, just know they are, of course, a unique individual. But get no points for this skill. One can probably always tell a sampled piano from an acoustic piano when listening to a recording of their favorite band (or at least we like to tell ourselves that we can always tell). Don’t argue, it’s perception. You can’t know what another being hears or doesn’t hear. The point is: can you pull it off - with a synth emulation? absolutely!

Just recognize that the success of the synthesizer since it became commercially available in the late 1960s, early 1970s has been tremendous.. and evolving and actually growing. Sample playback was so widely successful because of emulation. Early synthesizers were less concerned with emulation... at first they were totally unable to sound like much of anything other than sounding like ”synthesizer”

The success of the DX7 was one indication that *emulation* was what the (larger) market wanted.
Bleeps and bloops were for the longhairs and the professors, the mass market wanted to recognize the instrument sound. They wanted strings that sounded like strings, and brass that sounded like horns, and a piano that sounded like an acoustic piano...

Ranking the “ease of programming” for the DX7 on a scale of 1 to 10 was usually listed as a negative number.
Did that hurt sales? I don’t really think so. The fact that it was programmable was critical, yes... even though nobody programmed it (“It was too hard!” - according to many.
This was the oxymoronic situation. An all preset DX7 would not have sold... people wanted the programmability to refresh the instrument with *new* emulative sounds!

Just correct myself regards popularity...

I meant since the DX series. I do recall their ubiquitous presence on 80's hits. Back then, and since, it was probably the only Synth I actually knew the name of... being more concerned with guitars.

Reading various histories, it seems FM took a back seat come the 90's, although still present on some models, none of them saw the same success as the DX.

Anyways... certainly one of the reasons I chose the MODX7.... was that DX7 font logo. I recognised it instantly. I was looking at the Roland VR initially, but then when I saw the DX, it made me stop and look more closely.

Note that one advantage more recent FM implementations have (including FM-X in MODX) is the availability of effects. The DX7 didn't have effects - so the FM engine itself had to be used to create chorus or other "effects". Now you have the choice to add many layers of effects to FM-X in order to further shape the sound.

I've got a Lowrey organ. It's a transistor organ with a pretty harsh sound overall when using no effects. The organ includes an integrated Leslie branded speaker. Kicking this (effect) on makes all of the difference. Not that a tonewheel is considered an analog synth either - but it is considered to be "fat". And effects can definitely make the difference between thin-and-tinny to fat.

FM-X gives you a head start...

Researching on the web, I discovered that the base operator waveforms (All 1, All 2 etc) are "pre-combined" Operators.

To summarise, pre-FMx, only sine waves were provided, so most programmers would use 2 or 3 of the 4 operators to get to a Saw, Square or Triangle base, then have only 1 or 2 operators left for further modulation.

This is neat. In effect FM-X is providing 9 or 10 operators, if you count the All1/2, Odd1/2, Res1/2 as a 2 or 3 operator head start.

@Anthony

Going for classic analog emulations (and its extremities) with the FM-X engine is one of the reasons I'm playing with Motion Sequences. The amplitude envelopes of the MODX/Montage are not quite, to my sensibilities, right for making lovely phat Andromeda A6 style sounds. But with a little help from quite fast, but ultimately very simple exponentially dropping, Motion Sequences provide the right control mechanisms for amplitude/volume, filter and chorus changes that come closer to giving that lovely spring/sprung release kind of feel that analog has.

However, I'm still sidetracked at the moment, by the AMAZING effect that the Yamaha effects have on FM-X sounds.

Something to do with the "purity" of FM sounds means that the effects have a truly incredible effect upon the sound. And, here again, very fast, but quite long exponential decaying Motion Sequences impacting parameters of those effects lead to ever more intriguing and controllable sounds.

Nothing I've done is polished enough to demo or SoundMondo, I'm still discovering, but I think there's huge untapped potential in terms of shaping FM-X features with the Motion Sequencer, and then shaping the effects upon that sound with Motion Sequences, too.

I really wish there was an iPad visual editor for the complexity of the things I'm building, and to speed up jumping around to set all these things, as the more I know, the bigger my "programming" efforts are getting.

It's somewhat of a pity that a freakish TripHop or Drum and bass nutter hasn't gone bezerk with this thing, yet, to reveal the extents of what's possible. I don't have that sensibility, but do think this might be one of the most ideal hardware options for experimenting with those types of heavily transitioning sounds.

Perhaps if Yamaha could find the time to put a step sequencer in, that would attract that crowd.

Not that it matters much - except for the "pre FM-X" mention and how accurate that is.

As far as I know, the FS1R did not term its implementation "FM-X". So I would think the FS1R (circa 1998) would pre-date FM-X. The FS1R implemented all of the same options:

sine - The operator will generate a sine wave which can be used for additive or FM synthesis.
all 1 - Broad band — including all harmonics.
all 2 - Narrow band — including all harmonics.
odd 1 - Broad band — odd harmonics only.
odd 2 - Narrow band — odd harmonics only.
res1 - Resonant broad band.
res 2 - Resonant narrow band.

And then it also adds:

frmt - The operator will function as a formant for formant-shaping synthesis.

(which FM-X does not include).

