Menu Sidebar


Airwindows is one guy: me! I am an audio hacker and computer programmer from way back. I seek only to continue my life up here in Vermont, inventing things and putting them onto the internet, sometimes for free and sometimes for pay. My hope is that people richer than me (i.e. most people) don't rob me. My other hope is for another cup of Aeropress. One out of two ain't bad!

Kit B

Kit_B is an EXS24 sampler instrument, a heavy, ambient, darker kit with tons of multisampling and a very quiet background. It’s a little more open and ringy than Kit A. Check out the video for more!

If this helps you, support Airwindows 32 Bar And Grill by paying $10 with the blue button below. If not, delete it. If you want others to try it, don’t torrent the stuff, send people here!

Kit A

Kit_A is an EXS24 sampler instrument, a heavy, ambient, darker kit with tons of multisampling and a very quiet background. It’s a little cleaner and tighter than Kit B. Check out the video for more!

If this helps you, support Airwindows 32 Bar And Grill by paying $10 with the blue button below. If not, delete it. If you want others to try it, don’t torrent the stuff, send people here!

Righteous 2

Righteous2Demo is a universal binary Audio Unit plugin for the 2-buss.

It can be used on Mac DAWs like Logic and Ableton Live, Reaper and Digital Performer. It can be used in Pro Tools if you use the wrapper plugin Patchwork from Blue Cat Audio (79$).

And it is the first and only plugin that can do loudness maximization on YouTube AFTER they put in their ‘replay gain’ function.

And it does it perfectly honest and legit, too, by delivering what Google wants: but optimized.

YouTube is not going to stop turning down your music, by the way. They want to make people use auto-play and your loud masters get in the way. And they’re Google: they don’t have to listen to you, or to VEVO, or to anybody. It’s a level playing field and all the rules have abruptly changed.

Here’s how you cheat (except it’s not cheating: you’ll find some older classic tracks are coming across like gangbusters in the new YouTube regime)

Righteous was originally designed to force people to mix dynamically, at a lower target level than YouTube’s -13db RMS. (it’s now adjustable, so if they change it you can retarget! They won’t, though: their -13db parallels the loudness of full bandwidth pink noise, it’s kind of scientific as the loudest possible full-range dynamic sound)

Righteous also is designed as a final output stage for Pono, so it has an Airwindows fancy wordlength reducer set to 24 bit for such use (or any 24 bit export). It uses Naturalize out of Ditherbox, which is a $50 value already. It also includes a simplified version of Airwindows ADClip 4, without the techniques for sneaking clipped bass and treble energy back into the sound, and without the fancy bypassing when not actively clipping. That’s okay, all we really need is to soften the onsets and exits of clipping energy (most obvious at 44.1 or 48K)

That’s because Righteous, used as a loudness maximizer for YouTube, is set up to deliver full range clipped intense peaks on the hottest sounds. The key is to not try to limit or squish those ahead of time. Mix old school, allowing dynamics, make it big.

Then, once the framework’s in place and you’re bringing in the body and orchestration and lushness, be sure and set Righteous 2 at the correct target RMS level. It comes set for YouTube or you can set it for other systems if you know their target RMS. At this point you’re making decisions about how loud to make the fullness of the track.

If you hear Righteous 2 distorting, YouTube is already preparing to turn you down. Back off.

It’s that simple.

But it’s more than just an indicator. Because Righteous 2 applies its own saturation process (that dynamic peaks can punch right through to clip: true even for the original Righteous) it approaches YouTube’s target RMS value differently. It uses the energy lost from saturation to drive a warmth and bass shaper that gives even more fullness to the sound in a way that sneaks past the YouTube normalizers by being not as high-RMS as before… and because it’s the gentlest possible saturation, it broadens the ‘sweet spot’ for loudness a huge amount.

You can listen to the result in the youtube video I linked. There’s almost no audio except for the music, and that music is absolutely slammin’ and extremely competitive volume compared to what the normalizing is doing to modern high loudness masters.

Welcome to the future. By the way, just because you CAN use Righteous 2 to maximize YouTube loudness (no matter what they change: just target the RMS level they want you to hit) doesn’t mean you should. -13db is by no means obnoxiously overloud, but you can still allow for more open and dynamic sounds. You might like them! -13db is probably your best bet for not having your balances altered by YouTube, though.

Righteous 2 is $50.


PurestWarmDemo is a universal binary Audio Unit plugin, and another really extreme example of the Airwindows ‘Purest’ series of plugins. To explain why, I need to explain what it does. It’s for individual tracks, and I do NOT recommend using it on a full mix, even though like PurestDrive it is ‘invisible’ and at its best being non-obvious.

