3:27 I didn't even notice that i was wrong about this algorithm saying that there was no FM going on at the 2 carriers on the sides, even though I proceeded to immediately use it for the formant example lol 😂 but that idea still holds. If there is an algorithm with carriers that have no modulator, you'll have to think about what you would use the carriers for when there's no modulation happening.
German here. It just blew my mind! All these years, I thought it was actually written as 'substractive' instead of 'subtractive.' 😄 I can't even explain why that is. Great video btw. Thank you
@@cooksoni.a That was my first guess as well, but its actually not much different: subtrahieren or Subtraktion. No "s" in there, very odd 😄. Well, I learned something today, besides FM synthesis ;)
I'm only a third in, but so far this is an outstanding down-to-earth guide to FM. Kudos on making "the second video you should watch about FM." I appreciate the balance you've set between skipping the basics and not getting bogged down in the details, accessible but practical. Focusing on two-op to start and relating the behavior to better-known analog waveforms is a great way of building intuition. Thank you!
Thank you for this video! As a new owner of a DigiTone this is incredibly useful, and as a technically oriented person this satisfied my need to understand the behind the scenes as well as get a practical methodology.
I'm the same way, I'm in the process of making another video and just yesterday I said that I don't feel satisfied until I get to the bottom of how things work lol. Some people say you don't need to know the ins and outs to use FM, but for me at least, understanding how stuff works on a more fundamental level makes it way easier to utilize. I'm glad you found it useful!
Dude! You are godsent!!! You are absolutely right that there almost no comprehensive content on deeper fm synthesis techniques and theory, and the content I managed to find on the topic is made by dinosaurs of synth world, who still put all the quirky and cheesy sound design solutions of 70-80s on pedestal.. hardly anyone shows how to use fm in modern production comprehensively, pls give us more!! Instant sub!😊 Ps and I’m a Digitone owner, and soo many things in it’s design make more sense after watching your vid 😉
15:00 On using saw waves, a lot of psytrance style patches e.g. Forest Leads use saw waves with 0.5 or 0.25 ratio to the carrier (sometimes both carrier and modulator are saws). This adds a "croak" to low notes which sounds pretty cool when put through a filter. @olliepsy has some good examples
Yeah I spoke a bit brashly when I said that lol, I'm pretty particular about what I think sounds good, but there are definitely cases where using saws can be nice. I think the important thing is that the rising edge of the saw isn't completely vertical, and that's a feature of saws with very limited harmonics. it creates a little bunching up in the first part of the waveform which can sound really cool
Extremely helpful video bud. Learning fm and particularly operator, has been a bit of a final frontier for me and you nailed it, there aren't many example and use case and explanation encompassing videos. This was excellent. Instant sub :)
just stumbled upon a model:cycles for cheap and its cool to see it here, my purchase feels validated haha :) great video btw! takes me back to messing with the ratios on fm8 a decade ago heh
Interesting stuff. Informative as hell. I'm still trying to grasp FM in SurgeXT (love the synth, but as a moron, it's a lot to learn). Keep up the good work man.
Really interesting perspective for FM! Different to other videos in YT. As I have a Volca FM that has a classic DX7 layout.. I've found in other videos that "Ratio" shows up in Relative terms, as it is the relation Between the Coarse Level (octave) between two Operators, so you would have one in 0 in relation to -1, etc. And Output Level would be "Amount". Is that correct?? In any case it's evident that an integrated oscilloscope on FM machines would have been a good way to understand more this type of synthesis... That more often has become commonly used in a "preset based" workflow. (Yes I edited this because I found this explanation heh! but any comments on this would be appreciated!)
Thanks! Yeah so I've never used a DX7, but I think it allows you to set the modulator to be either a ratio or fixed frequency, and I think the ratio setting is just "frequency" and you choose what the ratio is by setting the coarse tuning. And yes, level would be the modulation amplitude/amount. All that is is the volume of the modulating signal, with more volume generating more harmonics. The same should be true on the VolcaFM, coarse tuning for the ratio and level for the mod amplitude.
Very nice video. One thing you did not cover in this video thaty I feel is important when programming FM patches is *Velocity to level* for modulators. This makes the patch more playable and can imitate behaviour of real life physics espacially when programming Metalic sonds, bells, percussion etc.
I’m going to have to watch this again when I’m ready for the difficulties that caught me by surprise the first time. I’m currently in the modular hardware world so some of your major concepts, such as ratios, aren’t facilitated by the equipment I have on hand. Or something.
@@tracyharms3548 so ratios are just multiples of a given frequency. For instance, 220hz is an A note. If you multiply that by two, it gives you 440hz. By 3 its 660, 4x is 880, etc. This idea of multiplying a given frequency by an integer reflects whats known as the harmonic series, and thats using those ratios in FM gives you a harmonic sound. Look up a video on the harmonic series on youtube and it might give you a better idea, its really interesting. But put simply, the concept of the ratio of one operator to the other is just what pitch each of the operators are at
@@cooksoni.a ”When your modulator is an integer multiple of the fundamental frequency then all of your sidebands will be in the harmonic series.” Does ‘fundamental frequency’ here mean carrier frequency?
@ it technically means the midi note that youre playing it at, because you can also change the carrier’s frequency by adjusting its ratio. But in most cases youll be using a carrier with a ratio of 1, so in that case yes. Because 1 times n frequency equals n, meaning that its the fundamental freq that is equivalent to whatever note youre playing it at
@ I imagined ‘ratio’ means the ratio of the frequencies of the two operators. If the carrier can meaningfully have a ratio that’s independent of the modulator’s frequency, I am lost.
