YamahaSynth.com Forums

This is the place to talk about all things related to Yamaha Synthesizers!
  1. Matthias
  2. MODX Series Synthesizers
  3. Thursday, 19 September 2019
Dear Forum,

after i had three weeks of fun using the presets of my new MODX6 and building first own performances, yesterday for the first time i spent some time with the FM-X engine of my MODX6, starting from the init performance and trying some fundamental things. Unfortunately, i discovered a surprising (not nice) effect: Even when i play with only one oscillator simply producing a sine wave (all other oscillators/modulators are at zero volume), strong clicking noise and distortions are induced when several notes are played simultaneouly. The effect, which seems to be an overdrive problem, is getting stronger the loader the signal is. For the algorithm with 8 oscillators (No. 1) about 6 notes have to be played simultaneously to start the clicking noise, for algorithms with less oscillators (e.g. No. 37 with 3 oscillators) it starts with 4 simultaneously played notes. When i add the second oscillator with the same (not detuned) sine wave, very strong distortion arises even when a second note is played at the same time. Tuning down the volume of the part seems to help a little bit, but can not remove the effect. Tuning down other volumes (perfomance, master volume) don't lead to an improvement. Has anyone also discovered this effect? Is this a malfunction of my MODX6?

Thanks for your answers !
Responses (6)
Bad Mister
Yamaha
Accepted Answer Pending Moderation
If you raise the Output Level of several Oscillators to maximum, it is normal behavior to start generating a sound that will eventually distort the system, this is no more unusual than if you take a resonant filter and turn the value toward maximum, it is possible to get the filter to self-oscillate (a condition where it generates its own ‘howling’ tone independent of Key-On). It is simply a behavior. Digital filters howl differently than analog filters (this too is a behavior).

You are not prevented from such things occurring. In one of the tutorials on Super Knob programming, in showing how it can adjust more than one parameter, a linked Assign Knob is turning up the Resonance, while another linked Assign Knob is turning down the output so that before it begins howling it’s output level is reduced. Conclude: at the edge of chaos is where some of the magic occurs... or you are not prevented from pushing the system beyond the musically satisfying tones, you can generate “noise” (non-periodic)

With Sine Waves the acoustic build up of energy, becomes like a torture test for sound systems. No inner movement, no constant dips an valleys in the energy distribution as would be found in Waveforms rich in harmonics; sine wave energy builds up rapidly. It perhaps explains why organs never developed a Sustain pedal (and it had a 2000 year head start). Listen closely for acoustic distortion, it is a part of sound design. And happens in acoustic instruments as well (but that is another discussion).

When it occurs can be different for the type of waves being combined and what pitches you are combining in the chord. Notice however, the behavior when your Init Sine Waves are combined in a chord is different from when combined playing a single note.

You might use the same number of Operators to create a mono sound, that when combined in a chord cause acoustic anomalies... the same number of Oscillators stacked and sounding a single pitch simultaneously, is not a problem. That’s a behavior.
  1. more than a month ago
  2. MODX Series Synthesizers
  3. # 1
Accepted Answer Pending Moderation
Let the algorithm as No.1 and tuning three more oscillators OP2/OP3/OP4 (all sine) to 99, the clicking noise begins when 5 notes are played simultaneously.
- Change the algorithm to No. 37 with only (!) OP1 active (sine, Volume 99), the clicking noise begins when 4 notes are played simultaneouly and strong distortion occurs for 5 notes.


All other Performance settings equal - if you used Algorithm 37 with OP1 (and only OP1) as Level 99 then this would be the same as Algorithm 1 with OP1 (and only OP1) set to level 99.

The only "special sauce" I could see possibly being employed between Alg 1 and Alg 37 is that Alg 37 only has 3 carriers while Alg 1 has 8. When you mix carriers together to create the final output - the volume of each carrier adds together. To compensate, under the hood, the instrument may trim each carrier to normalize the resulting volume. A "smart" way to do this would be to normalize according to the carrier outputs where a zero level carrier would be ignored. This way, Alg 1 and Alg 37 should be no different at all assuming using only one carrier. However, if a more quick-and-dirty method was used - then there could be fixed based on worse case. Where Alg -1 blindly trims each carrier more aggressively because there is more potential for additive effects (8 carriers total) where Alg 37 trims each carrier less because there are only 3 carriers total. This may or may not explain why Alg 37 seems to need more notes to create the distortion vs. Alg 1 configured with the same single-op configuration.

