FAQ & Terms of UseFAQ & Terms Of Use   Wiggler RadioMW Radio   Muff Wiggler TwitterTwitter   Support the site @ PatreonPatreon 
 SearchSearch   RegisterSign up   Log inLog in 

Notches in Omni
MUFF WIGGLER Forum Index -> Wiard  
Author Notches in Omni
I'm sure I'm showing my ignorance here but I've been doing some simple demos for the omni filter.

I've always liked a number of the middle settings - ie when you get two lights coming up. I always thought that when you have both LP and BP or both BP and HP then you were simply in an intermediate state.

Doing some sweeps and then looking on a scope I'm seeing a notch appear - I know that when you put filters of different types in series or parallel you then create new types but I thought this was a simple cross fade.
Any video of the scope? I'd be interested in seeing the effect.
hi, sorry I didn't see your post.

I'm not very technical - particularly outside of the synth world but I'm expecting to have a bit more time and I'll see what I can do.
I believe the core topology is State Variable which has already multiple outputs available. I remember reading that the Mode morphing is not achieved through basic crossfading between them. I'm not sure where I got the impression from, but I think it is done by dithering, using a high frequency signal (higher than can be perceived) to crossfade between the different filter modes, such that the ear perceives both modes simultaneously. These overlapping modes might be the reason for the notch?
Dr. Sketch-n-Etch
Blending LP and HP from a state variable filter to get a notch is an old and well known technique. It was employed on the Oberheim SEM. I just added the feature to my handmade 2164-based SVF design (it's dead easy to do).
[quote="Dr. Sketch-n-Etch"]Blending LP and HP from a state variable filter to get a notch is an old and well known technique. It was employed on the Oberheim SEM. I just added the feature to my handmade 2164-based SVF design (it's dead easy to do).[/quote]

Yeah, but that's not what the Omni Filter does at all.

All, just look at the 18 year old manual on for a block diagram that gives an idea of how the mode "morphing" is achieved. It may be dated (i.e. details may not pertain to all revisions, idk) but the principle is always ultrasonic rate switching between two adjacent outputs. Not much to do with crossfading nor with a real notch out (which, if I'm not mistaken, is provided separately via the "Allpass" aka Phaser mode - anyone? Take into account the Omni is 24dB, not 12 like a standard SVF). If there are notches in the intermediate states they are likely from arbitrary phase cancellations. 'Arbitrary' bc it's a bit messy to calibrate and these states aren't really defined in a proper theory as you tried here.

PS - I'm not clear on the waveform that does the switching. Grant sometimes referred to "PWMing" between the modes (other terms he used were seguing and time sharing), but the diagram states a sawtooth. Then again maybe the sawtooth is used with a comparator to get PWM? Anyone? I speculate that would be the easiest way. Comp. reference voltage = percentage of time sharing? Maybe someone tech savvy can figure out by reading the circuit board.
Dr. Sketch-n-Etch
What you're describing sounds like "rastering". Grant Richter published another circuit where this was done to "crossfade" between two signals by switching between them quickly with PWM. I've tried it and it works, although it has some issues.
There's a block-diagram on the site. It looks like it's actually a four pole design going into some kind of matrix/mux. I remember the original description being something like "ambitious OTA filter design" or the like. So it's probably four OTA-Integrator poles. (LM13600 I would imagine from that time or maybe 3080/3280.) Then there's a 50KHz oscillator feeding an ADC which is controlling this matrix. The outputs of the mux get summed for the final output.

Pretty cool idea.

I'm curious what the switch matrix is. Could be a CD405x, or it could be a fancier version like a DG-series equivalent. Or, maybe it's just standard CMOS switches configured as a matrix.

I thought about trying something similar to this at one point, (after seeing the diagram) but I got sidetracked with other things.

Edit: Oops. I missed the post above that already mentioned the manual/diagram.
Let me be a stickler here, not bc I was an expert (I'm not), just to be clear on what the module does... There are some known errors in the old manuals (especially the Waveform City one), and I think the terms "poles" and "matrix" in this diagram might be misleading. Please correct me if necessary, but isn't a matrix at least two-dimensional? The Omni mode morphing goes along one axis only, unlike e.g. the Boogie that can do more arbitrary morphing when used with a mixer. Maybe "poles" was used to refer to the supposed matrix' corners. If it's a matrix, it apparently gets de-multiplexed into a linear sequence, right? Otherwise you could go from any mode into any other. That would be cool! Now, if it's not a matrix I think it should say "4 modes" (4 individual mode outputs) instead of "4 poles"?
Hope this was intelligible enough to comment on, whether wrong or right! Intriguing subject.

Also, by the way, anyone know why the BP mode at the 12 dB output sounds almost the same as LP while it's certainly true BP at the 24 out?
MUFF WIGGLER Forum Index -> Wiard  
Page 1 of 1
Powered by phpBB © phpBB Group