15:04 - I think I remember watching an old SeamlessR video on RUclips where he briefly mentioned and demonstrated some details having to do with Phase vs Frequency modulation and saw waves - and it sounded interesting but I wasn't really grasping it right then and there. Would definitely be interested in you doing a video about that topic, especially with how helpful this video has been.
Interesting, I'll see if i can find that. Yeah I spoke a little brashly when I said that lol, because in most cases if there's a saw wave available in an FM synth, then the developers took the difference into consideration. But there are some synths that have FM capabilities that don't use a saw wave tailored for FM. That little tangent was mainly directed at people who want to try making an FM synth themselves. And I've seen people demonstrate stuff like using a real saw wave in video explanations about FM, along with other bits of information that isn't relevant to phase modulation, so yes FM vs PM video is forthcoming!
what is your opinion on sytrus as a synth? can it do everything that fm8 can do? I tried using the feedback thing in sytrus (by self modulating an oscillator with itself) and wasnt able to get the noise sound you were able to with fm8, I am assuming that the synth programmer most likely decided the noise was undesirable and then limited the maximum possible amount of feedback. Would you say that when making patches, the noise part of the feedback is useful at all? Thank you so much for making this video, I might pick up fm8 when it is on sale.
@@slowslicing I’ve never used sytrus, but that may be true that the developer didn’t give the FB the gain necessary to make noise. Depending on what DAW you use, there may be a stock FM synth that you could do it with, but my original DAW is Logic and you couldn’t do it there. I really love the noise that comes from high feedback, i think it can sound really pretty and of course you can use it to modulate in specific ways depending on how you use the envelopes. But keep an eye out for FM8 sales, i got it for $20 earlier this summer
Great information and great advice… Besides Gem I would also recommend Glanta by Fors and Vector Fm, which is a weird and fun approach to sound generation with FM Pop and lux are very interesting as well, both bye Fors, But Lux is much more of an additive approach, very fun nonetheless Once again, thanks and congratulations… already subscribed and waiting for more❤
Thanks! Yeah pop and lux are awesome. Lux is actually completely additive and not FM. But I love putting it in a drum rack because it has that pitch envelope built in. After I finish these other FM videos I have in the works, I really wanna do a video on additive synthesis, because it is HIGHLY underrated
@@cooksoni.a Please do the additive stuff! I had to move from LPX to Live and I miss Alchemy so much... Recently I read something about one OctoCell that looks promissing, but I couldn't find much information about it. Do you know it?
Also I'm with you on being a little sad seeing the Digitone II doesn't have the same accent of color that the OG Digitone has with the track buttons, screen, and white keys. I literally just bought one a couple weeks ago and was surprised by the II announcement, but it stung a little bit less bc of the fact that I think the OG still looks cooler lol.
Am I correct in guessing the presumption is that carrier and modulator will have their base frequencies changed in parallel so the ratio between their frequencies remains constant? (E.g. carrier frequency changed for a melodic sequence, modulator frequency changing in parallel to maintain the selected ratio.)
@@tracyharms3548 yes thats right. And about your other comment, it’s not a ratio between the two operators, its a ratio to the midi note. So if you tell it to play an A, that corresponds to 220hz for instance. So the ratios of the operators are ratios to that frequency, 220hz. A ratio of 1 is 220hz, 2 is 440hz, etc. That relationship is the same no matter what midi note you tell it to play. So if you have a sequence going, the ratios of the operators are always going to be in a relationship of the note youre playing times 2 or 3 or 4, etc
@ oh ok, well its the same with using like a modular sequencer that uses voltage to control the pitch. It’s just whatever pitch you tell the fm synth to play at, the ratios are going to be integer multiples of that pitch
That's true, it's just not a very impressive drum machine to me honestly. I like the tone of the operators on those soundblaster chips, but I never really liked the way they were configured in the built in drum mode. There's not a whole lot of fm going on in it, which confused me bc I thought the chip architecture allowed for something more complex, but I don't fully understand it. In my opinion it doesn't end up sounding very interesting
1:1 sounds and looks suspiciously like wave folding. This is further support for thinking fm is more like "multiplicative synthesis", rather than additive or subtractive. Definitions of domains seem fuzzy in my mind. Time vs frequency? Frequency can't exist without time, right?
I've noticed that too, and actually something very strange is that setting the carrier to a ratio of 0 and then increasing modulation actually just gives you straight up wavefolding. You can't usually do that in fm synths, but you can in Glänta. I'm still trying to work out why that happens lol. As for the time vs freq domain, they're just the two different starting points for synth algorithms. When you start with an oscillating waveform and alter its phase, that generates harmonics. But when you start with the harmonics themselves and move them around and change their amplitudes, then the culmination of those harmonics are what generate the waveform. In reality both the frequency and time domain are two dimensions mathematically representing the exact same thing (sound), but when you're making a synth algorithm you choose which dimension you're going to work in. Working in the frequency domain is by definition additive synthesis, and modal synthesis (often used in physical modeling) is probably the most common implementation of that.
@@benedictjohnson2176 well thank you but i wouldnt fancy myself a genius lol 😂 i cant recommend it currently because i havent used it, but i did order one. Maybe ill do a video on it when it gets here.
