Infinite LFO
does anyone know if there is an abstraction or external for a barber-pole type LFO - sometimes also referred to as 'infinite' or 'through zero' LFO?
I see the Shepard Tone generator example in the Pd patches folder but just wanted something I didn't have to hack...
thanks in advance
kc
I've actually never heard of this, but based on the description here, it shouldn't be difficult:
http://yusynth.net/Modular/EN/QUAD-LFO/index.html
You could just do it with just a few objects (going with some Max-style segmented patch cords here):
[phasor~]
|
|______________________
| | | |
| [+~ .25] [+~ .5] [+~ .75]
| | | |
| [wrap~] [wrap~] [wrap~]
| | | |
[cos~] [cos~] [cos~] [cos~]
thanks for this - I'll give it a shot - I was also looking at the Shepard tone patch in the Pd examples as well as the Risset tone explanation in Dodge-Jerse
are you familiar with the Bode Barberpole Phaser?
edit addition: this is a good starting point for a quadrature lfo but it is not infinite which will require some borrowing from the Risset glissando or something
I'm trying to make this work either in the time domain (w/ delays) or with allpass filters - but I've never been very good at DSP stuff so I'm more apt to attempt the delay patch
alternately, if I could get the [vst~] external to work in Pd I might be able to use this
http://www.olilarkin.co.uk/index.php?p=eseries
is there a known reason why Grill's [vst~] no longer works?
@anechoic said:
edit addition: this is a good starting point for a quadrature lfo but it is not infinite which will require some borrowing from the Risset glissando or something
Yeah but you can modify it from that pretty easily... Subtract 1 from the [cos~] outputs to make your "windows" (the total gain of the output will be -2) and use the wrapped phasor signals as modulators (modulating oscillators or delay lines or whatever). The discontinuities will be where the "window" amplitude is 0. And you can split it as many ways as you want, from 2 on up, so long as the fraction added to the [wrap~] is adjusted.
I haven't tried it, but I think this should work...
The Bode barber pole phaser uses frequency shifters, I think. It's not related to this at all.
IIRC they use a similar 'through-zero' type of control in the Bode barberpole phaser - not sure whether the freq shifter processing section of it uses that or not though...still need to grok how this would work with all pass filters or delays though 
see for some info:
I'm not actually clear on whether or not you are looking to make an infinite phaser, but I was kind of blown away that I'd never heard of this before, so I looked into it a bit.
I spent some time going through Bode's patent for the Barberpole Phaser (or, at least, I think this is the patent for it). The quadrature LFO is just part of the frequency shifter itself. In fact, while it gives a similar illusion, it's not really even designed like Risset's infinite glissando at all (fun fact: that was actually the first thing I successfully made in Pd). The infinite glissando essentially works like acreil's suggestion, in that you have several oscillators that are being windowed, and they are gradually changing in pitch relative to the phase of the window. I tried applying that idea to a flanger (easier to make than a phaser), and I didn't find it particularly convincing.
So I looked at the patent, and it works quite differently. Basically, you send the input through a series of all-pass filters (or a short delay-line for an infinite flanger). Then, you send that signal through a frequency shifter. When you sum the wet and dry signal up, the notches continuously go up or down (depending on if you shifted it up or down) in one direction. The amount of the frequency shift determines the time it takes for one notch to go from min to max.
Now, it's not outright said in the patent, but this is what I think is happening: it's a very novel way of using beating. When you add two sinewaves that are only 2 Hz apart, you hear volumn fluctuations at 2 Hz. In this case, all of the frequencies in the dry signal are shifted by the same amount in the wet signal, so they beat at the same frequency. It just so happens that the phase shifts of the all-pass filters make the beating controlled, so that the notches end up flowing across the spectrum at the rate of the beating. That's my guess, anyway (and I have no idea if that was even clear).
The attached patch is my attempt at making an infinite flanger using this method. [hilbert~] and [complex-mod~] combine to make the frequency shifter. To make this a phaser, just replace [cyclone/comb~] with a series of all-pass filters.
wow! thanks so much for doing this...yes I'm trying to implement a through zero phaser which is also called (erroneously) the barberpole phaser. As I understand it the same basic idea used in the Shepard Tone generator can be applied to both pitch and phase parameters. As DSP has never been my core competency I was trying to work out how I'd approach hacking the Pd example for generating Shepard Tone to shift phase.
My thinking was if I were to ramp several allpass filters and fade in/out their outputs at 120degree phases this might do the trick. But I hadn't gotten that far.
Essentially I'm trying to replicate a soundfile I DL'd which has a triplet of sines being continuously phased past a dry version of itself.
Again I thank you for your help - and I'll take a look at your patch! I'm sure I'll have some questions about it
love the infinite flanger!
@Maelstorm said:
I've actually never heard of this, but based on the description here, it shouldn't be difficult:
http://yusynth.net/Modular/EN/QUAD-LFO/index.html
You could just do it with just a few objects (going with some Max-style segmented patch cords here):
[phasor~]
|
|______________________
| | | |
| [+~ .25] [+~ .5] [+~ .75]
| | | |
| [wrap~] [wrap~] [wrap~]
| | | |
[cos~] [cos~] [cos~] [cos~]
Hello,
I tried doing cuadrature in other form:
[osc~]
| [r~ 1] [r~ 2]
| | |
[hilbert~] [hilbert~] [hilbert~]
| | | |
| [s~ 1] [*~ -1] |
| | [s~ 2] |
[outlet~][outlet~] [outlet~] [outlet~]
But in my patch something dont works, because if every outlet is conected to dac~ objet, there are not total silence.
I see in my patch that the output 1 is not canceled with output 3 and the same with the outputs 2 and 4.
I wish know another way of doing cuadrature...
Thanks!
[hilbert~] doesn't simply pass the input and give a second signal that is phase-shifted by 90 degrees. It sends the input to two all-pass filters that results in two signals that are 90 degrees out-of-phase with each other. So BOTH output signals are phase-shifted versions of the input. It's not really meant for this sort of thing. You'd be better off using the version I gave, if possible. It would also be less CPU intensive.
Thanks, what a mistake!
the reason because I am trying another way of doing cuadrature is due to I am trying to do a cuadrature lfo with the waveforms of saw, triangle, square and sine.
if is possible it seems to me very interesting such tipe of lfo
bye!
It's possible. Just replace the [cos~] objects with the shapes you want.
Thanks Maelstorm,
Has been easier than I thought at first.
Finally, the quadrature oscillator very easy to use and functional with 4 different waveforms.
Goodbye!
http://www.pdpatchrepo.info/hurleur/Complex_Cuadrature-LFO.tgz
Thanks Maelstorm,
Has been easier than I thought at first.
Finally, the quadrature oscillator very easy to use and functional with 4 different waveforms.
Goodbye!
http://www.pdpatchrepo.info/hurleur/Complex_Cuadrature-LFO.tgz
This thread is very cool!
@complex
What does the vol~ external do in your patch? Also from which library is it from? Best Regards
Thank you whale! Nice to see its a abstraction than an external
Oops! Looks like something went wrong!