This is the sort of thing where the generic instruction to "use your ear" to adjust volumes applies. Manny would likely experimentally figure out how the carriers are mixed at a more technical/under-the-hood manner because that's his nature. And the above "special sauce" paragraph would make sense to him. But that's not really important to effectively use the system. The general notion that turning up all volume dials to "11" is not bad advice. Experience and digging in will help you tame.

That said - it's probably worthwhile for you to learn if everyone sets their volumes to "11" in the same manner - if they see the same distortion. Just to make you "feel" less alone in your journey to grasp FM-X programming. Even it's easy to explain the theory - you may want to place your feet on firm ground of understanding. Placing some stakes in the ground. I do not have a MODX - and the distortion properties of MODX vs. Montage have been reported to be slightly different. So I won't run the experiment as it's a bit Apples-to-Different-Apples. (Granny vs Macintosh?)
  1. more than a month ago
  2. MODX Series Synthesizers
  3. # 2
Bad Mister
Yamaha
Accepted Answer Pending Moderation
Please try the linked tutorials to help understand why using Operators in the manner you are suggesting will overload the system... if your goal is musical tones — what you describe is severe. I’d say the behavior you are observing is quite normal. In additive synthesis (which is exactly what you are doing with 8 Operators as Carriers) you can accurately recreate the first eight harmonics in duplicating an instrument sound.

If all you do is raise the same pitch in each of the eight Carriers you’ll have an extremely non musical situation where 8 tone generators are doing the same thing. Same pitch, same volume for each of the eight sources... not a musical situation, I’ll explain why...

Having a sine wave tone source that you can tune for each harmonic is what you want in additive synthesis... but you can never have enough and you must Program each one (naturally). Manny Fernandez, in a recent Behind the Music podcasts, talks about a conversation with Wendy Carlos, who worked extensively with additive synthesis, she mentions the good new/bad news with additive. In big systems the good news is you get a periodic wave source (sinusoidal waves) for each harmonic, the bad news is, now you have to program them...

If you study the harmonic series of any known musical instrument there is severe attenuation between the levels of the fundamental and the next overtones in the series. And as you get to the third and fourth even more reduction in level. You’d be hard pressed to find a musical sound where the successive harmonics don’t get significantly softer as you go higher up the harmonic series (or down below the Fundamental)... by the time you’ve accurately reproduced the unique envelopes and Loudness of each of the first 31 harmonics and sidebands you’ll have fairly accurately recreated the instrument timbre of the sound you’re after.

So in nature you can understand that the Fundamental is usually the loudest harmonic (f x 1), in most cases when the second harmonic occurs it is not near as loud and is a whole integer multiple of the Fundamental... However, there are instruments, like Trumpets, which don’t reproduce the Fundamental at all. A trumpet family instrument begins on the second harmonic (the Octave to the Fundamental, f x 2) this is why the first harmonic leap for a Trumpet/Bugle is a musical 5th not an octave. Its lowest tone is the second harmonic.

In each sounded musical tone, each sounded Note, is represented the whole integer multiples of the Fundamental tone calculated from the instrument’s original pitch. These upper harmonic tend to be softer as you ascend... with few exceptions. There acoustic resonances that occur within instruments that can cause a specific harmonic in the series, jump up in volume over its neighbors... but generally, the higher you go up the series the dramatically softer the output level of that upper harmonic gets.

It is the relative volumes of these harmonics that become the “fingerprint” of the sound. From a single note we can identify the instrument. Our ear/brain can decipher the complex level relationship of the prominent harmonics in a sound. We can tell a Trumpet playing the Bb below middle “C” from a Trombone playing that same pitch because the Loudness array of the harmonics clearly identifies it. The brain can store, categorize and identify, countless thousands of such harmonic level mappings. And does it extremely well even in the untrained human ear. It is primarily the Loudness of the Fundamental that has us recognize the pitch of a musical tone... it is the mathematical relationship in Loudness of the whole integer multiples of that Fundamental that determines if we hear it as a musical sound or noise.

We even recognize non-periodic patterns of frequencies on just as high a level. You can easily recognize the relative volumes of harmonic and inharmonic components of a friend’s voice versus those of a stranger (even over a phone! which always has a terrible sound system). Your ability to appreciate musical relationships is severely impacted by the narrow range of frequencies reproduced by your phone (typically 300-3K) because the relatively low numbered harmonics (midrange) are all you can hope to get, it becomes increasing difficult to identify what you’re actually hearing over a phone when it comes to music. You would start to rely on secondary behaviors in an attempt to identify the sounds... you might guess brass horn, but without the clear details of the relative volumes of the harmonics, you’ll need a characteristic like a Slide note to recognize it as Trombone, and not as a Trumpet.

