Airwindows UltrasonX: Mac/Windows/Linux AU/VST

[jinxtigr]

TL;DW: UltrasonX is a method for rolling your own Console-type systems with total control over your ultrasonic filtering.

UltrasonX.zip(588k)

A little while back, I made a simpler 'Ultrasonic' filter, with the intention of making a lighter-weight utility that could be used where Ultrasonic might go. The idea was, Ultrasonic (which is already available) was too heavy: it uses five stages of biquad filtering and I thought doing the same thing with just one or two stages would be better.

Turns out you can do that, but if you stack them up you start losing the super-highs. The one or two-stage versions weren't steep enough: didn't bring out the super highs close enough to the ultrasonic zone. But, the five stage version is still just too much processing to put all over the mix. What do?

UltrasonX solves this problem (that maybe nobody but me had ).

This is a plugin that does any one of the five stages in Ultrasonic, one at a time. It's got settings for A, B, C, D and E stages. Each of these are a carefully calibrated resonance value, that add up to nice and flat and clear all the way up to the supersonic region.

Console7Cascade does something like this, internally, and gets a particular tone because its 'more resonant A stage filter) is before distortion, and the softer unresonant filters that compensate for this are after the distortions.

But here's the thing: if you want to make a Console topology where the channels aren't overly brightening, but you're feeding a submix buss that is more crunchy and shouldn't be hit too hard, you can do that too. This is likely going to end up my problem but now you can experiment!

To do that, construct a Console system, using an earlier Console or PurestConsole (the original) and see to it that there are exactly five instances of UltrasonX in each signal path. Maybe two on each channel, two on the submix, one on the 2-buss! Or, one on each channel before hitting the PurestConsoleChannel, one on the 2-buss before the PurestConsoleBuss, and then three more sprinkled between your 2-buss processing. And any of these can be the 'pre-brightening' or 'complementary darkening' ones, meaning you could have the brightness boost after summing for a bit of air in between your instruments, or early on for softening saturation and adding glue.

Or, you could use it anywhere, or on one single channel (that didn't have ultrasonic filtering) as long as you've got room to fit five individual filter plugins. Bottom line is, if your audio goes through ABCDE and comes out the other end, it should be correct. The rest is up to you

download 64 Bit Windows VSTs.zip
download Signed M1/Intel Mac AUs.dmg
download Signed M1/Intel Mac VSTs.dmg
download LinuxVSTs.zip
download Retro 32 Bit Windows VSTs.zip
download Retro PPC/32/64 Mac AUs.zip
download Retro PPC/32/64 Mac VSTs.zip
Mediafire Backup of all downloads
All this is free and open source under the MIT license, brought to you by my Patreon.

[thecontrolcentre]

thank you

[jinxtigr]

I think it may have been here that I first heard from people who weren't able to stack up UltrasonicLite or UltrasonicMed without losing highs, and wanted better. This is what that led to: a better understanding of how Ultrasonic's extended response happened, and a better system to use going forward.

So thank YOU

[kPere]

I just wanna step by to say you're a digital audio legend, and it feels great that such dedicated and skilled people are delivering just because. Thats the kind of world I want to live in.

[jinxtigr]

I just wanna step by to say you're a digital audio legend, and it feels great that such dedicated and skilled people are delivering just because. Thats the kind of world I want to live in.

It's the kind of world I want to live in too, that's why I do it go out and do likewise! There's waaay more and waaay smarter people than me doing likewise. I just have a specific sound to my stuff and will go to any length to get it. But you can go out and do just as well if the sound in your head is a different one. The important thing is the cooperation and being willing to try

[rafa1981]

Strictly speaking. I don't think that implementing a Butterworth as a biquad cascade allows placing non-linear processing between the 2-pole stages and still have an equivalent result.

I also don't think that 5 biquads can be called CPU intensive.

As this is a filter at a fixed frequency (no recalculations at real time) you can probably do the partial fraction expansion of the 10 pole filter in second order sections and to place the 5 stages (2 poles and 1 zero per stage) in parallel, which is good for numerical accuracy and for the CPU pipeline, as there are no data dependencies between the filters. It should run faster.

This partial expansion thing and converting to second order sections can be done e.g. with Matlab.

This assuming that the 10 pole filter has to be IIR of course, which IMO it shouldn't.

[jinxtigr]

But that's by intent: and I can't make it out of second order sections, or in parallel. Won't work. Has to be in series to make the larger, composite filter.

And absolutely you'll have nonlinear processing all through it, that's the whole POINT

And absolutely it has to be IIR. It has to run with no latency and will sound better than filters with pre-ring.

[rafa1981]

You have not answered to what I said, but to a strawman.