PurestWarm is simply the highest resolution asymmetrical saturation in existence. See the video: it’s best used on things like basses, where you might have a type of sound that peaks harder in one polarity than the other. When you put PurestWarm on a track, you pick which polarity you want to distort, and it applies the single most soft-textured distortion in existence, something that in other Airwindows plugins is responsible for adding huge fatness and boosts (try Density for a freebie example!) but in PurestWarm is used in the simplest possible form, only to restrict loud peaks. It’s not unlike PurestDrive (the experiment that started this line of inquiry) but there’s an implementation detail that’s at the heart of what this plugin is about.

When the saturation is applied, it’s done at 80 bit floating point resolution, and when we go back to 32 bit for the buss, there’s a form of dither (really noise shaping) to translate the higher fidelity 80 bit signal back to CoreAudio 32 bit. This is fundamental to how the Purest series plugins get their totally transparent sound, but it’s only relevant when you’re changing the audio, be it ever so slightly.

So, the dither only works on one polarity of the signal. For the ‘clean’ polarity, instead PurestWarm takes literally the input data word, and by that I mean the exact variable holding the input data, and passes it through to the output. Doesn’t even assign it to another variable, much less ‘multiply it by 1.0′ and call that the same thing (in floating point, that’s not always true if your 1.0 and your audio data are at different floating point scaling factors). PurestWarm literally goes into bypass and is not there at all, for one polarity of the output. For the other, it’s doing that ultra-high-quality saturation and noise shaping, at 80 bit.

I’m really happy with how this one sounds, even though I think it shouldn’t be used on full mixes. Too much warming! Keep it to pads and basses, though I’m not the boss of you. If you DID want to whack one whole polarity of your mix in the name of warmth, there is no more transparent way to do it, anywhere in digital or analog. And if your mix is coming out with all its transient spikes on one polarity, it might even make a kind of sense (though I still think it best to address that at the individual track level).

It’s a permanent fixture on my electric basses now: because it can be. Nothing about it hurts the sound, it just throws a whole bunch of warmth and makes the instrument more manageable and easier to mix. A true secret weapon, that will never sound like you put on an ‘effect’.

PurestWarm is $50.

ElectroHat (revised!)

ElectroHatDemo is a universal binary Audio Unit plugin on a mission: it intends to take the place of samples and vintage drum machines, turning any old rhythmic sound into a crisp synthetic and very electronic hi-hat! It’s been revised to work more elegantly with 88.2 and 96K sessions, and the tone’s been adjusted so it doesn’t keep on getting louder right up to the Nyquist frequency. Instead, it gives a warmer tone that’s still insanely bright and crisp, but acts like a hardware beatbox rather than a softsynth.

It will also do a decent snare if you like bright 606-style snares (look into Noise if you’re interested in something darker). The real heart of ElectroHat, however, is the hi-hats. This plugin gives you a huge variety (literally, 3000) of synthetic hat sounds each of which are adjustable for extra brightness. Pick between Syn Hat, Electro Hat, and Dense Hat and use the Trim to dial in the sound—pay attention to the nodes where the algorithm gets weird, as those can produce unique percussive sounds too—and use Brighten to control just how trebly the hat will be. Output Pad starts you off with a nicely attenuated volume. Don’t crank your hat too hard, it’s got a ton of sweet top end that should be treated with respect and can cut through any mix without getting in the way!

If you have trouble getting a track or software instrument trigger to work with ElectroHat, remember that you want to be feeding it the volume envelope of a hi-hat, as a source sound. For soft-attacked things you may want to increase the definition of the stick attack. Do that by using the free DigitalBlack2 plugin to gate your guide sound (route a sound to an aux to process it without altering the original tone if it’s audible in your mix) and then use the free Point transient designer plugin to add ‘pop’ to the attack. The video will show you how to do that, on even a muted synth bass playing psytrance 16ths! If you don’t like the way ElectroHat extends right up into the supersonic range, you can use the free Slew2 plugin to whack off the very top to your taste. This one takes amazing advantage of the broad range of Airwindows freebies to extend its usefulness.

Why trigger from audio tracks, rather than sequence stuff in DAW MIDI and softsynths? Because if your audio track is tighter than DAW MIDI because it’s recorded off a synth or drumkit using hardware sequencing (an analog drum machine, a vintage 808, an Atari computer, my Kawai Q-80 sequencer) then ElectroHat will be triggered off that tighter-than-DAW recording. DAWs hiccup at times, don’t always trigger stuff as perfectly as we’d like. ElectroHat can hit with sub-sample accuracy if the underlying track has that. There’s nothing else that can produce an ‘organic’, evolving hihat tone that is also totally electronic-sounding for modern music and tracks the groove that tightly. You can use this capacity to ‘sequence’ all kinds of things using just delays and echos from a simple guide track, all with hardware-sequencer tightness—the video shows you that, too.

