Two more interesting finds:

The triangle is not a straight sided triangle. It's seemingly kind of like the shape of a an inverted (1 - n) FM-X Odd1 created "square" wave.

0 (zero) for rate without extended range on is not zero.

And a bonus: Key Reset works very well, even for all notes played by an arpeggio. I wish this was true of Motion Sequence activation.

Disappointingly, there's no way to prevent looping of this LFO, otherwise it would make for a fantastic extra envelope.

I'm not sure I agree. I don't see amplitude responding in a square-ish way. I see a triangle attenuated some which would be normal. I'm using a crude measurement of an iPad software - so there's that too. However, the general shape is more triangle in amplitude than what you've described.

LFO Parameters, Measurement 1:

Scope Plot, Measurement 1:

LFO Parameters, Measurement 2:

Scope Plot, Measurement 2:

In order to set this up, I changed the AEG to 99 values across the board so when I let go of the piano key, the sound continues forever (hold). I changed to Monophonic so when I play multiple times the notes don't "stack". I changed from ratio to fixed and set the frequency to maximum so I would get less measurement error from slow moving sine waves. The waveform (spectral) was sine - starting with an initialized FM-X Performance. After setting the AEG and level (99, as default) on OP1 using the default Alg #1 - all carriers - I copied [SHIFT]+[EDIT] OP1 to all other OPS 2-8. The purpose for this was to work with maximum amplitude of the source. I also set the velocity offset to +64. I also turned off all effects although some effect levels were already not applicable starting with the initialized FM-X.

This is a carrier static output signal (output from respect of the carriers before LFO2). With no variation from velocity, trigger key pressed, number of keys, etc. Only LFO2 would have been causing changes in the amplitude and each measurement was taken without striking a new note (the note was setup to hold forever).

I could pull in better measurement instrumentation - but this seems good enough.

Also, I do agree that an "LFO Hold" feature would have some applications. That doesn't exist, though. It's not ultra high in my list of wishes even though I've run into this before.

@Jason.... thanks for posting those graphs.

@Andrew.... What you "detect" with a Human Ear, does not necessarily follow (or define) the waveshape of the LFO. I will explain.

First we need to look at "traditional" or Analogue LFO's. One can only assume in the Digital World, that "Traditional LFO" behaviour is modelled. In truth, in the Digital realm you can do whatever you like because there are no electronic limitations. However, my ears tell me that "modelled behaviour" is the path the Designers took. Jason's graphs would seem to support this. The waveform you can see in these graphs is known as "Hypertriangular", more on that later.

LFO's were an early attempt at automation. Essentially, they were uses to mimic a Human Hand twisting a (direct) Control Knob back and forth. This prevented the Human from becoming fatigued, freed up the use of a Hand, and allowed for repetitive consistency.

Typically they applied a varying "Control Voltage" (CV) to some other component of the System, that was actually in the (Audio) Signal Path. To be explicit, the LFO (actually a small "system" itself) was/is not in the Signal path.

So the Waveform you actually hear is being generated by the Component that is being subjected to the LFO's "command".

Looking at Jason's Graphs, what you can see (presumed) is a Triangle Wave LFO turning up and down a Logarithmic (aka "Audio") Potentiometer (I.e. an Analogue Volume Control "pot"). A Log Pot is designed like this, it is not "Linear".

The resultant waveform becomes what is collectively known as Hypertriangular. Another way this can happen is by "rectifying" a Sine Wave, that is turning all negative excursions of the Sine Wave into Positive Excursions using a grid of Diodes. Indeed, this "rectification" was sometimes employed directly on the output of a "Sine Wave LFO" to turn it into a "Triangle Wave".

In "non-musical", industrial systems, "correct" input and output waveshapes were perhaps mandatory, such that a lot more effort and expense was put into creating them (take for example Digital, On/Off, "1" vs "0" Square Wave communications systems).

In music, this extra expense was deemed unnecessary, and to boot, the imperfect waveforms were discovered to be, in fact more musical and "natural" sounding... such as musical "vibrato".

Input LFO vs Output Audio waveforms differ greatly, not only depending on the system employed, but also on what is actually being modulated.

Jason's example is Audio Volume (Signal Amplitude). Earlier I gave an example of Pitch Variation being generated by the Modulation of a Delay/Lag on the Audio Path.

In this case, the actual variation in Pitch is directly proportional to the "rate of change of Delay Time". So you end up with a mathematical "Quadratic" function thrown in for good measure (Google "Isaac Newton Calculus" 😉 ).

Now add to all this that LFO outputs, by nature, were "imperfect".

