What everyone is trying to tell you is — you need to actually go through the tutorial articles and actually play around with the parameters. Best way to learn.

If you don’t know what a Sawtooth wave is or what a Pulse wave is... STOP, do some research on your own.
If you have a background in analog synth, the description of Saw and Pulse are meaningful... but, if you have no background in analog (voltage control) synthesis, then as a musician Sawtooth and Pulse are just meaningless words that describe a wave shape. And may not be as helpful as you might first think.

Those who have experience with analog synthesis can lean on this background to “jump start” their programming. But if you are like most musicians, you have little concern for wave shapes and their impact. And like most musicians you wouldn’t know what wave shape you are listening to, ever.

You could be the highest paid, best oboe player on the planet, without ever knowing that a narrow pulse waveform is best when attempting to emulate an oboe with a synthesizer (why would you need to know this? — point is, you don’t).

What those with analog synth background have is just a strong sense of the fundamental starting point... much of the rest of analog synth programming is trial and error. Most folks have no idea how to program anything... they tweak and “discover” stuff, then they refine it, and tweak some more... later, a few will try to impress you with their knowledge, hiding the fact that a lot of it was first discovered by trial and error and pure luck. This is reality.

When we try to describe sound with words... although a poor substitute for hearing the actual thing... but nonetheless, helpful to approximate what is happening and therefore it helps get us started.

Sawtooth waves have all the harmonics (as a musician, if you don’t know about “harmonic” - then you need to look that up. It’s science, and doesn’t change throughout nature. As a musician you have no excuse... know the Harmonic series... the ear/brain uses it to recognize sounds. You recognize your mother’s Voice because the harmonic content is unique. You recognize the difference between a trumpet playing A440 AND A trombone playing A440 because, like a fingerprint, the harmonic content has a unique character. It’s HOW we recognize sound. If I were an impressionist, I would attempt to imitate your voice’s harmonic content ...

Pulse waves have every other harmonic (just the odd numbered ones)... it is the relative volumes of these harmonics that make the different sounds.

The drawbars of a Tone Wheel organ represent individual harmonics. Great to know all of this...
But so what? Unless you are curious enough, you will never figure out what you can make from them.

Sawtooth waves are used to build string and brass ensembles. Orchestral strings and Brass ensembles are made from waveforms with all harmonics. It is, in fact, hard to tell strings from brass if you do not hear the Attack portion or the end portion of the sound. Same as you could not tell the word “Squeeze” from the word “Pleasing” if I were to edit off the Attack and end portions. If you record someone singing a song with these two word in it, they are identified by the all important factors of how they start and finish.

The Attack and Release (synth parameters of the Envelope Generator) are used to create the Attack and End portions... and help us make sense of vocal sounds and musical sounds... The bowed attack — or how bowed strings rise up gradually from silence... versus how blown brass has a more abrupt beginning.... these help allow our ear/brain to recognize the difference. It’s sort of how the “s” sound of “Squeeze” comes up from silence; it rises more gradually than the rather explosive beginning of the “p” sound in the word “Pleasing”.

It is said that the ear/brain catalogs billions of these wave and envelope shapes, it allows us to exist in this world without completely freaking out. If you hear something that you cannot identify, your ear/brain will interrupt you from whatever else you are doing to try to figure out “what was that”. If it is one of the billions of sounds you already have cataloged it may not even give you pause. If it is a recognized sound but in a situation where is should not occur that will get your full attention, too.

So synth programming is about manipulating a few known functions, combined with a lot of trial and error. Pick the closest starting wave, and shape it using your Pitch Envelope, Filter Envelope and Amplitude Envelopes to shape it up.
Working with Samples (is much easier, because most of the work of matching the initial wave’s harmonic content is done) is different from working with other forms of synthesis, where creating the original wave is the initial WORK, then comes the shaping of the envelopes.

If you know that sawtooth waves are like the teeth of a saw... they rise slowly and drop immediately, or they rise immediately and slowly fade out, and are used to build strings and brass. While pulse waves... they are full On or full Off... and are used to build woody, hollow sounds when they 50% On, and 50% Off, and the narrower the pulse (the less time On compared to Off), the more nasal the result.

After awhile, it no longer is going to sound strange when you are told you can build a clarinet (woodwind) sound with a Pulse wave that is 50/50 On/Off (called a Square wave), and you can fashion a clavinet or an oboe (both nasal sounds) with a narrow pulse wave (25% On, 75% Off or 10% On and 90% Off).

