Home | ER-101 | ER-102 | ER-301 | Wiki | Contact

FM Sound Design


Yes, of course 0.2.12 started me right down the rabbit hole of building an FM synth on the ER-301. I just started simple. A 2 op sine wave FM algorithm, with op2 being the modulator and op1 being the carrier.

It quickly dawned on me that I’d want easy access to some parameters so that I didn’t have to dive to adjust them.

Even for two ops, it turned out to be quite a few parameters.

Independent ADSR for the carrier and modulator. Pitch for the modulator. Overall ADSR output level for the modulator. So that’s 10 for a 2 op algorithm. Maybe I could get away with AD only on the modulator. That’s still 8, before I start doing anything weird, and only for two ops. :slight_smile:

So I set them up inside a custom unit, and started adding all of these custom controls.

Don’t read this as complaining. I’m thrilled to death that I can do this in modular. Just throwing it out there for discussion. Is this probably the right approach? On the plus side, after building that algorithm it would be reusable.

One thing I found myself wanting to do was add a custom control right at the point where I go to assign it, rather than navigating back to the top.

Another thing I thought about was assigning some key ones to external offsets. In my small system, I only have 8 available right now… and they’re tied to units (e.g. Maths, Frames) that I might want to use for other purposes in a patch. So have to do that sparingly.

Just throwing it out there for discussion.


I was working on the same concept…with the same difficulties.
I was, back in the past, a huge fan of DX7 and i program it a lot.
Something we can do with some algorithm was to feedback an operator into himself, i’m thinking hard, there is a little white smoke out of my ears, but i don’t figure how to accomplish feedback on the modulator ( i managed to did it with the carrier and using a copy of out 1 into out 2 and then out 2 modulating my carrier ( not real feedback because of the slighty delay between the signals )).


since a few weeks i’m back at my dx7 and although i’m getting better and faster at programming it, i think the er-301 is still easier :slight_smile: patching feedback on the er-301 is something i haven’t really figured out, yes, i can route different chains or create custom units but for something like a simple low pass filter inside a delay feedback loop i think it’s a bit too complicated, i wish there was an easier way to do this or maybe i’m missing something?
anyway, the most important part of digital fm synthesis is in my opinion the envelopes and i would like to have direct access to every stage of them so i would choose external adsr envelopes (are there dx style envelopes in eurorack? i don’t think so and yes they are complicated but very very good for setting exact changes over time) so having only the operators and the vca’s inside the 301 makes it a bit more manageable. that would be 2 pitch controls and two vca controls for the simplest 2 operator setup. right?


I can’t think of a way to do it right now, but there is a feedback topological unit on the wish list, so maybe some day? :slight_smile:


For now though, if you were trying to emulate a DX7 algo, and wanted this feedback setup, you could just create a new oscillator to modulate that “op” and set it to the same frequency as the one you’re trying to modulate?

Oh yeah. The DX7 is a nightmare to program. Mine’s been sitting in a closet for ages now. Amazing the literally thousands of great patches (you can still find them if you search) that were created for this thing given the squirrely nature of FM and the interface it had to explore it. But hell, it was the early 80s and there was nothing like it!

I had to go look at the manual. The DX7’s EGs (despite their odd terminology) were just basic ADSRs that had a 2nd decay phase, right?


a second decay stage plus rate and level over each stage, where level 4 is the magic dronometer stage :slight_smile:

it´s completely different to the usual analog adsr envelope but incredibly versatile for all kinds of shapes, percussiv or long and evolving…


Completely possible I am missing something here. :slight_smile:

Outside of the additional 2nd decay stage (labeled R2 in the diagram), I don’t see anything about the DX7 EG that you couldn’t recreate with the following unit configuration in the ER-301:

ADSR --> Linear VCA --> Offset

As kind of an aside, one of [various] reasons I added a Mutable Instruments Frames to my rig was that I was hoping to use it as a very flexible, multi-segment envelope generator. I have not quite gotten around to fleshing that out. Perhaps I should bump it back up on my experiment list!


Also remember that there is absolutely nothing stopping you generating complex envelopes using an external step sequencer (so long as it has slide or equivalent) and recording the CV in the ER-301 - you want an 8, 12, 16, even 64 stage envelope - it really is no problem at all!!

If I remember correctly @NeilParfitt did a video demonstrating this, but I’ve no idea how to find that part!



Great suggestion!

Pretty sure it was @NeilParfitt’s video #12 on creating long envelopes.