ElectroHat can be the new go-to electronic sizzle! Its tones have huge variation within a range of crisp and bright synthetic tonalities with more character and smoothness than simple noise, it sounds a world apart from even round-robined sample triggering, and best of all it can groove like hardware because it literally inherits the exact timing of whatever you’re driving it with! Whether that’s recordings of a classic groovebox, or just you slapping your pant leg and micing it, the groove is perfect: and nothing is more important.

ElectroHat is $50.

Airwindows On (Mac) Pro Tools!

Every now and then you run across something that tells you: hey! You’re not alone, there are others who see things as you do!

For years now, I’ve been developing Audio Units, just the audio code. That way they always work and won’t break, and I can focus my attention on the sound. I always figured that someday, if people wanted it, someone (perhaps a DAW manufacturer) would work out how to skin the things for those who like knobs. Didn’t concern me, so I kept on caring only about the sound (and about improving things for my customers, getting them free updates, all kinds of stuff like that).

I was a little concerned about the way that Audio Units only worked in Logic… and Digital Performer… and Ableton Live… and Reaper… well, you get the idea. But there were two glaring omissions, Cubase and everyone’s favorite nemesis Pro Tools. And I am not the guy to make and maintain ports to platforms I can’t afford to develop for and don’t really understand. It’d be a recipe for just struggling and stopping, and that wasn’t okay.

Well, I’ve got a little announcement. It comes in two parts. First… Read More


NonlinearSpaceDemo is a universal binary Audio Unit plugin for algorithmic reverb. It extends on what its predecessor Space offered, by adding Nonlin behavior in two ways!

You’ve got a Sample Rate control that sets the overall size of the reverb tank. This isn’t meant to be switched while the reverb is live, so if there’s a problem bypass it and turn it back on: it should be OK.

There are Treble and Bass controls, which are not simply EQ on the output of the reverb: they tie into the guts of the reverb shaping the sound as it continues, so you can get many different reverb tonalities. The Treble is particularly good at setting reverb sounds into the distance, but even at full crank NonlinearSpace is very deep: compare it with other algorithmic reverbs. It has a character all its own, and now…

With the Nonlin control, you can do dial-an-eighties, or crazy unique things! This is a behavior normally found in isolated reverb algorithms, but I found ways to put it on a continuum. If you keep the liveness very low and ADD Nonlin, you get an Airwindows version of the classic gate-y Nonlin reverb sound, ideal for drums and keeping a mix from getting muddy. If you control your input sounds well you can do outrageous things with this! It seems particularly good at tom fills, and all the sounds are shapeable with the Sample Rate option and the Treble and Bass controls. This can be used on synthetic elements, too, not just live instruments!

Then, if you tire of that, try the reverse! Run a sporadic sound into a channel with NonlinearSpace, set Nonlin to -1.0 and crank up the Liveness, crank everything up until you have infinite sustain. What’ll happen is, the extreme Nonlin negative setting will force the output to not get too loud, but if you feed new audio in you’ll replace what was there. This is the ultimate tool for making eternal reverbspaces and ambient washes: if you’re using something like a volume pedal, you can literally paint in new sound as you go, balancing or replacing it on the fly, never getting runaway feedback. And since it’s a continuous control, every setting can be given a touch of nonlinearity just like you’re dialing in the reverb tone, helping you shape the spaces of the mix.

NonlinearSpace, like Space before it, wants to be on a stereo track, the better to fill up the atmosphere behind your mix. If you’re not finding it in the menus, check that you’ve got a stereo track for it to go on.

This is all on top of what was already a damn good sounding algo reverb anyway. Space was already an innovative design with many special Airwindows tricks in things like its comb-filters and allpasses, but NonlinearSpace takes it to a whole new level—and brings everything to the 80-bit processing of the Purest line of Airwindows plugins, which is great for reverb tones that are all made up of subtleties. NonlinearSpace is huge fun, and hugely effective, and I’m just loving it: I hope you do, too.

NonlinearSpace is $50, and people who owned Space get updated to NonlinearSpace for free :)


PurestSquishDemo is a compressor Audio Unit plugin for maximum transparency and organic squishiness! I’ve designed it for the 2-buss, but it’ll work for many things. It’s got two basic modes: you can use it very gently to ‘float’ the mix elements in a barely perceptible way, or you can lean into it harder for more textural effects.