Soon when you hear a sound, you will immediately “know” where to begin. You get there through “experience”.

Experience is a wonderful thing, but there is no shortcut to experience.
So you want learn FM-X or anything for that matter? ...you have to roll up your sleeves and dive in...

Oh, yes...and I will be able to do it thanks to people like you.

Keep on writing, please...

Fernando, you write well!

And I agree.

In an attempt to provide some info... FM is, at minimum, in terms of moon range of audio, half way there, even in its most basic modern forms.

However, MODX/Montage go way past the Solar System and out into galaxy wide options, through dynamic, programmable modulation (Motion Sequencer), arpeggios of immense complexity... and... both arpeggios and Motion Sequences being able to control EFFECTS, too!

FM-X, due to the nature of its sounds, is highly responsive to effects, and Yamaha has tamed an enormous number of them to put within the MODX/Montage.

Without animating them, just the static Effects alone put the MODX/Montage approach into the Solar System journey levels compared to the moonshots of things like the Digitone and OpSix. Although, in fairness, the OpSix is a Mars mission compared to the more mundane Moon mission of the Digitone.

In further fairness (and this is where things get a bit weird), this means FM-X is at least two orders of magnitude more powerful than the others because of the extra two operators AND the near complete integration with animatable modulations able to target both properties of the FM-X Operators/Carriers ... AND ... the values of the EFFECTS!!!

It's difficult to explain how much power and creativity this provides, as it's absolutely immense and nobody has done anything remotely close to explaining the breadth and depth this provides. Nearly 5 years into the Montage's life!

So it's much like a largely unexplored galaxy, at this stage. I've not seen anyone do anything even remotely approaching getting outside our mere Solar System when it comes to utilising these creative options.

Keep in mind, all of the above is within merely considering one part. That the Yamaha way is to then permit profound relationships with 8 directly controllable FM-X parts, each of 8 operators, each with 2 Effects ... AND... integrate that with arpeggio control (Arps can control lots of properties, too), as well as the Motion Sequencer, means a drawn to scale map of FM-X vs the others would render them so small as to be invisible.

Keep in mind, all of the above is within merely considering one part. That the Yamaha way is to then permit profound relationships with 8 directly controllable FM-X parts, each of 8 operators, each with 2 Effects ... AND... integrate that with arpeggio control (Arps can control lots of properties, too), as well as the Motion Sequencer, means a drawn to scale map of FM-X vs the others would render them so small as to be invisible.

Well said!

Now that you talk about parts...

Fantom has 16 zones (parts in Montage) with up to 4 partials (operators in Montage).
Montage offers 8 internal parts with 8 operators.
Let's forget for the time being those 16/8 external available parts in both...

I see then 64 stereo voices/waves in both...
Is there any advantage in any...?

I ask because I am so bad at kooking that I open the eggs with a can opener....

True, but I can stand in front of an old biplane telling everyone I want to reach the Moon...
At times, it is best to be told what you have at your disposal. You have a biplane there is clearly not enough. You have a biplane there that climbs up to 15000 feet is a step further. That's it...

I understand your biplane metaphor. Whilst window-shopping various synths and packages it is difficult to gauge just how powerful or capable a system is, until you have sat in the pilots seat and "flown" it at full throttle. Or at least, been taken for a test flight.

Like @ Dans original post... which is best? Again difficult to gauge unless the potential buyer has some idea of their ultimate application/purpose (are we going to the Moon or Crop Dusting?). The dilemma is, the buyer is unlikely to know the purpose or goal, without some "in the seat" experience. Reading marketing blurb doesn't really give you that.

For FM-X, "How high can it fly?", I think is readily answered in the video below. It's from Manny's FM Xpert Tutorials, Article 1. All you have to do is listen to the first 50 seconds... and realise that is not a Grand Piano, it is FM-X. No samples used in this video. Treat this as your test-flight.

In my opinion, this is testament that FM-X can fly very high indeed. As @Andrew says... Solar System high.

Manny's choice of a Grand Piano here reminds me of Kennedy's Moon Speech "We choose to do these things, not because they are easy, but because they are hard".

By all means watch the rest of the video, but don't expect to understand what all the parameters are doing right now. However you can understand each parts contribution to the overall sound because Manny shows you.

