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!


TL;DW: Dirty loud!


This one needs little introduction, but it’s been a Mac-only secret so I will introduce it anyhow :)

NC-17 was designed as the loudenator-killer. It uses the same technology as ADClip’s energy redistribution stuff, but on a soft-clipper, and instead of just feeding the energy back in, it uses it to modulate a Chebyshev filter: sort of complicated, but it’s definitely one of those Airwindows things. First to use such a soft clipper as the primary loudness maximizer, and secondly to use such a weird technology after it. Why a Chebyshev? To produce second harmonic. Why do that? To feed deep bass back in despite the loudenating. Okay, so the whole idea is pretty strange.

What happens?

Firstly, the whole tone changes, whether you’re pushing loudness or not. Check that first. You might immediately dislike the result, or if you seek ‘glue’ maybe you’ll like it, or use it to replace some other ‘glue’ effect. If you’re still with NC-17 after checking that, proceed to turn it up (or mix hotter into it). You’ll find no specific ‘break point’: instead, it just gets dirtier and dirtier the harder you push. The whole texture of loudness dynamics is altered, so you get the loudness cues of distortion but with a bass foundation mere distortion won’t permit, and a continuous spectrum between ‘more or less clean’ and ‘impossibly too loud’.

This one will also handle cleaner synthesizer tones, EDM, the kinds of sounds that reveal the artifacts of other loudenators ruthlessly. With NC-17, instead you get a slight ‘grungening’ but then it refuses to break up in the normal sense, just stretches to fit. (this one might be the one you want on drum submixes, too)

There’s nothing quite like it, and now it’s Mac and PC VST, and free.

By free, I mean that it’s supported by Patreon. I’m happy to say that this month we got back up above $700 (it might not look like it but that’s what they sent me, so it counts) so there will be an extra plugin from the list (with total sales under $700). Next week we’ll have the ‘1.5 dB’ version of EveryTrim (aka HermeTrim, for a supporter who asks for stuff I can do :) ) and then I’ll find something nice off the list to include.

The next goal after that is open source (I sincerely hope I can get the LinuxVST working by then) and I expect the Patreon to get there somewhere around the New Year. Help that happen by spreading the word in an informal sort of way, about all this! Keep it appropriate, but remember: I don’t advertise, spam, or bug people, so word of mouth is the only way this works. It doesn’t work nearly as well as bugging people, but it does work in its way, and that way is my way :)

Thanks for being there for me, and I hope you enjoy NC-17!


TL;DW: Left/Right, Mid/Side, and Master in one plugin.


This one’s by request: while I was getting NC-17 ready for its big day, I thought I’d bring out a nice little utility.

EveryTrim is like PurestGain only more so: it is very simply every basic stereo trim you can have (in loudness terms, anyhow!) You get left, right, mid/side, and a master level control. It works on stereo tracks only, as mid/side is meaningless without stereo.

It’s also efficiently coded, suppresses denormal numbers, and uses the same noise shaping to the floating point buss you get in PurestGain. If you need a nice basic gain trim that does all those things (and nothing fancy: Wider is much more sophisticated, and I’ve got an idea for a still more sophisticated stereo-widener plugin that is in the works) then EveryTrim will come in handy. Begone, dull pan-pots! EveryTrim will also be simpler than using EdIsDim and MidSide just to adjust mid/side balances: while you can do that with that pair of plugins, they’re really for doing processing between them using another plugin (any plugin, doesn’t have to be M/S). With EveryTrim, you can tweak mid and side levels directly, in a more obvious way.

If you like getting handy free plugins from me… or if you like the way my plugins support back to Win7 and earlier, and MacOS 10.6.8 and earlier (while still working on current platforms), you should support my Patreon. I’m operating outside the normal plugin market by doing this, but the better it does the more I’ll be capable of, and the more fancy plugins I’ll produce. If I can stay over $700 when next month begins, I’ll pick another plugin from the bottom of the big ‘Kagi Shareware List’ in addition to putting out NC-17. If I get over $800 I’ll begin open sourcing these plugins… and I’m still working on getting the Linux build to work. It’ll be an .so file, and when those become possible I will update every plugin I’ve made VST so far, to be also LinuxVST. (if anyone’s building LinuxVSTs, I’d love some pointers!)


TL;DW: Weird flangey little modulation effect.


Here’s a further experiment along the lines of Chorus and ChorusEnsemble! This one is more in the ‘unique because it’s kind of lame’ category. I like being able to do this sort of thing, because in this 2017 plugin business, everything you do has to be the hippest trendiest most popular thing or you’re basically doomed to get squished like a bug.

But hey! I’m doing a Patreon, not a ‘business’ of selling ‘hit plugins’ that are ‘the best plugins’. And therefore, nothing’s stopping me from putting out something that’s not an emulation of some famous hardware manufacturer’s property (and putting them out of business, eventually). My stuff doesn’t have to be the target market for what people have learned to want the most over the years (often for good reason). Heck, my plugins don’t have to have a reason! And most importantly, my plugins don’t have to succeed. They can exist (and be updated, etc) even if only a few people out there like them… or even if unpopular people like them.

