Understanding Through Zero FM

I must admit I am uncertain where the opening up of TZFM is evident in the modules which Brian mentioned. I’m aware that I have been producing FM synthesis (although unknowingly prior to reading the SOS article and watching the video!) but I’m unsure of how there is something different in the Ladder Filter, Sine Osc, Aliasing Saw, and Aliasing Triangle. Is f0 on the frequency of these units indicative of the change here?

I believe you’ll have to modulate the hertz control, that is, the f0 control. that gives you linear thru-zero FM. modulating the V/Oct control would give exponential FM. (EDIT: previously I said exponential thru-zero FM, which is not the case, of course!)

Thanks for this, I love to hear demos and comparisons without specific musical context, really inspiring.

Yes, that’s what I thought, but just looking at the modulation indicators movement when feeding in a modulating triangle wave, it looks to be still exponential i.e. the modulation indicator of f0 drops far lower beneath the bias setting than it does on the upward movement.

Sorry if my language not precise or technically correct, but does that make sense?

(actually on reflection it appears to be the inverse of that which was described in the video Joe offered)

Rest assured that f0 modulation is linear. Only the fader display itself is scaled exponentially, because otherwise you would not see any modulation.

1 Like

FYI
There is no such thing as thru-zero exponential FM because the exponential function is always positive.

Please allow me to recall a previous post I made that clarifies:

http://forum.orthogonaldevices.com/uploads/default/original/1X/6817d1c45999831e931ed0378bc13d503d98e070.png

Thanks Brian,

But now, unlike in your diagram below, the f0 function is able to modulated? (and that’s the unlocking part?)

Exactly!

See, I’m learnin’ day by day! :slight_smile:

So in that instance where I have a triangle LFO modulating f0 and on the lower scope it shows as equidistant about the midline yet on the f0 modulation the bulk of the modulation occurs beneath the the bias amount, very little above it, what’s going on there? Do I need to offset or attenuate that modulation signal coming in?

Only the fader display is scaled exponentially. That is why the modulation looks lopsided. A linear display of frequency is almost useless. Can you see why? Imagine if the fader tried to display 0Hz to 12000Hz in a linear fashion? Where would 300Hz be on the fader? How big would a modulation index have to be before you can see it?

This makes very much sense. Here I go, giving advice without the credentials – sorry for any fake news from my part! :slight_smile:

EDIT: I did wonder about the modulation display but didn’t give it that much thought because it sounded right. Also, unstable release. But good catch!

OK, got you. And this is why the lopsidedness changes in relation to the fundamental frequency, less lopsided movement the higher you go?

Yes, that is right. Exponential (also known as octave and log) scaling will compress higher frequencies and expand the lower frequencies.

Many thanks.

1 Like

from Teezer Through Zero VCO - Richard Brewster's Electronic Sounds

“The negative-frequency version of a waveform is simply a time-reversed replica of the original waveform. When a VCO is modulated through zero frequency, the waveform slows down to a stop and then speeds back up in the reverse direction.”

And here is another thread on Muffs that explains it well:
https://www.muffwiggler.com/forum/viewtopic.php?t=136716

I didn’t find a video easily so I just made a quick one. I’ll share anyway since it illustrates it on the ER-301. As an aside, it doesn’t appear the O|D implementation actually runs backwards- it is still a ramp up at negative frequencies.

1 Like

O snap! Thanks for bringing that up @Joe. Looks like my last round of optimizations removed the phase inversion for negative frequencies. Fixing…

1 Like

Tsk… can’t leave you guys alone for 5 minutes without you go having all these amazing conversations, making videos and finding bugs to fix :joy_cat:

Talk of all the presets made me think of the Volca FM (I’d still like to pick one of these up one day) and how it can load DX7 patches, from the product page:

“The volca fm perfectly reproduces the sound engine of the classic digital synthesizer that made the world aware of FM synthesis. Its six operators and 32 algorithms provide complete compatibility, and it can even load files in the SYS-EX/SYX format of the Yamaha DX7.”

I wonder how feasible it would be to convert the SYS-EX/SYX patches into ER-301 patches?

1 Like

…and fixed in v0.2.13! Sorry about that.

2 Likes