"I really like a ratio of 1:2" My friend, I hope you like EBM because that genre is pretty much entirely derived from grumpy musicians using 1:2 ratios
I think you’re mistaking how ableton operator’s coarse pitch adjustment corresponds to the modulator to carrier ratio. Each adjustment to coarse pitch is one semitone off the midi note played. So +12 coarse on the modulator would give a 2:1 ratio to an unmodified carrier because one whole octave is doubling the frequency. In this video you say coarse set to two means a ratio of two. That is also why you were confused about adjustment up to 41, which is still just a ratio around 3.5:1
I am new to operator/Live at large, but I don't think you're right. Ableton says this on their instrument reference page: Osc Coarse Frequency (Coarse) - The relationship between oscillator frequency and note pitch is defined by the Coarse and Fine parameters. Coarse sets the ratio in whole numbers, creating a harmonic relationship. Maybe you're thinking of fine? that would make more sense for semitones
Thanks! It's Ess Mattisson. He used to work at Elektron and had a big part in designing the Digitone and Model:Cycles. Now he's making plugins and M4L devices at Fors, their website is fors(dot)fm (youtube is weird about links in comments)
Oh you know what, I forgot that Operator doesn't have the correct phase, it's a weird quirk. And I think that the feedback paths are not independent from the level of operator you're feeding back, so you won't be able to get the feedback level high enough before using it to modulate another operator. So unfortunately it seems like you can't do this trick in Operator, sorry about that.
Great video. I own the FM8 but have used it in exactly one song. I was surprised how well it does pads. Its not something i expected from an FM synth. But normally if i am going to do any audio rate modulation its in phase plant. SO my question is (and i cant find a clear answer) is what kind of FM synthesis does FM8 and others like DX7 use. is it linear FM? or is it actually Phase modulation ? my guess is linear FM where we are actually modulating the frequency and not the phase.
No, FM8 uses linear phase modulation, and that's actually what every single FM synth uses. FM is a misnomer, just because that's what Yamaha patented it as in the 70s. But actually, there's no such thing as a true FM synth, it's all phase modulation. There are a couple reasons for that, the main one being that phase modulation much more computationally efficient than FM, and it allows you to do feedback which you can't do in FM (or at least it won't behave the way that it does in phase modulation). Linear FM is usually used to describe voltage controlled oscillators, as a opposed to exponential and thru-zero FM. Linear FM is most similar to phase modulation, but you need a ton of gain on the modulator to get the classic FM sounds. Thru-zero FM is usually a better choice for doing analog FM sounds. So yeah, it's like that meme of the astronauts and he goes "it's all just phase modulation?" and the other one says "it always has been" and kills him lol
@@cooksoni.a ok thanks for clearning that up. I watched a video a long time ago when i was learning phase plant, where someone said "phase plant can do true FM synthesis not just phase modulation" which has probably caused a lot of wasted hours trying to get FM sounds from phase plant by modulating the frequency instead of the phase. Appreciate the video, it visually laid things out better than I have seen before. I spent the last hour or so with FM8 and phase plant open just trying to replicate sounds from FM8 in phase plant so i could have a better understanding of how the ratios and all that stuff works. So thanks again. looking forward to whatever comes next
@@PeteGunnShow Okay that's interesting. Well whenever you can modulate the frequency of an oscillator at audio rate then you de facto can do FM, but there are a number of caveats there because of the different ways of scaling frequency, like I mentioned. And the difference of course becomes apparent when you use waveforms with straight angles. But when something is marketed as an FM synth, it's pretty much guaranteed to be phase modulation. Unless somebody makes something in Max or something, like I've seen an FM synth Max tutorial before and he just used FM instead of PM. But thanks for the compliment. I'll try and get my ass on that
@cooksoni.a Rad. Thanks for the explanations and insights. I think I have 3 different styles in my rack. Captain big O vco has inputs for exp and Lin fm, and the happy nerding fm aid does phase modulation. It actually expects a sawtooth as the input, LoL. Please do a vid on that. BRB, subscribing.
DX7 uses what's now known as "Chowning-style" FM (due to its creator and relatively simple signal path). Also, Linear & Exponential FM are simply different ways the carrier osc processes the modulator onto its frequency: one example of how this differs is that Linear maintains the carrier osc's pitch and only affects timbre/harmonics/etc., whereas with Expo. FM the modulator(s) will change the _overall pitch_ of the carrier/voice as its frequency changes (causing pitch envelopes or vibrato effects). Both linear & exponential FM are applicable to the analog & digital realms, and very often FM synths are capable of both. This is NOT specific to hardware or "Voltage Controlled Oscillators". Now, Through-Zero FM is its own separate thing, and was originally only possible in the digital realm (and only really capable in analog domain as a hybrid circuit using clever switching circuits combined with a bipolar VCA, and it still has some differences & compromises). And just to add some further clarification on some of OP's opinions; FM & PM are distinct forms of synthesis with clear definitions - different from eachother but with some overlap, and capable of similar sounds. This is shown by the fact that you can modulate the phase-position of a wave cycle *or* its overall frequency in the same way, and end up with different results (you can also modulate both separately). FM does NOT automatically equal Phase Modulation in the technical sense, as OP states! Look into some proper PM synths (like retro Casio) to see some differences. Furthermore, Feedback is 100% possible in FM and is a pretty standard technique these days. I don't understand where OP cot this idea, as original feedback modulation simply meant a modulator osc's output being "fed back" into its own frequency (usually only by a partial cycle, I believe?) *before* affecting the carrier osc. But there are also different types & applications of FM Feedback. I believe OP's heart is in the right place, and the passion seems to be there, but there are still clear lapses in his overall understanding of not just FM synthesis, but some wider concepts of synthesis as a whole (and some of its terminology). This has led to them making some inaccurate information in the video and comment replies, likely due to innocent mixups and/or not having the "full picture" for some of these technologies. So I just wanted to clear some of it up - no hate here!