Seems like this topic and the other one on understanding linear TZFM are starting to converge a bit on “can I recreate the DX7 inside of the ER-301”… :slight_smile:

I’ll put my thoughts here. So, yes, this is essentially what I set out to do last night. Recreate one of the DX7 algorithms in the 301. I guess my purpose in doing so was more for comparison’s sake. How well can the ER-301 nail it?

In the grander scheme of things, I’m of a divided mind on being able to do something like load a DX7 SYSEX patch directly into the ER-301. On the one hand, it would be really cool. On the other, I already have a software synth that can do that, and in addition to taking FM to the next level by offering non-sinusoidal waveforms, multi-segment repeating envelopes, a slew of FX, etc., it’s UI for working with FM is pretty splendid. Don’t know that the ER-301 would be able to rival that on it’s LCD display.

BUT, where I think the magic really comes in here is that FM8 is still a fixed architecture synth. The ER-301 is definitely not. So if I wanted one of my operators to be a sample player instead of a sine, saw, or pulse wave, I’m just a couple of button presses and one big beautiful rotary encoder away. :heart_eyes_cat: And that’s just one example of the countless ways you could change the DX7’s architecture and create something brand new in the FM space.


BUT, where I think the magic really comes in here is that FM8 is still a fixed architecture synth. The ER-301 is definitely not. So if I wanted one of my operators to be a sample player instead of a sine, saw, or pulse wave, I’m just a couple of button presses and one big beautiful rotary encoder away. :heart_eyes_cat: And that’s just one example of the countless ways you could change the DX7’s architecture and create something brand new in the FM space.

This is absolutely fascinating… Its like a playground for user designed FM algorithms.


I don’t think it will be a real feedback unless i modulate this modulator with the carrier, i try it immediatly
Edit : I’m stupid, i can’t do that …


a normal adsr has only one stage that is “amount” or level eg. the sustain stage, all other stages are time based and you can’t change the level, that’s the main point of the dx envelopes, that you have the normal rate, speed control but also each level on your hand.


I don’t think I ever knew this, and I am unclear how this works, so please humour me;

Lets look at one section in isolation, say the Attack phase, as I understand it Attack is the time it takes to go from zero to maximum, zero and maximum being defined higher in the chain by the amount of bias and gain applied to the VCA - yes?

So what constitutes setting a level on the Attack phase?


It would not be real feedback, of course. But if whatever op you’re trying to feedback on itself has a constant frequency, and you modulate it with another oscillator with the exact same constant frequency, it would be the same net result, I think.

I’m looking at the diagram above.

L1 is the peak level of the attack phase - could be set with a linear VCA
R1 - regular attack time in ADSR
L2/R2 - corresponds to the additional decay phase - not possible
L3 - set with sustain level in ADSR
R3 - set with D parameter in ADSR
R4 - set with R parameter in ADSR
L4 - set with an offset

Am I still missing something? Still looks possible to me to cobble all but the extra decay phase together from units in the ER-301. :slight_smile: May not be as straightforward to configure. And I guess if you’re going for a different shape than this - that might have been possible with the DX-7. Not sure it would be using this setup.


Kichhofer is totally right, in fact the dx7 was my first synth, i think i bought one of the first in France. So years later when i use another one ( JP8 ) iwas very surprised to don’t find those level on the ADSR.


In fact when you have multiple segments as on a dx7 it make sense to have level for each point.
You can sculpt the exact shape for your envelop.
I loved to do very slow evolving sounds with my dx.
A thing you can do is to have the first level L1 smaller than the L2 or L3 or L4.
The control forge give me this sort of possibility in the modular world.


Ah - I get it now. :slight_smile:

Like I said, DX7 has been in a closet for quite some time.


Very interesting. This is one of those things I thought I might be able to accomplish with Frames (I’m not sure I can’t but it’s not as straightforward as I’d hoped). Actually DX7 wasn’t so much on my mind at that point as Absynth, and the wonderful endless envelopes you can create - unlimited points, and you have control over timing and level of each point.

Will have to look further at Control Forge. Monome Teletype looks like it might also be able to do some EG gymnastics like this too but I need to understand it a little better. Neither of these two are lightweight purchases. :wink:


ha that’s great, for alot of users it was the other way around, they came from the classic adsr and couldn’t handle the dx approach.
…maybe brian builds us one of these for the compact and lightweight fm synthesis setup :blush::


Really, that looks like just about what you need to “conveniently” explore the vast world of FM. :smiley_cat:


No, unfortunatly it doesn’t do the same thing at all, tomorrow i’ll make you a little video you can see and hear the enormous difference.