FM'ing 3 or more different Oscillators ( Is it Possible )

Anything modular synth related that is not format specific.
Post Reply
User avatar
PietroC
Common Wiggler
Posts: 114
Joined: Fri Apr 05, 2019 2:17 pm
Location: Montreal

FM'ing 3 or more different Oscillators ( Is it Possible )

Post by PietroC » Sun Oct 06, 2019 5:40 am

Hello
Little Noob Dilemma
Let's say i have 3 Oscillators
i wanted them all to have the same amount of frequency modulation

Option 1 ( The Only way i know of )
Mult my FM Signal to each FM Input for 3 OSC
Adjust each attenuator independently

Option 2
Is FM'ing Externally possible to achieve ????

Example
3 OSC--->Mixer
FM'ing the SUM

Thanks To All for the Countless Help and Knowledge
( Also I believe this post should have been in Modular Synth Discussion but topic classifying is a bit confusing to me )

User avatar
authorless
Super Deluxe Wiggler
Posts: 2335
Joined: Tue Feb 17, 2009 10:41 am
Location: Wilmington, NC

Post by authorless » Sun Oct 06, 2019 6:44 am

To make this a little more clear, you want three carriers all modulated with the same modulator by the same index?

I would send the modulator through a VCA or attenuator then mult the output to each carrier. That way you have a single control for all three indices.
"Wait a month and buy from a fellow wiggler who's desperate for new modules." - Johnisfaster

"It's oscillators are so precise and lifeless it's actually a digital modelling analog synth." - nadafarms

Arneb
Veteran Wiggler
Posts: 671
Joined: Thu Jul 04, 2019 4:50 pm

Re: FM'ing 3 or more different Oscillators ( Is it Possible

Post by Arneb » Sun Oct 06, 2019 7:13 am

PietroC wrote:Option 2
Is FM'ing Externally possible to achieve ????

Example
3 OSC--->Mixer
FM'ing the SUM
I'm not exactly an FM expert, but I don't think this works. IIUC FM is not a form of waveshaping - it physically works by manipulating the VCO itself, not the waveform generated by the VCO.
Last edited by Arneb on Sun Oct 06, 2019 7:14 am, edited 1 time in total.

User avatar
ranix
Super Deluxe Wiggler
Posts: 1770
Joined: Wed Jul 13, 2016 9:39 pm

Post by ranix » Sun Oct 06, 2019 7:14 am

Not sure what you mean by "FMing externally" but check out the Happy Nerding FM Aid. It's my favorite module

User avatar
oberdada
Common Wiggler
Posts: 234
Joined: Tue Nov 01, 2016 12:06 pm

Post by oberdada » Sun Oct 06, 2019 9:20 am

Why is it important to have the same amount of modulation?

If you feed one modulator into the FM inputs of three oscillators, it would be really important to adjust the carrier frequencies of all three oscillators, whereas the difference due to minor variations in modulation amount would be less noticeable.

Mixing three oscillators and using that signal as FM input to a single oscillator is a completely different setup.

FM of sorts can be applied to any audio signal. You have to use a delay, then modulate the delay time.

User avatar
cornutt
Fig juggler
Posts: 1383
Joined: Sun Jan 25, 2009 10:06 pm
Location: Rocket City USA

Post by cornutt » Sun Oct 06, 2019 9:34 am

oberdada wrote:
FM of sorts can be applied to any audio signal. You have to use a delay, then modulate the delay time.
Technically that's phase modulation rather than frequency modulation. It can achieve similar effects, though.
Sequence 15 -- sequence15.blogspot.com

User avatar
starthief
Super Deluxe Wiggler
Posts: 4587
Joined: Tue Oct 04, 2016 10:39 am
Location: St Louis, MO
Contact:

Re: FM'ing 3 or more different Oscillators ( Is it Possible

Post by starthief » Sun Oct 06, 2019 11:00 am

The first method (multing the modulator to 3 carriers) is the correct way. Other methods are going to give you interesting results though and they're worth trying.

As for "external FM", or rather PM, here are your choices, in approximate decreasing order of accuracy and sanity:

- Imitate phase modulation with a delay. No feedback, 100% wet, short delay time and attenuate the modulation amount a lot. The delay has to doppler shift (not crossfade) with delay time changes, and it has to accept audio rate modulation -- some delays will slew or filter the CV and that won't work. Disting's delay algorithm works for this.

