Modular synthesis can be a complicated topic - there are hundreds of modules and a practically infinite number of ways to patch them together. Lots of channels will review a module, or give you surface-level details on how it works, but this channel aims to dive deep and really explain how things work.
Great video, this cleared things up a lot for me. Just out of curiosity, is that spectrum display reacting to the audio from your modular or is it an animation? I noticed that none of my analog modules are capable of producing a sine wave as pure as shown in the graph, even the self-oscillating filters have a few higher harmonics at around -60 db, which makes it a bit more difficult to see what's going on with the side bands.
@nilskozstum8256 - It is being generated by the audio coming out of the modular rig, but since I wrote the code, I get to make decisions about how big frequency spikes have to be before I include them in the animation. Generally speaking though, I'm not intentionally cutting anything out. The only sine wave I've really a real extra frequency component in is the Nonlinear Circuits CEM 3340, and in that case you can distinctly see a little sharp peak right at the apex of the waveform. I'll have to check things again though.
@@SoundVoltage thanks for clearing that up! I guess it's down to how the sine is generated, my main oscillator that I use for FM is the Instruo Cs-L, where one of the oscillators is a triangle core and the other a saw core, which then get filtered/sent through comparators to generate the other waveforms. You can clearly see on the spectrum that the same harmonics that you get in the core waveform are still present, just much more toned down. I don't mind this at all, I love the sound of these oscillators. I was expecting the filters to be cleaner then they are though. Cleanest sine I could get from MI Plaits, unsurprisingly digital will get you the most mathematically pure sine.
I still don't understand the thing about the FM radio tbh. Like why does it need FM in the first place? Which problem is solved by applying FM to some music before sending it out there so car radios can pick it up? or is it before or after? what's each stage like?
@Beatsbasteln - It's not really right to think of it as "applying FM before sending it out". Without any audio signal, the FM transmitter would just output a constant sine wave at the frequency of the station -- say 90.3 Mhz. That pure sine wave is modulated by whatever the audio is. It's not something done ahead of time, the transmitter does it on the fly, wiggling the oscillator a little faster or a little slower than 90.3 Mhz depending on the pitch and amplitude of the signal. That is literally what it means to broadcast in FM.
Very nice series, thank you for your effort! I'm very curious though, what's the story with the carrier frequency dropping & disappearing? Is there another video perhaps about this? Alternatively I'd appreciate if you could share any relevant reading material about that. Thanks again!
@Act1v1st - Yes! That is really interesting. I will be digging into this, but it's a couple videos out still. In the mean time, you can check out Bessel functions -- they basically describe how the sideband levels evolve over time. Also, there's an element of destructive interference that happens when those sidebands go down below zero and "reflect" back over sidebands that are already there. I'll start discussing that last one just briefly in the video that should be out tomorrow morning. The details get involved though, so it's going to take just a bit to get everything lined up for it.
I retuned FM filters (ie the RF transmission component) years ago on work experience, never expected EQ knowledge to come in so handy, literally just turning bolts to dial in a BP on a scope. Great idea for a series, looking forward to jumping down the rabbit hole!
First of all, beautiful video! This may well be the clearest explanation of FM creating sidebands I’ve seen. I tried to work through Bristow’s text back in the 80s but didn’t have the math depth to really get the Bessel functions. It took a broadcasting tech class in college for me to really get it. What are you using to display the sidebands? Looks great!
@Bent6 - Thank you so much! Actually, the sideband spectrum display - along with a lot of the waveform displays, etc. - are all animations that I've done myself, in Python and Javascript.
Recently found that cool DX7 emulator. Decided to practice FM with a more approachable device first. Now we're here. Maybe off by a week, but still good timing ;)
tbh i don't think I would have understand it if I didn't already before I saw the video. The text to speech and different information on the screen makes it very confusing even when you know what's going on.
you mentioned "DX", they are actually PM synthesis not having any sidebands below the route and the same go's for most digital FM synths. they all should be called PM synths.
Great series! The minimalist design works really well here and the pace is just right. Agree with tychoclavius, that when what is written is different to the spoken, my brain goes brrrr. Another thing that came to mind, perhaps you could additionally display the oscillation below the frequency analyzer or on another "slide".