Ultimately, what it all boils down to is whether the output result is acceptable or unacceptable to the requirements of the Task at hand. In Music, this translates to whether, subjectively, the output sounds "desirable" or not.

Testament to this is the advent of early "Bucket Brigade" Modulation Pedals (Chorus, Flanger, Delays, Phasers, etc) in the late 70's. Electro Harmonix "Electric Mistress" and Roland "CE1" got their LFO design "right" by happy accident, where many others fell by the wayside (often using exactly the same components, just different implementation). Working Electric Mistresses will now cost you a lot of cash. Similarly the "famous" Roland Jupiter and Juno "Choruses", which are now "modelled" as accurately as possible in modern Digital synths. Modelling "analogue imperfection" it turns out, is very hard indeed.

@Jason

It was a piss take.

That's exactly (somewhat) what the Odd1 "square" waves look like, if you 1-n them.

Similarly, calling the All1 a "saw" wave is a stretch.

The output is not reflective of the modulation itself. What I mean is the output is not a 1:1 to the movement of the LFO. If you take a look at level - you see that levels 0-25 are much finer than levels 75-127. Level, as a parameter, doesn't output a linear result. However, the triangle is a linear modulation. So the end result is a byproduct of the level parameter. And still the LFO is really progressing through a triangle shape.

That's why you see the rounding out near the crossover point where "silence" is in the middle of each triangle shape.

At a Depth of 80 and Operator Depth Ratios at 7 - this gets in the range of not squelching and provides a more "expected" result.

Still a triangle modulated level (parameter). There isn't a mode to make the level parameter linear. Although this "could" have been adjusted for the LFO only - I think reuse is king here and not in a bad way. Some system to build user LFOs with various curves could also allow the user to create a more linear output and get to the place where you expect. Although there may be a complaint of why you need a curve to get linear and a straight line to get a curve. ... it's due to the behavior of level itself.

@Jason I get what you're saying... paraphrased:

"It might be a triangle wave, but it's being used to do this... [insert maths and outcome]"

That outcome, with the Pitch, is probably different with Filter Modulation, where there appears to be some rounding (easing) at extremities and at the midpoint. Which makes me think they do 0 to 1 and invert it for -1 to 0, rather than a calc through -1 to 1.

edit:
Here's an exaggeration of what I hear happening to the Filter at extremities:

Also should add:

When there is a visual representation of a triangle wave used, I'm more inclined to expect linear modulation than those willing to make excuses for it not being that.

It's gotten to the point where both terminology and graphical representations should be presumed peculiarly Yamaha's way of defining some impregnable level of what's going on rather than what is actually resultantly occurring.

@Andrew... please see my earlier post.

It seems you believe/expect the Audio Output to mirror whichever LFO waveshape you have chosen.

In reality, it is very unlikely that the LFO waveform will be "seen" at the output, but it is also musically undesirable. Exaggerating the point, hearing the LFO will sound robotic, clinical, mechanical... "un-natural".

In reality, getting an "output" to mirror the "input" of an external signal (LFO in this case) is actually quite complex and hard. Thankfully not required in basic applications like musical effects.

If you listen to a "hard" (max Depth) Tremolo, the effect is sometimes described as "throbby". There's a lot of mood, character and "swampiness" to that "musical" sound. Behind that real output is, in effect, a Square Wave... an electrical "Switch" of some sort turning Fully Off, and Fully On.

The effect of that On/Off influence on the Audio path is subject to something called Hysteresis... Slewing, Lag or Sag in common parlance. The Audio Signal is unable to "react" immediately to the change in "influence", I.e. the ON/OFF of the LFO. (Have a listen to Ry Cooder - Paris).

With modern technology, and especially with DSP, if we want the Audio to react exactly in Sync with the LFO, we can. The question is... would you really want to?

Take the "Square Wave Tremolo" as an example. If you turn the Audio Output to react exactly to the LFO you end up with a "Slicer" effect. This is a very stuttery, percussive, "in yer face" effect. Useful for adding highlights and "impact" to a section of music (listen to Bon Jovi - Keep The Faith). But... I can pretty much guarantee you would not want to hear "slicing" throughout a whole song.

There is a "Slice" effect in the Insert FX of the MODX. There are also numerous traditional "Tremolo" type effects. You can experiment, and be your own judge of preference.

Despite having the power of DSP, it is clear Yamaha intentionally modelled the behaviour of Traditional LFOs (as per "Tremolo"). This is exactly what you would expect. If they didn't do this... they would likely have quite a lot of unhappy customers.

Don't overthink it. Don't be upset because your Oscilloscope App does not show a perfect Triangle Wave superimposed on your Audio. Simple... if you don't like the sound, try a different option... Sine, Triangle, Square, Saw Up, Saw Down etc. Options. There's hundreds of them.