I don't have as much experience with physical modeling as I do with FM, but I really love the sound of physical modeling synths. I'm planning on doing a video on modal synthesis (which is used in a lot of physical modeling algorithms) because it just sounds so cool. I don't know how much advice I could offer in sound designing with it, but there are so many directions you could go with it, and a couple of misconceptions that I see about it online, so I definitely want to do a video on that
@ Yes!!! Modal synthesis is my favorite part of physical modeling synthesis, and getting a deep dive into that in the same way you dove into 2 operator fm to just really have the fundamentals down would be so appreciated!!
@ What modal synth do you like using/ would be using for the video? I currently use Mass in Opal by Fors when I want something simple, or Objekt when i want something more complex, but i’d love to know about more!!
@@rainbowkittycat627 Have you heard of Fors other modal plugin called Tela? It's absolutely amazing. I personally think it's the best sounding modal synth out there, I think it has to do with the internal gain staging in the code because it just sounds so much more clean than any other one I've messed around with. Other than that, Ableton's native Collision is also a modal synth that sounds nice, Intellijel came out with a eurorack module that uses a very similar synth engine as Collision (they worked with the co-developers who worked on Collision), I don't know if it's just placebo but I think the eurorack module sounds better to my ears. Mutable's Rings is another classic one that I'm sure you can find in software form too, I don't think it sounds as good as the others I mentioned but you can use it as an actual resonator which is nice.
@@bubbleandsquawk I think actually it’s closer to physical modeling, the waveforms are just sine, saw, and noise, and those get sent to a “waveguide” which i think can be either a positive or negative comb filter. It sounds awesome tho, that one is also pretty one of a kind. If i get better at programming i really want to make a drum plugin reminiscent of that
@@cooksoni.a I’ve got one! I suppose it isn’t actually an FM synth per se now that I think about it, but it is capable of some very basic 2-op fm sounds via the sine modulator. There is no feedback and the modulator doesn’t have an envelope of its own but you can kind of work around this with motion recording. Also, each voice is bi-timbral and as you said there are noise waveforms in there so it’s pretty flexible. The waveguide is cool but you only get one of them and a send for each voice. If I use it at all it’s typically on one voice only.
I just have a fascination with abacuses lol, and I found that huge one for $20 at a thrift store in college so i had to get it 😂 it's probably a 70 year old teaching aid
What software is this? Nobody is asking in the comments so the question is probably stupid, ... but I couldn't find you mentioning it. Am I going crazy??
You're good lol, I mentioned it at the beginning of the sections but I was using Max for the first demo, then the plugin FM8 by Native Instruments, then Ableton Live's Operator instrument.
@@cooksoni.a Thx a lot! "Max" didn't mean anything to me. I hav been around the live sound synthesis world quite a while now and somehow it never crossed my paths. I noticed the word "max" in what you said, but thought it hopeless to google for "max" ^^. Thx for clearing things up for me! Have a good one!
@@ajmgdaj yeah no prob! And yeah Max is a visual programming language/environment by the company Cycling '74, and they partnered up with Ableton so you can make Live-specific plugins called Max for Live Devices. There are a handful of official Ableton plugins that come with it which were made with Max, I'm not sure if it's like that for the standard version but if you have Live Suite then you're automatically able to load MFL devices , edit them, and make your own. Interestingly actually, Live was supposedly prototyped in the Max language back in the 90s lol
@@SeattleMarc thank you. But I think I did, I tried to demonstrate that understanding FM boils down to understanding the timbral variations you get with only 3 parameters in a 2 operator configuration. I don’t think that’s any more difficult than understanding subtractive synthesis.
What business have ratio and amplitude to fm. Spread out harmonics, add harmonics? Wth that supposed to mean? Why spreading out harmonics related to frequency modulation, that two different things, how aplitude related to harmonics..that messy
3:27 I didn't even notice that i was wrong about this algorithm saying that there was no FM going on at the 2 carriers on the sides, even though I proceeded to immediately use it for the formant example lol 😂 but that idea still holds. If there is an algorithm with carriers that have no modulator, you'll have to think about what you would use the carriers for when there's no modulation happening.
beyond how excellent this tutorial has been for me so far, you seem like a really kind person. just something about your energy
Thats so sweet lol, thank you ❤
German here. It just blew my mind! All these years, I thought it was actually written as 'substractive' instead of 'subtractive.' 😄 I can't even explain why that is. Great video btw. Thank you
@@janbert4927 haha thats awesome. I figured the german word for it might be closer to “substractive,” is that the case?
@@cooksoni.a That was my first guess as well, but its actually not much different: subtrahieren or Subtraktion. No "s" in there, very odd 😄. Well, I learned something today, besides FM synthesis ;)
@@janbert4927 Very interesting lol 😂 I wonder how that got started
I'm french and I realized I did the same error :D. But it actually makes sense as the french word is 'soustractive'.
Or maybe we confuse it with "subscription"
I'm only a third in, but so far this is an outstanding down-to-earth guide to FM. Kudos on making "the second video you should watch about FM." I appreciate the balance you've set between skipping the basics and not getting bogged down in the details, accessible but practical. Focusing on two-op to start and relating the behavior to better-known analog waveforms is a great way of building intuition. Thank you!
This is the stuff Bitwig is so good at. Thanks for inspiring me to sound design some new stuff :D
Yay for Bitwig
was going to say: Bitwig would be an awesome tool to demonstrate these concepts.
Buy Opsix native and this video is really good for learn the basics of every fm synth and explore different possibilities, really nice 👍
Many thanks, very helpful for Digitone users.
Yes please to a video on frequency vs phase modulation!
Thank you for this video! As a new owner of a DigiTone this is incredibly useful, and as a technically oriented person this satisfied my need to understand the behind the scenes as well as get a practical methodology.
I'm the same way, I'm in the process of making another video and just yesterday I said that I don't feel satisfied until I get to the bottom of how things work lol. Some people say you don't need to know the ins and outs to use FM, but for me at least, understanding how stuff works on a more fundamental level makes it way easier to utilize. I'm glad you found it useful!