Looking at how the sidebands are symmetric, it just occurred to me that the ones to the left of the carrier are actually undertones, and that this must be one (just one) of the reasons FM timbres often sound inharmonic.
@modalmixture - Ah! We're going to get there! You're right, they are undertones, but in video #4 I think, you'll see something cool where actually the undertone becomes the fundamental for a new 'virtual tone'. But that only works when the carrier & modulator frequency are in a nice ratio to each other. When they aren't? Inharmonic city. Glad you're enjoying the series. #3 should be out on Friday morning I think.
@rockrobertson5776 - That keyboard was surprisingly decent! Cool that you still have yours. It deserves a video all it's own, I wonder if I can find one locally...
really enjoying this so far. good call on breaking it into shorter videos. even short videos you have high information density with this topic so it gives people who are following along food break points to kinda onboard each set of concepts
I found out about this module - and this video! - only a few days after building my own stripboard copy of the old Mutable Instruments module "Kinks", where the binary input section is essentially the same as the Max and Min here. Odd coincidence. Thank you for describing what's going on so well! (The first section of "Kinks", the unary input section, outputs the inverse and the half- and full-wave rectified versions of the input, so yeah.) (I just ordered MM's "Dividers" PCB/panel set as well, so that'll be a fun build.)
Since this is meant to be a complete overview of FM, i think its worth clarifying the whole "deviation range" thing. The way it was described it made it sound like all FM needs to have the carrier wave completely in the middle with equal deviation on other sides. FM can be done with the modulator being a unipolar wave which would only raise (or lower) the pitch of the carrier. Plenty of complex oscillators (ie Verbos Complex Oscillator) work this way, so i definitely think its worth noting at some point in the series. Looking forward to the rest of the series and future videos from you in general!
@speed0 - That's good feedback, thanks. I was going to come back to deviation in the context of exponential FM where the carrier sits about 1/3 the way through the deviation rather than centered -- but I didn't know about the verbos, I'll mention that. Thanks.
Really enjoying these. I've been testing these examples using a VCO2600 as carrier (utilizing it's fm inputs) and the two independent outputs of a cre8audio chipz.
These are so valuable! I'm learning a lot. You really can't overstate the value of an oscilloscope. Seeing what's happening visually makes all the difference. I do agree with the other commenter about text on the screen that doesn't match what you're saying. Some great advice I heard a while back about presentations is not to make the audience read and listen at the same time, cause they can't do both. What I do now is make each slide either a picture or a single word, or at least as few words as possible, specifically avoiding sentences.
I'm not a fan of videos with spoken tekst and different written text at the same time, but I was a brave little boy and managed it. Thanks for the excellent visuals and explanation!
I've known for a while that FM sounds 'nicer' when you use integer divisions of the carrier frequency for the modulating frequency, but that makes a lot more sense now that I know about the sidebands.
This is good, but you did the thing that scientifically literate, and especially mathematically literate people do, you jumped from one thing to another too quickly. What is it that FM Radio is really doing? After all, it isn't broadcasting sine tones! I get it about the broadcast width etc, but how does it carry songs by Elton John and The Wu Tang Clan? Not in the same way a DX7 carries sine tones, presumably.
@teebeedahbow - the two are really very similar in terms of the actual modulation. In the DX-7, you have one audio rate sine wave modulated by an audio rate signal (another sine wave); on the radio you have a much, much higher-than-audio-rate sine wave modulated by an audio signal (Elton and Wu Tang - the crossover we've all been waiting for). Of course the DX-7 turns into audio, which vibrates the air and on into our ears; whereas an FM transmitter wiggles at the electromagnetic field, which is then picked up by an antenna. It's a fair criticism though, and maybe I should come back and do a longer video just on radio -- I didn't want to distract too much from the synthesis side of things, but radio definitely deserved more time. Thanks.
@@SoundVoltage Thanks very much for the explanation. I really enjoyed your video. I was just disappointed with myself that I could make the jump from FM synthesis to FM radio.
@NiamorH - Maybe I'll do a whole separate video on that. I didn't want to take people away from the synth side of things, but it is definitely worth more than the 60 seconds I gave it.