- Something like FM Aid or Sinulator. This is also a kind of phase modulation. It's meant to work with sawtooth waves for the input but it will still do something interesting with most signals (not square/pulse).

- You can imitate FM Aid by mixing your carrier and modulator and running the result into a wavefolder, or better yet, a sine lookup table -- something like Piston Honda's external input, WMD Geiger Counter or perhaps Starling Via Scanner. The relative gain of your carrier and modulator are going to be very important here.

- If you have a VCO that will go down to low LFO rates but will phase modulate strongly enough, or which supports thru-zero FM, you can use it as a "silent carrier" -- or really, like a wavefolder! -- and mix your audible carrier and modulator into the modulation input. I've done that with Synth Tech E352 and E370.

User avatar
ranix
Super Deluxe Wiggler
Posts: 1770
Joined: Wed Jul 13, 2016 9:39 pm

Post by ranix » Sun Oct 06, 2019 11:33 am

oberdada wrote:Why is it important to have the same amount of modulation?

If you feed one modulator into the FM inputs of three oscillators, it would be really important to adjust the carrier frequencies of all three oscillators, whereas the difference due to minor variations in modulation amount would be less noticeable.
When stacking modulators it becomes important to sync the carriers and the modulators all to the same source, or you get detune that injects unpleasant inharmonic content. Like, if you have one carrier and one modulator and the modulator is a few cents off that's not a big deal. But if you have a carrier, a modulator affecting the carrier, and another modulator modulating the modulator, the effects start to get noticeable. Or in anything more complex than that. Filtering out the unpleasantness (I've been yelled at around here before for calling this "noise") is tough to do without also filtering the high frequencies you want to hear.

In my patches I work around this in various ways. This is where music theory starts to matter. I don't know how to communicate these ideas very well. I'm not really a musician myself I just watched a bunch of youtube videos... but I'll try.

Sometimes I want more than one carrier running at the (exact) same frequency. Sometimes I want a carrier running (exactly, no detune) a fourth or a fifth up or down from the first carrier. Or a modulator that's exactly a fifth up or down from the carrier, or from another modulator. On a DX style synth, these intervals are represented by frequency ratios, i.e. a fifth is a 2:3 (or a 3:2) ratio. The carrier could be peaking exactly 2 times for every 3 times the modulator peaks, or vice-versa.

One way I get an exact 2:3 ratio is by sending an oscillator's square wave output through an audio-rate clock divider. The oscillator is running at frequency 2. I can get a square wave with frequency 1 from the /2 output of the clock divider.

To get 3, we have to get a signal that's running exactly 3 times the speed of 1. Since our 1 is a 50% square wave we know it contains odd harmonics, so each harmonic is going to be i.e. 3 times, 5 times, 7 times, etc. the frequency of the fundamental (aka 1). So we can get 3 by sending the square wave to a band-pass filter that isolates only the third harmonic. That gives us a sine wave running at frequency 3, which we can use as our modulator. I have a state variable filter I use for that which seems to work well.

If you're using an FM Aid it's driven from our oscillator which is running at speed 2. The modulator will be the output of our band-pass filter, which is a sine wave of speed 3. That gives us a perfect 2:3 ratio between the carrier and the modulator.

If you're using an FM oscillator instead of FM Aids you could use an output from your first oscillator to sync the FM oscillator. This FM oscillator would be your carrier which would be running at speed 2. But I don't know how sync would affect the modulation of such an oscillator - if the sync interferes with the modulation you'd have problems. This isn't a problem with a module like the FM Aid, since the modulation of the FM Aid does not affect the base frequency.

There are an awful lot of things you can do to get various frequency relationships between signals using the above tools, and an awful lot of frequency ratios you can get from them. And you can use clock multipliers instead of dividers, and multipliers and dividers sent through filters track at different speeds, and the phase of any synced oscillators matters... there are lots of considerations and options.