Dude! You are godsent!!! You are absolutely right that there almost no comprehensive content on deeper fm synthesis techniques and theory, and the content I managed to find on the topic is made by dinosaurs of synth world, who still put all the quirky and cheesy sound design solutions of 70-80s on pedestal.. hardly anyone shows how to use fm in modern production comprehensively, pls give us more!! Instant sub!😊
Ps and I’m a Digitone owner, and soo many things in it’s design make more sense after watching your vid 😉
Easy to follow and practical, thanks !
I finally understand what each algorithm goes and how to use them !
Thanks for this intermediate overview!
I really needed a guide on how to utilize the algorithms more deliberately than randomly.
Cheers, dude 🎉
Love the format of this. It's not easy to teach intuition, but you've done a great job. Subscribed :-)
Your channel should grow fast! Excellent stuff.
Dude, thank you! Just what I needed to still chern some love out of my digitone! Very cool content! Keep it up!
Great video - from a big 2-op FM fan :) subscribed
Fantastic video! Thank you!!!
15:00 On using saw waves, a lot of psytrance style patches e.g. Forest Leads use saw waves with 0.5 or 0.25 ratio to the carrier (sometimes both carrier and modulator are saws). This adds a "croak" to low notes which sounds pretty cool when put through a filter. @olliepsy has some good examples
Yeah I spoke a bit brashly when I said that lol, I'm pretty particular about what I think sounds good, but there are definitely cases where using saws can be nice. I think the important thing is that the rising edge of the saw isn't completely vertical, and that's a feature of saws with very limited harmonics. it creates a little bunching up in the first part of the waveform which can sound really cool
great video, learned a lot!
Extremely helpful video bud. Learning fm and particularly operator, has been a bit of a final frontier for me and you nailed it, there aren't many example and use case and explanation encompassing videos. This was excellent. Instant sub :)
Phenomenal explanation brotha
Great FM teaching!
Thanks from Berlin 👋🏻
About damn time !
Quite educative! ❤ Loved the pace
just stumbled upon a model:cycles for cheap and its cool to see it here, my purchase feels validated haha :) great video btw! takes me back to messing with the ratios on fm8 a decade ago heh
really well explained, loved it !!
Excellent work. Thanks for the effort you put into it. 🙂
Cool, a video on FM synthesis !
Interesting stuff. Informative as hell. I'm still trying to grasp FM in SurgeXT (love the synth, but as a moron, it's a lot to learn). Keep up the good work man.
Thank you, great video!
Nice explanation!
Thanks for making this.
Really interesting perspective for FM! Different to other videos in YT. As I have a Volca FM that has a classic DX7 layout.. I've found in other videos that "Ratio" shows up in Relative terms, as it is the relation Between the Coarse Level (octave) between two Operators, so you would have one in 0 in relation to -1, etc. And Output Level would be "Amount". Is that correct?? In any case it's evident that an integrated oscilloscope on FM machines would have been a good way to understand more this type of synthesis... That more often has become commonly used in a "preset based" workflow. (Yes I edited this because I found this explanation heh! but any comments on this would be appreciated!)
Thanks! Yeah so I've never used a DX7, but I think it allows you to set the modulator to be either a ratio or fixed frequency, and I think the ratio setting is just "frequency" and you choose what the ratio is by setting the coarse tuning. And yes, level would be the modulation amplitude/amount. All that is is the volume of the modulating signal, with more volume generating more harmonics. The same should be true on the VolcaFM, coarse tuning for the ratio and level for the mod amplitude.
instantly subbed
Very nice video. One thing you did not cover in this video thaty I feel is important when programming FM patches is *Velocity to level* for modulators. This makes the patch more playable and can imitate behaviour of real life physics espacially when programming Metalic sonds, bells, percussion etc.
I’m going to have to watch this again when I’m ready for the difficulties that caught me by surprise the first time. I’m currently in the modular hardware world so some of your major concepts, such as ratios, aren’t facilitated by the equipment I have on hand. Or something.
@@tracyharms3548 so ratios are just multiples of a given frequency. For instance, 220hz is an A note. If you multiply that by two, it gives you 440hz. By 3 its 660, 4x is 880, etc. This idea of multiplying a given frequency by an integer reflects whats known as the harmonic series, and thats using those ratios in FM gives you a harmonic sound. Look up a video on the harmonic series on youtube and it might give you a better idea, its really interesting. But put simply, the concept of the ratio of one operator to the other is just what pitch each of the operators are at
@@cooksoni.a ”When your modulator is an integer multiple of the fundamental frequency then all of your sidebands will be in the harmonic series.” Does ‘fundamental frequency’ here mean carrier frequency?
@ it technically means the midi note that youre playing it at, because you can also change the carrier’s frequency by adjusting its ratio. But in most cases youll be using a carrier with a ratio of 1, so in that case yes. Because 1 times n frequency equals n, meaning that its the fundamental freq that is equivalent to whatever note youre playing it at
@ I imagined ‘ratio’ means the ratio of the frequencies of the two operators. If the carrier can meaningfully have a ratio that’s independent of the modulator’s frequency, I am lost.
great video!
great video. really great. lets stay in touch.
I like the modulator at a quarter carrier freq, then run through a folder. Pipe that through a lpg, and we're in contrabassoon territory.
15:04 - I think I remember watching an old SeamlessR video on RUclips where he briefly mentioned and demonstrated some details having to do with Phase vs Frequency modulation and saw waves - and it sounded interesting but I wasn't really grasping it right then and there. Would definitely be interested in you doing a video about that topic, especially with how helpful this video has been.