Antony, I'm discussing (here) the behaviour of Filter Modulation.

Nothing else.

I think you're overthinking it, and bringing in many other aspects of many other parts of sound(s).

Hence your last paragraph.

[quotePost id=116844]
Also, I do agree that an "LFO Hold" feature would have some applications. That doesn't exist, though. It's not ultra high in my list of wishes even though I've run into this before.[/quotePost]

A switch, both visual and in code, is all that's required to add this, and it'd benefit arps and patterns, immensely.

As it is now, there's a bug - in playing an arp vs replaying a recorded Pattern of an Arp (now it's just notes) responding differently to the other triggering mechanisms for LFOs.

@Andrew....

I work in the big corporate global Telecommunications Industry and God help me if anyone found out that secretly I am also a musician.

I try to partition this aspect of my life away from the day to day befudgery of corporate idiots... and believe me... in the Comms/IT business they are packed up to the ceiling.

So please excuse me for addressing persistent musical realm queries with some background technical insight. I have been kind and spared you and anyone reading from "the math".

The point is, I view my MODX as another musical "instrument" along with my guitars, my amps, my effects pedals and all auxiliary equipment that allow me to create and enjoy creating, music.

The MODX ticks all the boxes in this respect.

By contrast, my Dell Laptop PC is a frequent source of irritation. It actually prevents me from doing my job on a regular basis, and I find myself consistently having to frequent their "Support Forum". For real, genuine, others have suffered, problems.

My image of you, is sat with a MODX, with the screws and top panels removed, ribbon cable taps, hooked in test equipment with real time Wireshark analysers, whose only enjoyment is to rip it to shreds and find anything that doesn't fit your microscopic myopic view of the world.

Have you actually tried to make any music with the MODX since you bought it?

How many other Synths do you have laid bare on your experimental bench?

How many other Synth Journal/Forums do you frequent in order to pour your hatred of the world onto unsuspecting patrons?

To be honest Andrew.... you are overthinking it, and that's being polite. Everybody here knows it. Very few people though have the gumption to say it.

I suggest you buy a Dell Laptop. You will have a riot over on their support pages. And maybe give us some peace in the meantime.

You love speaking for others and leaping to conclusions (there's a mat for that), I'll give you that.

Perhaps you're more willing to overlook and/or heap praise on idiosyncratic software's (fixable) shortfalls, flails, issues, uncertainties and the laborious and irksome UX of an otherwise quite good product.

I'm not.

I hope Yamaha hopes to do better, and the stuff I write about is just the "tip of the iceberg" that I think either could be better explained (this thread, for example) for greater general empowerment of a more common audience, or made better to use, and I can be bothered flowing forth about it.

Not a lot of thought in it, at all. Mostly strains of streams of consciousness shaped somewhat into responses (to the MODX experience) articulated (poorly, often, I'll happily admit, because I just spew it out, rather than overthink it. I type very fast, think very slow, respond with knee jerks and can lift heavy things) whilst waiting for a computer doing something else.

Almost everything I write is written with a good faith wry smile. You don't read any of it that way, which is a pity. Am sure you could and wish you would.

[quotePost id=116868] And maybe give us some peace in the meantime. [/quotePost]
That would instantly increase the quality of this place by a significant degree.

[quotePost id=116869]You love speaking for others and leaping to conclusions[/quotePost]
Obvious conclusions shared by many.

[quotePost id=116869]Perhaps you're more willing to overlook and/or heap praise on idiosyncratic software's (fixable) shortfalls, flails, issues, uncertainties and the laborious and irksome UX of an otherwise quite good product.

I'm not.

I hope Yamaha hopes to do better, and the stuff I write about is just the "tip of the iceberg" that I think either could be better explained (this thread, for example) for greater general empowerment of a more common audience, or made better to use, and I can be bothered flowing forth about it.[/quotePost]
A deluge of unarticulated complaints, insults, pompous postulates about how a professional musical instrument should be designed coming from the one member of this forum who can't make or play 10 seconds of actual music.

[quotePost id=116869]Not a lot of thought in it, at all. Mostly strains of streams of consciousness shaped somewhat into responses (to the MODX experience) articulated (poorly, often, I'll happily admit, because I just spew it out, rather than overthink it. I type very fast, think very slow, respond with knee jerks and can lift heavy things) whilst waiting for a computer doing something else.[/quotePost]
Ah, so that's why.