FM-X can create harmonic and inharmonic sounds, musical tones and noises. Learning some of the rules to create musical tones is worth exploring. In between the whole integer relationships of the overtones, the relationships that leave fractions creates another whole area of musical sounds... often described as “bell tone” relationships. Ear-catching in their oddity — when exploring FM seek out its uncanny ability to do, accurate in every detail, bell tones... check out the mathematical relationships that characterize a sound you describe as a “bell”. Check out the tuning of Modulator to Carrier on an FM-X “bell”
  1. more than a month ago
  2. MODX Series Synthesizers
  3. # 3
Accepted Answer Pending Moderation
Dear Bad Mister / Jason,

thanks for your support and your tips!

Jason, i think it will not be necessary to upload a special performance because the effect can be checked very simply by using the "Init Normal FM-X" Performance. When i enter this performance (not altering any of the standard parameters) and simply do i.e. the following things:

- Let the algorithm as No.1 and tuning three more oscillators OP2/OP3/OP4 (all sine) to 99, the clicking noise begins when 5 notes are played simultaneously.
- Change the algorithm to No. 37 with only (!) OP1 active (sine, Volume 99), the clicking noise begins when 4 notes are played simultaneouly and strong distortion occurs for 5 notes.

I would be grateful if you could check for me if this is a normal behavior. If yes, this would mean that when using a simple sine wave with Vol. 99 for algorithms as No. 37 it is not possible to play polyphonic using more than 3 notes simultaneouly.

Thanks and regards!
  1. more than a month ago
  2. MODX Series Synthesizers
  3. # 4
Accepted Answer Pending Moderation
Maybe upload a Performance with issues to Soundmondo and leave the link here.

At least you could get feedback if others with MODX hear the same artifacts as well as tips or reason why you get these results.
  1. more than a month ago
  2. MODX Series Synthesizers
  3. # 5
Bad Mister
Yamaha
Accepted Answer Pending Moderation
It is not unusual on early attempts to program your own sounds, particularly, in the realm of FM-X... It takes a while to relax and recognize that if you max out Element Levels you can over modulate the signal. Sine waves are like a torture test because of the steadiness of the signal, none of the variance in Loudness as the harmonics change and interact, as in other more complex waves. There are no harmonics in the Sine Wave... if your programming looks like lots of Element Levels at 99, don’t be surprised at the results.

Also a key to FM programming is tapering the effectiveness of Operator interaction... the “scaling” you can have a Modulator:Carrier interaction that needs to get softer as you go up the keyboard, or vice verse... learning to scale the results is a very important part of controlling output level across the keyboard.

All 8 Operators set as Carriers turned up all the way is not a pretty sound and does not lend itself to chords or music making. (Almost as useful as all 8 Operators set as Carriers set all the way down). In between these extremes lies a literal cornucopia of sonic possibilities. Even just using this “additive” Algorithm.

I recommend studying some of the Factory FM-X programming to get an idea of how Element Levels are used, and how their scaling is shaped left to right on the keyboard... it is not a race to make each Operator at or close to maximum... remember the more a Modulator influences the Carrier the more harmonics generated... at some point all harmonics = noise (technically speaking). Recognize that you can easily over modulate things — as you get more experience you’ll find that old adage “less is more” applies to FM-X programming. Shaping the response with Envelopes is a key... something that is very chaotic sounding, bordering on noise, can be made into an articulate attack transient with the proper short Envelope. That same chaotic sound with a long Envelope will sound like some kind of distortion.

The sound of a hard mallet striking wood is a short blip, the long wooden tone of the marimba bar is quite different, if you switch the Envelopes of these two components (Operators) it becomes an entirely different instrument sound.

There are several very extensive tutorials (articles/videos) on programming FM-X from which you can learn about Operators — there are always eight available. When they are Carriers they Output audible waves, when they are Modulators they influence the Operators they are connected to...

Mastering MODX: An FM-X Exploration (Part 1 of 4)

Manny’s FM-Xploration (Article 1 of 5)
  1. more than a month ago
  2. MODX Series Synthesizers
  3. # 6
  • Page :
  • 1


There are no replies made for this post yet.
Be one of the first to reply to this post!
2021 © Yamaha Corporation of America and Yamaha Corporation. All rights reserved.