Interesting, I'll see if i can find that. Yeah I spoke a little brashly when I said that lol, because in most cases if there's a saw wave available in an FM synth, then the developers took the difference into consideration. But there are some synths that have FM capabilities that don't use a saw wave tailored for FM. That little tangent was mainly directed at people who want to try making an FM synth themselves. And I've seen people demonstrate stuff like using a real saw wave in video explanations about FM, along with other bits of information that isn't relevant to phase modulation, so yes FM vs PM video is forthcoming!
I also have memories of watching and not understanding him talk about that lol
what is your opinion on sytrus as a synth? can it do everything that fm8 can do? I tried using the feedback thing in sytrus (by self modulating an oscillator with itself) and wasnt able to get the noise sound you were able to with fm8, I am assuming that the synth programmer most likely decided the noise was undesirable and then limited the maximum possible amount of feedback. Would you say that when making patches, the noise part of the feedback is useful at all?
Thank you so much for making this video, I might pick up fm8 when it is on sale.
@@slowslicing I’ve never used sytrus, but that may be true that the developer didn’t give the FB the gain necessary to make noise. Depending on what DAW you use, there may be a stock FM synth that you could do it with, but my original DAW is Logic and you couldn’t do it there. I really love the noise that comes from high feedback, i think it can sound really pretty and of course you can use it to modulate in specific ways depending on how you use the envelopes. But keep an eye out for FM8 sales, i got it for $20 earlier this summer
thank you!
ive been using Helm for like, years now, and its mod knob uses FM, and i think it's the easiest way I've understood it
Great information and great advice…
Besides Gem I would also recommend Glanta by Fors and Vector Fm, which is a weird and fun approach to sound generation with FM
Pop and lux are very interesting as well, both bye Fors, But Lux is much more of an additive approach, very fun nonetheless
Once again, thanks and congratulations…
already subscribed and waiting for more❤
Thanks! Yeah pop and lux are awesome. Lux is actually completely additive and not FM. But I love putting it in a drum rack because it has that pitch envelope built in.
After I finish these other FM videos I have in the works, I really wanna do a video on additive synthesis, because it is HIGHLY underrated
@@cooksoni.a
Please do the additive stuff!
I had to move from LPX to Live and I miss Alchemy so much...
Recently I read something about one OctoCell that looks promissing, but I couldn't find much information about it. Do you know it?
Also I'm with you on being a little sad seeing the Digitone II doesn't have the same accent of color that the OG Digitone has with the track buttons, screen, and white keys. I literally just bought one a couple weeks ago and was surprised by the II announcement, but it stung a little bit less bc of the fact that I think the OG still looks cooler lol.
Am I correct in guessing the presumption is that carrier and modulator will have their base frequencies changed in parallel so the ratio between their frequencies remains constant? (E.g. carrier frequency changed for a melodic sequence, modulator frequency changing in parallel to maintain the selected ratio.)
@@tracyharms3548 yes thats right. And about your other comment, it’s not a ratio between the two operators, its a ratio to the midi note. So if you tell it to play an A, that corresponds to 220hz for instance. So the ratios of the operators are ratios to that frequency, 220hz. A ratio of 1 is 220hz, 2 is 440hz, etc. That relationship is the same no matter what midi note you tell it to play. So if you have a sequence going, the ratios of the operators are always going to be in a relationship of the note youre playing times 2 or 3 or 4, etc
@ Since I am approaching your tutorial with no involvement of MIDI, this accounts for some of my confusion. Thanks for your instruction, regardless.
@ oh ok, well its the same with using like a modular sequencer that uses voltage to control the pitch. It’s just whatever pitch you tell the fm synth to play at, the ratios are going to be integer multiples of that pitch
Model:Cycles isn't the only one Blast Beats by Twisted Electrons is another fm drum machine
That's true, it's just not a very impressive drum machine to me honestly. I like the tone of the operators on those soundblaster chips, but I never really liked the way they were configured in the built in drum mode. There's not a whole lot of fm going on in it, which confused me bc I thought the chip architecture allowed for something more complex, but I don't fully understand it. In my opinion it doesn't end up sounding very interesting
1:1 sounds and looks suspiciously like wave folding. This is further support for thinking fm is more like "multiplicative synthesis", rather than additive or subtractive.
Definitions of domains seem fuzzy in my mind. Time vs frequency? Frequency can't exist without time, right?
I've noticed that too, and actually something very strange is that setting the carrier to a ratio of 0 and then increasing modulation actually just gives you straight up wavefolding. You can't usually do that in fm synths, but you can in Glänta. I'm still trying to work out why that happens lol.
As for the time vs freq domain, they're just the two different starting points for synth algorithms. When you start with an oscillating waveform and alter its phase, that generates harmonics. But when you start with the harmonics themselves and move them around and change their amplitudes, then the culmination of those harmonics are what generate the waveform. In reality both the frequency and time domain are two dimensions mathematically representing the exact same thing (sound), but when you're making a synth algorithm you choose which dimension you're going to work in. Working in the frequency domain is by definition additive synthesis, and modal synthesis (often used in physical modeling) is probably the most common implementation of that.
As an FM genius, do you recommend the new Digitone 2 for FM sounds? Thank you
@@benedictjohnson2176 well thank you but i wouldnt fancy myself a genius lol 😂 i cant recommend it currently because i havent used it, but i did order one. Maybe ill do a video on it when it gets here.