PurestSquish is one of the Purest line of plugins, which means certain things. It doesn’t emulate any known hardware device, it just brings the squishiness and fluidity to the sound while trying very hard not to touch the tone color in ANY other way. These are transparent plugins, be careful about over-applying it just to hear an effect. It also means PurestSquish is implemented using 80-bit math internally, with carefully designed algorithms to apply the effect with as few math operations as possible. This is somewhat controversial as the noise floor of these ‘superfluous math operations’ I’m avoiding is very low, but typical plugins waste a lot of CPU overprocessing. The Purest line is about going the opposite direction, as hard as possible.

Finally, one of the things PurestSquish has is a highpass on the sense circuit for bass. Using this, you can bring hugeness to the lows of the mix. Since it’s just on the sense it’s not in the audio path, but because the compressor ‘turns down’ what’s a gain boost going in, the Bass Bloom control can be used like it was a boost: apply the compressor, and then work in enough Bass Bloom to have a giant low end. This, plus the way PurestSquish preserves textures going through it, makes it a fantastic compressor for cranking up the size and scale of electronic music—or any sort of music, really! But the highpass Bass Bloom stuff comes from EDM.

Shoot this out against any compressor, from Airwindows or otherwise. It really has its own sound, it really is exaggerated as far as squishiness and transparency are concerned, and there is a real usefulness to applying just only the compressedness to a mix. You can get your color or ‘classic tone qualities’ in lots of other ways. PurestSquish is distilled squishiness, concentrated to keep the flavors of your mix pure while it changes the texture.

PurestSquish is $50.


DeEss3Demo is the latest DeEss: returning to the unrivalled sound of the first classic Airwindows DeEss, but with 80-bit internal math and a new Sense Monitoring switch, by request! This is the ultimate De-Esser, now with new ease of use and deeper tone.

This plugin’s made a lot of interesting friends… very competitive, significant friends who seized onto the original DeEss as a secret weapon. It seems that if you really knew what you were doing with the filters and such, you could get the thing to tame sibilance with incredibly little damage to the audio. It took expertise, because there was no metering or monitoring then: you had to be able to hear supersonics and you needed the experience to know what you were aiming for.

The second version tried to simplify the effect and make it more approachable, more a one-size-fits-all effect. This fizzled. Less experienced people weren’t reassured, and the power users weren’t excited about losing functionality, even sticking with the old version (which was updated to 64-bit to ensure everyone could keep it working if they wanted to).

Live and learn! DeEss3 is EXACTLY the first DeEss, right down to some of the quirky behaviors that gave it its tone, but with an 80-bit buss and noise shaping to the 32-bit output word. It actually is better than the first DeEss, in the only way it can be: more accurately reflecting that first algorithm, at higher fidelity. And it’s got a new control that makes it WAY easier for both pros and newbies: use the sense monitor control to adjust your settings, so you can hear exactly what it’s taking out, against a backdrop of impressive silence. That by itself is a powerful argument for just how good DeEss3 is: you can hear how little it touches the rest of the audio.

The video will show you how to work DeEss, so this secret-weapon plugin can finally be less of a mystery. Normal people can use it effectively now!

DeEss3 is $50.


PurestAirDemo is a universal binary Audio Unit plugin for enhancing treble up near the Nyquist limit! It specifically tweaks that sonic area, wherever it is, with special Airwindows algorithms. The idea is to get ‘air band’ boost without the coloration of a normal equalizer, and in line with the Purest style of plugins it’s done with a really simple and elegant algorithm.

There’s also an ‘Air Limit’ control, and what this does is kick the previous control down if the energy gets too high. It’s not properly a de-esser by itself, but it does work to moderate the brightness effect. You can use it without air-band boost, to function as a tone softener.

PurestAir is really clean and maintains the tone quality of what’s passing through it, adding little color of its own. Part of the secret to that is the way it operates relative to the Nyquist limit. What that means is, PurestAir is a different plugin depending on what sample rate you use. If you’re at 44.1K it’ll have the most obvious effect as it’ll be working at around 22K. If you’re at 48K, it’ll center on 24K and so on. Still functional at 96K (with a boost at 48K but some effect in the audio band), probably no longer useful on 192K. Be aware that this isn’t a normal EQ algorithm and is dependent on your sample rate.

PurestAir is $50.

Older Posts


handsewn bespoke digital audio

Kinds Of Things

Shopping Cart Button