A historical tidbit for those that don’t know…
The original developer of Surge was Claes Johanson, who then went on to start Bitwig. When he did that he made Surge open source.
6 Likes
Is there a way to attach .vcv files here? I have a nice patch to show it to you. Lets see if this works!
FM2Example.vcv (3.3 KB)
Thanks, that works fine. I will play with it.
1 Like
Just to clarify, when I said it might have been a mistake to compare two FM modules, I meant the FM-OP and the FM2. But your patch does a good job of comparing the FM2 to the FM3. Thank you.
yeah i think to make that patch out of fm-op rather than FM2 I need 6 fm-ops rather than 2 fm2s. that’s what I was trying to show. But I not sure.
1 Like
Ah, now I understand 
I probably like the FM-OP so much because it has few parameters and is easy to musically control as a self-contained voice. I think FM-OP is a good beginner FM module as it is so simple and easy. So, I would characterize the XT FM modules as significantly more complex than the FM-OP. In that regard, I might suggest FM-OP to a newbie as what to explore first and the XT FM modules what to explore next, but that is just my my personal thought on the matter. It will take me longer to fully explore the XT oscillators.
Right! That also explains why we use more CPU than FM-OP (which is what had me looking; and then I realized we were doing more than 3x the work 
)
FM-OP is more akin to the sine oscillator in surge than it is to FM2 or FM3 is the takeaway (but the sine oscillator has a different set of goobits like multiple waveforms and built in filters and unison).
But FM-OP is cool now I’ve learned it!
2 Likes
FM2Example.vcv (4.5 KB)
oh and here’s an expanded version of that patch which is actually almost starting to sound like something musical. Chuckle.
3 Likes
It does. Thanks!
How do you get 3x? One has two oscillators and an adsr, the other has three osc?
we are running 3 oscillators, 2x oversampled, with downsampling filters, yeah.
But also maybe I don’t understand FM-OP. It seems to only have one sine oscillator in it. Where’s the second? We are taking about the BogAudio module called “FM-OP” right which has like a ratio at the top and an FM input with a depth?
1 Like
Yeah FM-OP is just one oscillator right?
You need two of them (or more) to actually do FM.
1 Like
You only need one oscillator to do FM, so long as you can patch feedback.
1 Like
oh, you are right. the only FM is the self feedback. It is 8X oversampled, however.
1 Like
Well, I certainly learned a lot from this discussion. The FM-OP manual says:
A sine-wave oscillator and simple synth voice designed to allow patching up the classic FM algorithms (using multiple instances).
Guess I never read the manual. Now I need to learn what the classic FM algorithms are.
So I have just been using the internal feedback FM… I suppose the FM without the OP…
Sorry for taking this off-topic, but I appreciate everyone taking the time to educate me.
1 Like
This should help.
Crap scraper. That’s a link to Opulus, which really is quite something.
1 Like
Reading more, I was using a single FM operator and no algorithms other than perhaps the self feedback. 
1 Like
So the DX7 and their ilk were collections of connected operators routed to one another in algorithms. For instance the DX7 had 6 “operators” ( in the DX7 case a sine wave with an envelope that could self modulate and take FM input and be tuned to a ratio of the base frequency) which inter-modulated in one of a variety of ways. You would have a card that looked like this and pick different ones.
If you want to play with this, there’s really nothing better than the plogue DX7 vst which shipped last year, but arturia has a good DX7 and dexed is a really competent open source one.
In this language “FM-OP” is an operator in that. A single node in that graph. To make one of those you would wire up 6 FM-Ops with their FM routed in certain ways and you would start to get DX7 style sounds.
The Surge oscillators can be cast in that light.
- FM2 is a 3 operator model which would look like 1 → 1, 2 → 1, 3 → 1 1 → out. The oscillator 2 and 3 have independent depths and a C:M ratio which is fixed at an integer.
- FM3 is a 4 operator model which is 1 → 1, 2 → 1, 3 → 1, 4 → 1, 1 → out. Operator 2 and 3 are at an arbitrary C;M ratio (doesn’t have to be integral) or can also be set to an absolute frequency; Operator 4 is always absolute.
- Because surge is like C++ it counts starting at 0 so “FM2 is 2 modulators, FM3 is 3 modulators” but the DX7 counts the “1” hence the difference in labeling (but the DX7 is not called the DX6 since the 7 means something else but I don’t remember what)
The surge Sine oscillator is a bit more like a single node but from a TX81-Z rather than a DX7, in that it has wave shaping using ‘quadrant shifting’ where you move quadrants around to get odd shapes. It also has unison and a built in simple filter to tame some of those pointy shifts.
All run at 2x oversample.
Hope that helps with surge!
7 Likes
Thanks, that helps a lot. I’ll have to play with the Arturia DX7 VM since I have it. Also, thanks for reminding me of Dexed. I was trying to remember that name since I have played with it in the last couple of years.
Yeah, this all helps with Surge, at least it helps me.
Guys, what an amazing plugin! Thanks ever so much. They look and feel great and the UI is wonderfully consistent. Fantastic job with this. I’m two thirds into testing the VCO’s only - man, there’s a lot of modules to take seriously here - and this is my initial feedback. There might be a couple of bugs in there. Let me know if it’s too much. It might come across overly negative but that’s only because I’m trying not to rave about all that I love about them
The VCO’s in general
- Major kudos on the theming system!
- Love the unison modes, they just lift everything to a higher level.
- In unison modes they can deliver a quite high output level. I wonder if an output level knob would be appropriate, where there’s space. Otherwise we really need attenuators right after the VCO.
- I prefer the “warm” character for all the VCO’s.
- I’m not a fan of the dragable labels for Oct and Uni. I think at least a click on them should pop a menu like the wave selection labels.
- I would really like a DC filter on the outputs, as many of the VCO’s have quite a DC-offset in some modes. Could be optionally engaged from a context menu, if it uses CPU.
- The modulation system is great! Super flexible and compact way of doing it.
The Classic and Modern VCOs
- They both have quite a DC offset when using unison+detune.
- Very free of aliasing, wonderful.
- Modern VCO starts to use more and more CPU when you crank up the unison. Classic VCO does not. Bug?
- They both sound great but I actually prefer Classic VCO for its tonality.
The Sine VCO
- Love it
- Uses only a tiny bit more CPU in unison.
- Lots of DC when using feedback.
- Heavy aliasing with feedback over a certain amount.
- LO/HI cut doesn’t do anything.
The Wavetable VCO
- Great! Does eat some CPU in unison but otherwise lean.
- Would love a longer lead-in on the start of the range of the Formant knob, it can be from 0 to very abrubt sometimes.
- Some of the waves are great, some don’t have much musical relevance or just seem off, with hardly any variance in them. Scratching my head sometimes.
The String VCO
- Fantastic, love it!
- The attack gets quite clicky though, wish there was a way to soften it a bit.
- High DC offset on some settings.
- The constant* exiters seem difficult to make use of musically.
The Alias VCO
- Has a lot of aliasing, maybe that’s the idea.
- Not for me
The FM2 and FM3 VCO’s
- Very clean and efficient, sounds great, love it.
- High DC offset on some settings.
The Twist VCO
- Sounds great!
- Can’t stop it droning and use the trig input. Feels like a bug.
7 Likes