Speaking of ease to program - the FS1R ranked even lower than the DX7 due to higher complexity coupled with an even less approachable interface. The current generation is certainly much easier to program.

it seems FM took a back seat come the 90's

Because sampling started to drop in price. The sampled piano in the Korg M1 sounded more like an actual piano than most piano patches in a DX7. Any sample based synth I can think of uses subtractive synthesis which is what people who wrote or tweaked their own patches were familiar with from analog synths whereas FM required more time to get through the learning curve. As users of Yamaha AWM synths know, the combination of detailed samples and good digital filters can create a lot of convincing analog emulations.

Another reason was Yamaha held the FM patents, although they did license some 4-op technology to Korg who released the DS-8 and 707, FM synths with subtractive-like programming interfaces. Other mfrs had to adopt other digital synthesis technologies to work around the Yamaha patents and most chose samples in ROM. Even Yamaha went that route with machines like the SY22, SY77 and SY99 which mixed FM and samples much like Montage/MODX do today, and the SY55 which was sample based entirely.

@ Brian...

I didn't realise Yamaha held the FM Patents. That obviously explains a lot of the history I referred to.

If I was experienced enough to spot a trend right now (and I'm not), I see a lot of "Synth Modelling" technology being peddled.

I am familiar with modelling from a Guitarists perpective, which is getting to a point where pro's can't tell the difference between a real Vintage Marshall and the Modeller. Given a choice though, they would still rather have the Marshall. I have a "modelled" Electro Harmonix Electric Mistress... its the only way I could get that sound (without paying $1.5-2k for 44 year old orginal). It works, but I can still hear the difference (something is missing).

The problem with "models" is they are limited to whatever it is the programmers decided to "include" in the model. Admittedly this could be a long list of rare Holy Grail Synths, but it's like eating food someone else has prepared. I like the idea of FM being a presentation of raw ingredients with which to make your own meals. That makes FM still fresh in my opinion, we just need to get away from the cheesy pop sounds it got stereotyped with. :p

You generally get a lot less included in a sample although it can (and up until recently always has - for tranditional instruments) sound better. By less included I mean in terms of knobs and dials to turn in order to change the character. And the sampling is "stuck" with vibrato and other specifics unique to that recording session. I'm not sure models present a huge problem of "inclusion" vs. samples. They both have their opportunities for ball-and-chains.

I like the approach of having a ROMpler coupled with other synthesis engines. "Engines" (as in plural) may be greedy - but it beats hording equipment to have as much under one tent as technology, finances, and my back (weight) allows.

Note that one advantage more recent FM implementations have (including FM-X in MODX) is the availability of effects. The DX7 didn't have effects - so the FM engine itself had to be used to create chorus or other "effects". Now you have the choice to add many layers of effects to FM-X in order to further shape the sound.

@J

Your post got me thinking.

What if... effects could be applied to the Modulator output, before it hit the Carrier?

Total bananas could ensue.

:p

Yes, I was looking at that and was reminded (by - in all honesty - what I do think is a great GUI interface for the effects routing) that effects only apply to carriers. I created a real wacky FM-X Performance where sliders 1-3 affect the sound drastically of carrier (slider/op #4). And then have a "straight forward" lead type square wave in op #5 so I can mix crazy with something that is more static (pitched, not noise, familiar). Changing the levels of ops 1-3 make carrier #4 sometimes sound like a swirly ring modulator/phase/crazy-town sound - or sometimes a mellow sound and all kinds of surprises between. And carrier #5 allows for me to quickly go from noise to a "lead" sound or at least adjust #4 (noisy potentially) to #5 for times I want to come out of crazy-town "wall of sound" to do something more melodic and up-tempo.

During this sound design, I looked at the effects routing to see if I could selectively assign effects to operators 1-3. And, no, you can't do that. The interface is clear that only carriers participate. That was OK by me - even though I do agree that effects can add another dimension to the modulation.

Don't forget that each operator gets envelopes that can be used instead of employing an effect (that you can't do). That's part of what my program does - mostly uses pitch envelopes to modulate the modulator. I actually began this journey because when constructing the Performance my early experimentation showed that twisting the DATA DIAL for feedback (op #1) - it made my initial FM-X Part sound like it was speaking "Yama" - with a little softness on the "m". There was BM's mention of someone programming vocals through FM and I started walking that line. I automated the feedback knob twisting with motion sequence and had what sounded like vocoder without using the vocoder effect or using the formant effect. Then I branched out from there and got a very dynamic FM-X Part with easy controls using the sliders.

I think it is important to spend some time, early on, doing what you're doing walking through the FM-X tutorials and learning classic FM construction. Getting more of a foundation on what can be done when two operators interact (just a single modulator and single carrier). You can actually take this VERY far without effects, motion sequence, and controller assignments to destinations. Then motion sequence and effects become gravy on top of what can already stand alone very well even without it.

I've managed to work my way through Manny's series of FM-X Tutorials. I can't say at this stage that I have completely digested all that knowledge.

The tutorials move from basic fundamentals to advanced Super User very quickly. However I have them all bookmarked and will go through them all again at a slow pace.

For anyone else interested, there is probably just as much to learn in there about advanced Control Assigns and Motion Sequencing. His use of multiple Assign and Sequence control curves to step or sweep through exact parameter values across all Operators and all Parts is mind blowing.

And yes.... getting FM-X to emulate Classic Analogue synths is a big part of the content, especially once you get to the end of Article 3, and through 4 and 5.

Big thanks to Manny if he's reading this!