Hope this answers your question.

Downloaded; thank you, Antony.

The Opsix has less preset algorithms than FM-X. However, Opsix allows for user algorithms. This really opens up lots of possibilities not available in FM-X. User algorithms allow for each modulator to apply feedback to itself. FM-X has fixed feedback only one operator which receives feedback. Opsix can assign every operator with self-feedback. This is more similar to the implementation in the Reface-DX with respect to feedback flexibility. Beyond that, user algorithms allow for you to assign the output of one operator to any other operator. There is assignable chaining (op n -> op n+1). And there is also arbitrary carrier assignment. Meaning that any of Opsix's 6 operators can be configured to produce audio output (i.e. carrier). Due to the flexibility, you can assign carriers as modulators for other operators. In this case - the control surface will show a blended color for operators that are both modulators and carriers. FM-X doesn't do any of this. This is more of a feature available in software FM engines such as FM8.

Having options like this is awesome. To have only one fixed feedback on only one operator is often limiting when programming a new sound comprised of a few operators from 2 or 3 other FM-X PARTs using an algorithm that fits, especially if a couple of the pieces I want to use from those several sounds are using a feedback operators. I often end up wasting/burning an additional PART just for 2 to 3 operators of sound, because no algorithm has multiple feedback operators or not enough carrier/modulator stacks. And in too many of the Montage algorithms the 2 additional operators (usually 1 & 2) are merely single carriers, which are usually useless for most use cases when I am programming via combining parts of algorithms/operators from several different FM-X PARTs. So for the most part the Montage only has 6 useful operators for most cases when I am programming an FM-X sound. It would be nice not to have to burn an extra FM-X PART just to add a carrier + modulator stack!

I would like to see Yamaha provide additional options in FM-X on the Montage. 1. Adding new algorithms with the additional 2 operators stacked & not as single carriers, would be awesome! 2. Having multiple Feedback operators in a single Algorithm would be a huge bonus as well, and would be extremely useful when programming.

I can't express how much having these additional features/functionality in the Montage FM-X engine would be so useful!

^^ 100% agreed. Darryl!

If an operator has nothing above it (in an algorithm diagram) it should be possible to adjust its feedback.

@Darryl...

The way I see it, a less powerful 6 Op FM Engine has to introduce more flexibility into its operators to compete with a more powerful 8 Op, multi-Part FM engine.

Using the video above, the "Grand Piano" was built from just two 8 Op Parts, Parts 1 and 5.

The other six 8 Op Parts were used to build what Manny calls "eccentricities and stuff" - keybed thunk, wound and straight string enharmonics, etc.

If I recall correctly, Manny is a Professor of FM Synthesis. Therefore Manny has the knowledge and experience to use the leftover 6x 8 Op parts to "ice the cake".

I have to wonder how much of what he builds is literally beyond the wits or understanding of most "normal" keyboard users.

Simply, having 64 Operators to play with is way more than most users will ever have the need or inclination to use.

Adding optional feedback on each of those operators, and adding user defined algorithm paths, is likewise just adding more functionality that most Montage/MODX owners will never use or even understand how to use in a meaningful manner.

It is potentially casting more pearls before swine.

I read in one of those Yamaha articles (I think FM101) that Feedback loops were useful in a pinch (running out of operators), but once you got to large (8) operator algorithms, and multi algorithm performance capability, feedback becomes redundant. It substitutes for a 2 operator 1:1 Ratio, 1:1 waveform stack. They conserve operators, if operators are limited.

All that said, I notice many posts on ideascale that request multi-feedback and custom algorithm options to be built into a future software release. I suspect that these requests are not driven by user encountered limitations, but merely by observation that 3rd Party Product X has this option and Yamaha doesn't.

There's an old saying in the UK, "Give an Inch, and people will want a Yard". Keeping up with the Jones's. They don't need a fan driven air dried brick and clay pizza oven in their back yard, but by gosh, if Jones has one, they want one too.

Look at it from Yamaha's perspective. There's already droves of users complaining about the over complexity of the FM-X engine and its user interface, while a seemingly equal proportion of users are complaining it is not complex enough.

Maybe Yamaha will add such functionality. They have included it on previous, smaller, FM engines; it isn't that they don't have the means. Then maybe the world will be a better place?