"I really like a ratio of 1:2" My friend, I hope you like EBM because that genre is pretty much entirely derived from grumpy musicians using 1:2 ratios
I think you’re mistaking how ableton operator’s coarse pitch adjustment corresponds to the modulator to carrier ratio. Each adjustment to coarse pitch is one semitone off the midi note played. So +12 coarse on the modulator would give a 2:1 ratio to an unmodified carrier because one whole octave is doubling the frequency. In this video you say coarse set to two means a ratio of two. That is also why you were confused about adjustment up to 41, which is still just a ratio around 3.5:1
I am new to operator/Live at large, but I don't think you're right. Ableton says this on their instrument reference page: Osc Coarse Frequency (Coarse) - The relationship between oscillator frequency and note pitch is defined by the Coarse and Fine parameters. Coarse sets the ratio in whole numbers, creating a harmonic relationship.
Maybe you're thinking of fine? that would make more sense for semitones
is there a digitone ii somewhere in this video, or is it just in the main picture for the video.
No, I capitalized on it for views. 🫶
Cool video. Who is S Madison?
Thanks! It's Ess Mattisson. He used to work at Elektron and had a big part in designing the Digitone and Model:Cycles. Now he's making plugins and M4L devices at Fors, their website is fors(dot)fm (youtube is weird about links in comments)
I used feedback on Ableton Operator but it can't create sawtooth wave
@@longgianghai5420 what king of setup are you using? Lemme know what the settings and algorithm are
Oh you know what, I forgot that Operator doesn't have the correct phase, it's a weird quirk. And I think that the feedback paths are not independent from the level of operator you're feeding back, so you won't be able to get the feedback level high enough before using it to modulate another operator. So unfortunately it seems like you can't do this trick in Operator, sorry about that.
Great video. I own the FM8 but have used it in exactly one song. I was surprised how well it does pads. Its not something i expected from an FM synth. But normally if i am going to do any audio rate modulation its in phase plant. SO my question is (and i cant find a clear answer) is what kind of FM synthesis does FM8 and others like DX7 use. is it linear FM? or is it actually Phase modulation ? my guess is linear FM where we are actually modulating the frequency and not the phase.
No, FM8 uses linear phase modulation, and that's actually what every single FM synth uses. FM is a misnomer, just because that's what Yamaha patented it as in the 70s. But actually, there's no such thing as a true FM synth, it's all phase modulation. There are a couple reasons for that, the main one being that phase modulation much more computationally efficient than FM, and it allows you to do feedback which you can't do in FM (or at least it won't behave the way that it does in phase modulation).
Linear FM is usually used to describe voltage controlled oscillators, as a opposed to exponential and thru-zero FM. Linear FM is most similar to phase modulation, but you need a ton of gain on the modulator to get the classic FM sounds. Thru-zero FM is usually a better choice for doing analog FM sounds.
So yeah, it's like that meme of the astronauts and he goes "it's all just phase modulation?" and the other one says "it always has been" and kills him lol
@@cooksoni.a ok thanks for clearning that up. I watched a video a long time ago when i was learning phase plant, where someone said "phase plant can do true FM synthesis not just phase modulation" which has probably caused a lot of wasted hours trying to get FM sounds from phase plant by modulating the frequency instead of the phase.
Appreciate the video, it visually laid things out better than I have seen before. I spent the last hour or so with FM8 and phase plant open just trying to replicate sounds from FM8 in phase plant so i could have a better understanding of how the ratios and all that stuff works.
So thanks again. looking forward to whatever comes next
@@PeteGunnShow Okay that's interesting. Well whenever you can modulate the frequency of an oscillator at audio rate then you de facto can do FM, but there are a number of caveats there because of the different ways of scaling frequency, like I mentioned. And the difference of course becomes apparent when you use waveforms with straight angles. But when something is marketed as an FM synth, it's pretty much guaranteed to be phase modulation. Unless somebody makes something in Max or something, like I've seen an FM synth Max tutorial before and he just used FM instead of PM.
But thanks for the compliment. I'll try and get my ass on that
@cooksoni.a Rad. Thanks for the explanations and insights. I think I have 3 different styles in my rack. Captain big O vco has inputs for exp and Lin fm, and the happy nerding fm aid does phase modulation. It actually expects a sawtooth as the input, LoL. Please do a vid on that. BRB, subscribing.
DX7 uses what's now known as "Chowning-style" FM (due to its creator and relatively simple signal path). Also, Linear & Exponential FM are simply different ways the carrier osc processes the modulator onto its frequency: one example of how this differs is that Linear maintains the carrier osc's pitch and only affects timbre/harmonics/etc., whereas with Expo. FM the modulator(s) will change the _overall pitch_ of the carrier/voice as its frequency changes (causing pitch envelopes or vibrato effects).
Both linear & exponential FM are applicable to the analog & digital realms, and very often FM synths are capable of both. This is NOT specific to hardware or "Voltage Controlled Oscillators". Now, Through-Zero FM is its own separate thing, and was originally only possible in the digital realm (and only really capable in analog domain as a hybrid circuit using clever switching circuits combined with a bipolar VCA, and it still has some differences & compromises).
And just to add some further clarification on some of OP's opinions; FM & PM are distinct forms of synthesis with clear definitions - different from eachother but with some overlap, and capable of similar sounds. This is shown by the fact that you can modulate the phase-position of a wave cycle *or* its overall frequency in the same way, and end up with different results (you can also modulate both separately). FM does NOT automatically equal Phase Modulation in the technical sense, as OP states! Look into some proper PM synths (like retro Casio) to see some differences.
Furthermore, Feedback is 100% possible in FM and is a pretty standard technique these days. I don't understand where OP cot this idea, as original feedback modulation simply meant a modulator osc's output being "fed back" into its own frequency (usually only by a partial cycle, I believe?) *before* affecting the carrier osc. But there are also different types & applications of FM Feedback.
