easiest way to display filter behaviours

as shown in vadim's last talk
https://www.native-instruments.com/en/reaktor-community/reaktor-user-library/entry/show/12687/
he prooved some anti-aliased visuals of his filters (i say anti-aliased visuals, cause i never saw it that way in reaktor)



beside the aa-visual stuff,: what would be the easiest way to reflect filter behavious visually in real time? (not sending white noise through an fftanalyzer) and how?

This one, oh yes i am pretty hot to know about.
I know, or I might guess that the visual display is still realized by a copy of the filters core macro, BUT wouldnt it be the greatest to have a visual representation of every changements you would do in core.
(ok, i know this wont be possible for now, but just a thought)

 

as i love to auto-reply my posts... (i think it can be always usefull for someone) the minor anti alias issue on visuals in this post solved itself by just re-opening the ensemble.

(is this a bug on the new multi display which was introduced september 2016 or normal primary behaviour? i couldnt tell)

so here an example on the factory 3 band eq with object type -6 on the left and -5 on the right. (for me, any changements i set on multidisplay only took place by not only restarting the instrument through its integrated primary switch (which created weirdo plane artifacts in half of the square display horizontally) but finally restarting the reaktor engine did the job..)



there were other visual artifacts while changing values on multi display (and even in this example there are by looking closely to what is going on on the left side of each square, but for what i want to achieve, this is fairly enough.)
i didnt find many examples in the UL about the new aa integration in mdisp, but Paul Rogat got some and did it beautifully with his LFO's and other recent blocks (which i can recommend to everyone)

i wasnt able to change the line width at all btw. Paul Rogat's LFO's use a setting of 2 for LW, but nothing i tried did change my curves in LW so far...

but still, i am more interested into having a great visual representation of real-time changements in filter macros. "Real-time" in the sense of while/during i change its structure.

And i do think, i am not the only one who loves to see what is really going on - "visually" - while messing up fitler macros.

so let's see how far this post could spread out interest

 

Use the LW input on Multi-Display for the curve thickness

 

Vadim said in the other thread to look in the Factory VA Filter.
In there you'll find the use of a transfer function macro for that particular filter type, and the supporting code required to throw it onto the screen via a multi-display.

 

to change the line make the following changes in both of these macros and add the value module to the line width:

 

the issue i am focusing on:

here's an fft analyzer served by white noise interpretating the factory integrated VA FILTER in primary. (eyes on the 2 screens on the right)
is it useless to think that multi display coud also somehow represent the real harmonics recreated by the resonance?

 

THIS DID THE TRICK!!! "value module" thanks bolabo! (viewing last 3 posts just right now)

 

fairly exaggerated example.

it changes the plane grid. where does this grid come from?

[edit: ok i set idx on mdisp plane on 0 for the following, which therefore wont show up "plane"]

 

yes he did link me on va filter. but why is that a "particular filter type" ? /beside not knowing which of all macros are "transfer functions".

to summarize my post:

Is it possible to visualize changements in the filter structure directly/ means "real time" and coherant to any possible scenarios like feedback oscillation or other wonders ?

 

The harmonics are not created by the resonance, but by the nonlinearities in the filter. It's just that at stronger resonance settings these harmonics become more prominent. However, they are strongly dependent on the input material, thus there is no one curve (independent of the input) which can show them.

 

This was just one example of how to plot the amplitude response, illustrating a general principle. The TF macros contained in the library complement the filters themselves, so that for most of the library filters there is a ready corresponding TF macro. Please be aware that TF (tranfer functions) exist only for linear filters, for nonlinear filters they can only approximate the filter behavior and even that not for all cases (particularly self-oscillation is "out of range"). Actually in the talk itself the transfer function display was intended for linear filters, which I was demonstrating at an earlier point, for nonlinear filters, I'm just showing "some linear curve of a similar nature", very qualitatively.

 

thank you, Vadim! now it is pretty much clear!

I asume, that I never looked into TF and actually thought it was just a "converting macro" after a vcf core macro to do decent display behaviours. Just now I discover that in the library you prevent a TF for each VCF.

Comin back to the title of this thread, converting all vcf macro modulations possible into a TF is a no go for me.
I would dream about to have a visual response in real time to every changements in vcf macros. in a great average.

here is what i came up with to demonstrate a try out to come closer to this:




each row captures/interprets the same response in 3 different ways.

1st TF, 2nd is made use out of what i found in UL from 2010 "visual filter by adam hanley", 3rd is FFT Analyzer on white noise.

the 2nd is not perfect, but i get much more average information out of it than on WN into FFT Analyzer. (btw, is there a way to round/smooth this last one?)

the 2nd wont reply correctly on any self-oscillation behaviour in its state of now, btw.

So back to the topic again, since i dont believe NI willl come up with a vcf macro into vcf (tf) macro converter, would it be possible to have something like this: (which sends sine sweeps on clk rate and re-interprets):



I THINK, THIS ONE WOULD BE THE COOLEST GOAL FOR ANY CODER IN REAKTOR. TO ACHIEVE

 

btw, if anyone wants me to upload the ens example of this last graphic. put a thumb up

 

just kiddin
(for the pro's - i didnt zip this file - it is small enough)


btw, factory VA FILTER got a bug: type and slope were mixed up.