But seriously, how likely is it that you will ever need it? Or more directly, how likely is it you will ever get round to even consider using it? More probable I suspect is you will select a Preset and just run with that.

There are some draw backs that Manny alludes to in quite a few of his articles.... the lack of a "Phase Option".

He suggests that due to the nature of FM (any FM, not just Yamahas FM) the Bessell function kicks in and creates non-uniformity in the energy distribution amongst harmonics. According to Manny, this is both a blessing and a curse. He ponders that if FM engines allowed Operator Phase to be configured relative to other Operators, the Bessell function could be more easily controlled or mitigated. This would reduce workflow, but you'd still need the requisite knowledge to make beneficial use of it.

I notice there are some basic phase options in the OpSix, but not comprehensively implemented as per Manny's recommendations. So on the OpSix, its a "neat trick", but not much else.

What it all boils down to is gathering real experience with FM. If you are solely focusing on becoming an FM Expert, I'm sure either MODX/MONTAGE or the OpSix will happily carry you on that journey.

But it won't be until you have been round the world a few times that you can be truly informed of your expectations and desires regarding your preferred FM "vehicle".

Antony, read Darryl's comments a little more closely.

He is describing things simply because he well understands them.

He might know more than most of us about FM, and then some.

@Antony...

Maybe someday I will take a bit of time and go through Manny's videos; however for the foreseeable future, I will likely continue programming FM-X the way I have been doing so since over 30 years. I have my DX7II-FD on a stand adjacent to my Montage, and plug it into the Montage A/D Input for comparison purposes only (in real time), as the conversion of all my programming I did on the DX7II over the years doesn't translate perfectly on the Montage, so it requires a few tweaks to get the FM-X PART(s) to sound exactly like the DX7II. Notice I continue to type DX7II & not shorten it to DX7, because layering is essential for me, and the DX7II is basically 2 x DX7's in one + it has Unison functionality that makes each of those 2 layers sound more like 4 to 8 layers each, so basically a DX7II can sound like up to 8 to 16 layers (at a great cost to Polyphony), but a DX7 can only do & sound like 1 layer because that is all it was designed for.

I try to utilize FM-X on the Montage as much as possible and try to have close to 50% of the sounds to be FM-X PARTs, because it doubles my Polyphony on the Montage, and I can usually make the FM-X PART(s) sound exactly like the AWM2 PART, or at least close enough that most people can't tell the difference.

I notice many posts on ideascale that request multi-feedback and custom algorithm options to be built into a future software release. I suspect that these requests are not driven by user encountered limitations, but merely by observation that 3rd Party Product X has this option and Yamaha doesn't.

Maybe Yamaha will add such functionality. They have included it on previous, smaller, FM engines; it isn't that they don't have the means. Then maybe the world will be a better place?

But seriously, how likely is it that you will ever need it? Or more directly, how likely is it you will ever get round to even consider using it? More probable I suspect is you will select a Preset and just run with that.

I can assure you, my intentions with wanting multi-feedback and having the 2 additional operators stacked instead of having 2 almost useless carrier operators, are not because some other product X has these options. TBH, I don't even know what other products can do with regard to FM-X, except for what has been posted in this thread, mainly by Jason. I legitimately could use these things, especially the way I program FM-X.

Sometimes I do program the odd FM-X PART from scratch; however the vast majority of my programming for the past 3 decades on the DX7II & now FM-X is to find/use existing FM-X PARTs/sounds and combine them into one FM-X PART if possible, to get the sound that I'm going for. It's all about finding sections/operators of 2 or 3 FM-X PARTs that when combined in one single Algorithm (PART) & a lot of tweaking, give me the sound that I am trying to create. That's how over 90% of my FM-X programming is done. I kinda think of it as taking a bunch of pieces from different Sounds/Algorithms & layering them together in a single Algorithm that fits. Usually I can find an algorithm that does fit, but sometimes 2 or more of the pieces that I am trying to combine/layer either total up to 8 operators (none of which are just a carrier operator, so all those FM-X algorithms on the Montage are useless in these cases, and are basically only 6 useable operator algorithms...same as on my DX7II) or they each have feedback operators, whereby I can't combine or layer them into a single algorithm and need to burn/use an additional FM-X PART that essentially only uses between 2 - 4 operators...so it's a bit wasteful.

Hence why for me, these enhancements would be extremely useful when programming FM-X PARTs.