User avatar
PietroC
Common Wiggler
Posts: 114
Joined: Fri Apr 05, 2019 2:17 pm
Location: Montreal

Re: FM'ing 3 or more different Oscillators ( Is it Possible

Post by PietroC » Mon Oct 07, 2019 2:49 am

authorless wrote: I would send the modulator through a VCA or attenuator then mult the output to each carrier. That way you have a single control for all three indices.
Reverse Engineering ( Thats Beautiful ) I did not think of that
It would simplify the process of adjusting evenly 3 attenuverters

So i guess reverse my question to be clearer

Oscillators are only Frequency Modulatable Internally
I would not be able to Group many voices then apply a tremolo after they are mixed

but
Arneb wrote: IIUC FM is not a form of waveshaping - it physically works by manipulating the VCO itself, not the waveform generated by the VCO.
This pretty much clears up my question
Makes sense since cv signals cannot be configured to target only pitch
Hence the FM Input on Oscillators
starthief wrote:
ranix wrote:
Thanks So Much for all this info
Im new to this and your explanations are very clear but obviously
I have much Learning / Explorations to do

My Dilemma Began cause
- i have only on Buff Mult
- I want to avoid further purchases
- Curiousity
- Trying to get the sounds perfect for sound layering / sculpting

Weapons Of Choice
https://www.modulargrid.net/e/racks/view/951326

User avatar
Blairio
Super Deluxe Wiggler
Posts: 1906
Joined: Mon Oct 07, 2013 9:50 pm
Location: Glasgow, Scotland

Post by Blairio » Tue Oct 08, 2019 9:35 pm

Yamaha's 6 operator DX7 had about thirty different 'operator as carrier or modulator' algorithms.

Even the current Reface DX has twelve such 4 operator algorithms,

I guess that in theory (at least) any of these multiple operator algorithms could be implemented in Eurorack?

User avatar
cornutt
Fig juggler
Posts: 1383
Joined: Sun Jan 25, 2009 10:06 pm
Location: Rocket City USA

Post by cornutt » Wed Oct 09, 2019 12:02 pm

Blairio wrote:
I guess that in theory (at least) any of these multiple operator algorithms could be implemented in Eurorack?
Sure. You can do it with any VCOs, if you have enough. HOWEVER... you will only get "mathematically correct" results if your VCOs have linear FM inputs. Now, "artistically correct" is another matter, and you can get many interesting things using the ordinary volt/octave inputs. So grab some VCOs, patch them up, and see what you get.

I'll add this: In FM, if you use harmonic-rich waveforms (pulse, sawtooth) as operators, the results you get will trend strongly towards noise. Start with sines and triangles. Remember, the DX7 operators only produced sine waves.
Sequence 15 -- sequence15.blogspot.com

User avatar
pricklyrobot
Veteran Wiggler
Posts: 531
Joined: Sun Mar 06, 2011 9:09 am
Location: Austin, Tex-ass

Re: FM'ing 3 or more different Oscillators ( Is it Possible

Post by pricklyrobot » Wed Oct 09, 2019 12:21 pm

PietroC wrote: Oscillators are only Frequency Modulatable Internally
I would not be able to Group many voices then apply a tremolo after they are mixed
If tremolo was what you were after, this approach would work, but FM is basically vibrato. So you either have to modulate the pitch at the source (VCO), or you do it with a delay line (this is how Vibrato FX units do it).

User avatar
Blairio
Super Deluxe Wiggler
Posts: 1906
Joined: Mon Oct 07, 2013 9:50 pm
Location: Glasgow, Scotland

Post by Blairio » Wed Oct 09, 2019 12:21 pm

cornutt wrote:
Blairio wrote:
I guess that in theory (at least) any of these multiple operator algorithms could be implemented in Eurorack?
Sure. You can do it with any VCOs, if you have enough. HOWEVER... you will only get "mathematically correct" results if your VCOs have linear FM inputs. Now, "artistically correct" is another matter, and you can get many interesting things using the ordinary volt/octave inputs. So grab some VCOs, patch them up, and see what you get.

I'll add this: In FM, if you use harmonic-rich waveforms (pulse, sawtooth) as operators, the results you get will trend strongly towards noise. Start with sines and triangles. Remember, the DX7 operators only produced sine waves.
Yes, the DX7s (and all their antecedents) operators only produced sine waves, but the Reface DX operators can produce a range of waveforms from square through sine through sawtooth. The result is harmonically richer than with straight sines. I spent years programming DX's back in the 80s. The Reface DX can achieve timbres more complex than the 4 or 6 operator DX7's and DX9's etc in very short order.

Post Reply

Return to “Modular Synth General Discussion”