[quotePost id=116869]Almost everything I write is written with a good faith wry smile. You don't read any of it that way, which is a pity. Am sure you could and wish you would.[/quotePost]
You got it backwards.
If your "good faith" ???????? is not coming across, it's your fault, not others'.

[quotePost id=116870]Frankly, I fail to see how anyone can discern someone else's emotional, or mental, state based on mere text.[/quotePost]
In this particular case it's easy, cause the subject pours his generous soul into each post, in good faith, for everyone to read.

@Bill!

I think you're overthinking it 😉

"But some of us like to peek behind the curtain and try to understand what is going on. Is the curve linear or exponential or something else. Did my ears just get 'lucky' or can I use my understanding to more easily find other combinations that work well?"

You C!

I.C.

We all C!

From 6502 assembler to the ST and shining C... to distant shores.

The original question was what's being modulated. And for amplitude, level is being modulated. And although amplitude is, in turn, the result - I can see how it could be interpreted as incorrect marketing/labeling to say an oscillator is modulating amplitude directly. This isn't true - what is being modulated is level. In the shape of each waveform. And, as a consequence, the final outcome has "acceleration/deceleration" characteristics intrinsic to the impact of level to the final amplitude.

Determining level's relationship to amplitude is fairly simple. Set the level to 0 and create a source/destination pair under Control Assign where the an assignable knob is the source and say, for FM-X, the FM-X parameter of operator level is the output. Assuming algorithm 1, you can turn on every operator since they are all carriers. Assignable knobs have a menu where you can see the actual value. This goes for Super Knob as well - so either will serve. You can plot the graphs of input value (assignable knob value 0-127) to output (amplitude as captured by a scope).

However, I really think a piece that's missing here is the scale. If you notice the graphs are are linear in scale. Audio is not a linear phenomenon.

Source: Random search engine search

Logarithmic potentiometers are used as volume controls in audio equipment because the response of the human ear to the loudness of sound is also logarithmic.

So a simple change in the graphing method used would give a better adjusted look at what's going on. Which lands us back to amplitude as related to human perception.

Logarithmic slopes on a linear graph (scope) are perceived by us as being linear.

The "log" line above is the triangle waveforms ascending behavior while the "lin" line above is what we perceive. Thus landing us back to "perfect" triangle rises and falls when we compensate for our own wiring.

Back to marketing - I think the labeling is fair as-is. It just takes a little while to get to why.

@Bill... if I understood you correctly, I think we are on the same page.

When you described the "summit of understanding" you are correct. When providing technical background, one has to be careful not to lose the audience.

In some cases a peek under the hood can help musically, such that the musician feels more in control of their output.

Synths are by nature very technical instruments. On the other hand, not understanding the technicalities does not prevent somebody from making music with them.

Before owning a Synth, my only musical outlet was the Guitar.

On Guitar I realised that Modulation was, and still is, a massive source of confusion for a lot of players. Even a simple Chorus Effect, is beyond the patience of some players.

A common scenario is a Guitar player buys a "Chorus" pedal, to sound like his Guitar Hero. Having spent a considerable amount of cash on said Chorus, he discovers that he still sounds nothing like his Hero, and publicly "damns" the Pedal as a P.O.S.

The reality very often is, he just hasn't set up the Chorus controls correctly.

Setting up the controls is easier said than done. If you cannot relate what you hear, to what is "written" on the Controls, then you are subject to a lengthy trial and error process.

By contrast, if you have some visibility of what's going on "under the hood", you stand a better chance of dialing in the right sound, in a timely manner.

Modulation pedals are famously non-intuitive. Add to this, if you buy 10 different brand Chorus pedals, no two will sound the same. There is no Standard, nor should there be. Equally, no user has the right to demand some arbitrary standard (other than it works safely).

My "knowledge" of Modulation stems from studying David Gilmour (Pink Floyd Guitarist).

In my opinion, he produced some fantastic sounds from his guitars, which seemed tantalisingly easy to achieve, but in truth were the result of a dizzying array of Modulation... he is famous for it (Andy Summers is another example).

In order to "get close", at some point I realised I needed to deep dive into the various forms and brands of Modulation he used.

Gilmour also did not limit himself to just using "Guitar" Effects. He sometimes passed his Guitar through "Designed for Synth" auxiliary equipment... because he had greater control over the LFO!! (No wonder I could never reproduce those sounds).

I suppose the underlying theme in what I have portrayed is that there is rarely any direct correlation between an LFO waveform, and the Audio Output*. Moreover, what really matters is whether you like or dislike the Audio result.

You can't blame the LFO, or the people who made it. All you can do is try things differently.

(* Usually there is a 1:1 correlation to the "speed" or frequency of the LFO. Even so, it may not always be apparent).