What it all boils down to is gathering real experience with FM. If you are solely focusing on becoming an FM Expert, I'm sure either MODX/MONTAGE or the OpSix will happily carry you on that journey.

But it won't be until you have been round the world a few times that you can be truly informed of your expectations and desires regarding your preferred FM "vehicle".

So as I started out with, I do have quite a bit of experience with FM-X, and I'm probably in the minority of people who utilize FM-X on their Montage or MODX almost 50% within a Performance.

This Performance that I recently completed utilizes all 16 PARTs in the Performance I created. I programmed all the sounds/PARTs from scratch and ~65% of what you here is the DAW (Pro Tools) based sequencing that I recorded, and the other 35% is the PARTs that I play on the keyboard live:

https://soundcloud.com/dclowe/montage-justin-bieber-intentions

7 of the PARTs are FM-X, 7 are AWM2 & 2 are DRUM/AWM2's. One of the Drum PARTs is just the Touch Tone Phone sound you hear every now and then throughout the song, which I could have done as an FM-X PART & made it sound exactly the same, but for the sake of time & since I already had that sample, I chose to load that .wav file into a DRUM PART instead.

Out of the 7 FM-X PARTs you hear, 4 of them represent 2 sounds/PARTs that I was trying to program; however because of the limitations of the Montage's FM-X engine with regard to single feedback operator in an algorithm, I couldn't combine/layer the 2 sounds into a single algorithm, but also because I needed 1 extra carrier/modulator stack, and the only Algorithm options that would have fit (if they could do multiple feedback operators) had Operator 1 & 2 side by side as single carrier operators, when I need just 1 more stack!

Just as one more example with regard to how I program FM-X, so that you get the full picture of why this would be so beneficial to me and probably to many of the other people who are requesting/'Up Voting' these FM-X enhancements on IdeaScale, here is another SoundCloud clip that is an Audition I recorded for a C7 Piano library that I sample via SampleRobot and Programmed based of the CFX Stage piano originally:

https://soundcloud.com/dclowe/montage-c7-strings

If you listen to the layers of strings/voices closely, I have orchestral strings, choir voices and a synth based multi-unison voice strings all layered.
The multi-unison voice strings that I had programmed close to 30 years ago, by finding 3 sounds/voices on the DX7II that I took sections of their algorithms, found an algorithm that would fit and combined them into one single Voice. Thankfully the operators from 2 out of the 3 voices didn't have the feedback operator & I was able to find an algorithm that they could all fit on and be combined/layered. It's algorithm 16 on the DX7II-FD and on the Montage it's algorithm 26. Note, I could probably enhance it even more, but for the type of sound, the single carrier operators 1 & 2 on the Montage are completely useless, so their volumes are at 0. If I recall correctly, the 3 sounds I used to create 1 single voice on the DX7II were angels, ghosts & pan flute, and with utilizing Unison Poly with Unison Detune set to 4, it made it sound like closer to 6 to 8 voices layered. On the Montage, in order to keep it to just 1 PART instead of 4 PARTs(all detuned from each other) I had to use the 'Symphonic' effect to emulate the Unison layering functionality on the DX7II...

Nor will you find an engine that comes close to doing as much as the FM-X engine in the current Yamaha synths.

It is not quite true. AFM on SY-TG/77 and SY99 can do things which can't be done in FM-X (more waveforms in operator, free input, free feedback, free algorithm, noise input, AWM sample input, more complex envelopes with looping...). Also FS1r has 8 additional noise operators...

Somebody from a different forum sent me a .pdf document... an "A-Z" of FM Synthesis... it was a University Thesis, intended as a beginner to expert guide. You're head would explode after about Page 6. And there were around 200 A4 pages of small font to get through.

Could you please send a link to this document?

@Daryl.... apologies if what I wrote sounded like a personal retort, it was certainly not meant to be. I was just hoping to add my voice and general opinion to the whole discussion by picking up on some points you made.

I appreciate after what you've replied that you are an experienced "FM-er". So in your case, valid requests. You are, with hindsight, one of the people I refer to that has "travelled the FM Globe".

I still suspect however, that a lot of wish list requests are driven more by "grass is greener" consumerism, than by "working musician hardship".

I sincerely hope that Yamaha do "upgrade" the FM-X Engine. How much benefit that will provide to the majority of users will no doubt continue to be debated.

Cheers,,

Tony.