I believe OP's heart is in the right place, and the passion seems to be there, but there are still clear lapses in his overall understanding of not just FM synthesis, but some wider concepts of synthesis as a whole (and some of its terminology). This has led to them making some inaccurate information in the video and comment replies, likely due to innocent mixups and/or not having the "full picture" for some of these technologies. So I just wanted to clear some of it up - no hate here!
Do this with physical modeling synthesis please!! (My fav synth is Objekt)
I don't have as much experience with physical modeling as I do with FM, but I really love the sound of physical modeling synths. I'm planning on doing a video on modal synthesis (which is used in a lot of physical modeling algorithms) because it just sounds so cool. I don't know how much advice I could offer in sound designing with it, but there are so many directions you could go with it, and a couple of misconceptions that I see about it online, so I definitely want to do a video on that
@ Yes!!! Modal synthesis is my favorite part of physical modeling synthesis, and getting a deep dive into that in the same way you dove into 2 operator fm to just really have the fundamentals down would be so appreciated!!
@ What modal synth do you like using/ would be using for the video? I currently use Mass in Opal by Fors when I want something simple, or Objekt when i want something more complex, but i’d love to know about more!!
@@rainbowkittycat627 Have you heard of Fors other modal plugin called Tela? It's absolutely amazing. I personally think it's the best sounding modal synth out there, I think it has to do with the internal gain staging in the code because it just sounds so much more clean than any other one I've messed around with. Other than that, Ableton's native Collision is also a modal synth that sounds nice, Intellijel came out with a eurorack module that uses a very similar synth engine as Collision (they worked with the co-developers who worked on Collision), I don't know if it's just placebo but I think the eurorack module sounds better to my ears. Mutable's Rings is another classic one that I'm sure you can find in software form too, I don't think it sounds as good as the others I mentioned but you can use it as an actual resonator which is nice.
KORG volca drum is also an FM drum machine!
@@bubbleandsquawk I think actually it’s closer to physical modeling, the waveforms are just sine, saw, and noise, and those get sent to a “waveguide” which i think can be either a positive or negative comb filter. It sounds awesome tho, that one is also pretty one of a kind. If i get better at programming i really want to make a drum plugin reminiscent of that
@@cooksoni.a I’ve got one! I suppose it isn’t actually an FM synth per se now that I think about it, but it is capable of some very basic 2-op fm sounds via the sine modulator. There is no feedback and the modulator doesn’t have an envelope of its own but you can kind of work around this with motion recording. Also, each voice is bi-timbral and as you said there are noise waveforms in there so it’s pretty flexible.
The waveguide is cool but you only get one of them and a send for each voice. If I use it at all it’s typically on one voice only.
Very interesting toys you have there! What's the abacus for? lol
I just have a fascination with abacuses lol, and I found that huge one for $20 at a thrift store in college so i had to get it 😂 it's probably a 70 year old teaching aid
@@cooksoni.a I was thinking you were maybe organizing drum loops on it or something. It is really cool though! I'd love to have one cause, aesthetic!
What software is this? Nobody is asking in the comments so the question is probably stupid, ... but I couldn't find you mentioning it. Am I going crazy??
You're good lol, I mentioned it at the beginning of the sections but I was using Max for the first demo, then the plugin FM8 by Native Instruments, then Ableton Live's Operator instrument.
@@cooksoni.a Thx a lot! "Max" didn't mean anything to me. I hav been around the live sound synthesis world quite a while now and somehow it never crossed my paths. I noticed the word "max" in what you said, but thought it hopeless to google for "max" ^^. Thx for clearing things up for me!
Have a good one!
For ppl like me: The software is called Max/MSP by a company that seems to be associated with / a grooup from Ableton
@@ajmgdaj yeah no prob! And yeah Max is a visual programming language/environment by the company Cycling '74, and they partnered up with Ableton so you can make Live-specific plugins called Max for Live Devices. There are a handful of official Ableton plugins that come with it which were made with Max, I'm not sure if it's like that for the standard version but if you have Live Suite then you're automatically able to load MFL devices , edit them, and make your own. Interestingly actually, Live was supposedly prototyped in the Max language back in the 90s lol
why put the dn2 in the thumbnail?
@@christdolphin69 For fun
Thanks for the video. Very educational. But you haven't exactly dispelled the myth that FM synthesis is hard and complicated.
@@SeattleMarc thank you. But I think I did, I tried to demonstrate that understanding FM boils down to understanding the timbral variations you get with only 3 parameters in a 2 operator configuration. I don’t think that’s any more difficult than understanding subtractive synthesis.
curious why the opsix/native was not even...there...
I've never used them
:-( well, I won't buy elektron
@@gossamyr Wow thank you for letting me know!
Breathing and POPPINg in the mic.
Ärm feeling like NF?
I don't think you can pop into a built in laptop mic lol
.5 : 1 is great but .25 : 1 is juuicy
low your gain "&or" comp tresh or shut your window o/ (conseil d'un ingé son français)
Your video is awesome, really helpfull. But I must say one thing. Watch your vocal fry, the video is long and it gets tiring to hear
@@quesadillasdemedianoche well thank you but thats how i talk, so feel free to not listen lol
@@cooksoni.a I did not mean to sound rude, but I guess there is no way to say it in a comment box that it does not sound somewhat awkward. I'm sorry.
@@quesadillasdemedianoche it's fine, sorry I was rude. I just can't really change the way I sound
Ahqaaaaaaajahaaaaa your laugh
Cloneboy
What business have ratio and amplitude to fm. Spread out harmonics, add harmonics? Wth that supposed to mean? Why spreading out harmonics related to frequency modulation, that two different things, how aplitude related to harmonics..that messy