Kind of punk, or something (stay tuned for some major DIY Airwindows stuff coming down the pike along those lines).

So, here’s Ensemble. It’s a weird, unique little sound. It was meant to be a big pad thickener with great richness and depth. Well, you can throw on a bunch of bass, but it’s more like ‘cheesey string ensemble synthesizer from the 70s’, and that by accident, so it’s not even a specific (branded!) string ensemble synthesizer from the 70s. It’s kind of an annoying sound, I think.

It’s free, so if you think you might have use for that, have fun with it! I’m off to make something else :)

ADClip 7

TL;DW: The ultimate Airwindows loudness maximizer.


Finally! The newest version of Airwindows’ famous ADClip is out, and it’s Mac and PC VST for the first time, and it’s free! If you have dense, busy mixes and you want to push their loudness to the max, this is the one you want. And, oddly, if you’re mastering for streaming services or iTunes and don’t want to do a thing besides convey your mix at a set LUFS and intersample peak level, this is still the one you want :)

Here’s how it works.

You’ve got a boost control, a soften control, an enhance control, and a popup or multi-function control with three options: normal, gain matched, and Clip Only. These are all interactive, so I’ll explain them in the context of that ‘mode selector’ popup.

In normal mode, boost makes things louder. This is a clipper, and no more polite method can ever make things louder than a clipper: you need to either give it a busy and complicated mix to mask the clip artifacts, or use it to snip off non-tonal percussive peaks, at which it excels. A clipper does NOT produce ‘limiter-like unvarying block of sound’, and you shouldn’t try to achieve that. A clipper gives you punchy but LOUD, and tries to retain all the dynamics you’re feeding it, rather than smooth them out for a ‘clean’ sound.

The soften control manages the way that high frequencies enter and exit the clips. It algorithmically reshapes the edges of your clip, stopping it from getting digital glare and fizz. This is the heart of ADClip (also present in my simpler clippers, not counting One Corner Clip, which is still upcoming).

The soften control also balances the outputs of two separate energy-fill algorithms, one for bass and one for highs. This was the response to a certain other loudness maximizer that launched proclaiming clippers were dead, and which is still promising its version 2 (and some bugfixes) while ADClip has gone far beyond it. Turned out the secret of that one was an elaborate way of massaging clipped-off loudness back into the signal, in a way that was supposed to be transparent but ‘cracked’ into artifacts when pushed too hard. The algorithms were presumably very sophisticated, which tends to just make the breaking point more obvious when you hit it.

The Airwindows version is a completely different, cruder and more direct version of doing the same thing, so when you slam the heck out of ADClip7, you get a deep bass slam that ‘overhangs’ a bit to add weight, and still fits into the clipped output. And you get the softened, analog-style clips to add ‘heat’ and overtones, but you also get a high-mids reinforcement that normally just highlights bright transients that would otherwise be lost to the clipping. And the Enhance control lets you go between purely ‘analog clip’ energy, and these added reinforcements.

The gain-matched mode has two uses. You could use it to ‘set your slam level’ in a way guaranteed not to trick you into thinking louder is better: it turns the output down, so instead of hearing everything get bigger and louder, you just listen for the point that the clipping’s adversely affecting the sound. Then you can flip back to normal mode, if that’s what you wanted (maximum loudness without blatant grunge). Or, you can ignore the slam and use gain-matched mode as I demonstrate in the video: ADClip7 already suppresses intersample peaks when they’re part of clipping, already reinforces energy lost to clipping, so you can use it in conjunction with a tool like Youlean’s loudness meter to dial in a specific intersample peak level for iTunes or other such picky streaming services. It’ll work like padding the output. and if you’re already in the ballpark LUFS-wise, ADClip is a far more sophisticated tool than just limiting and then padding the output to get to your ‘true peak’ target.

The last mode is Clip Only, and rather than selecting the various algorithm outputs individually, this version of ADClip gives you them at their respective loudnesses, combined. That means you can engage this mode to hear ONLY the clips, and check that you’re not hearing any recognizable ‘scrunch’ of continuing clippage. But since you’re also hearing the enhance outputs, you can adjust softness and enhance level to balance the stuff being introduced to the sound. My recommendation is to set the controls so no one type of artifact predominates: it’s not necessarily great to throw in a bunch of ‘enhance’ bass just because you can. If you’re hearing that much of it that you’re tempted to use it as an effect, you’re definitely also over-slamming your music.

So my recommendation in 2017 heading to 2018, is to use ADClip7 in gain-matched mode, to keep that ‘true peak’ measurement within the Mastered For iTunes requirements. I’m sure not everybody will stick to that, but I’m happy to say it is actually quite good for doing that, and if people want to smash stuff with it and enjoy the bass thud, that’s their affair. Remember a clipper makes stuff dynamic and punchy, not ‘flattened out’: use a limiter if you need dynamically flat, or perhaps both. In this modern era of replay gain, I’m going to suggest that dynamic and punchy is where it’s at. Learn the lessons of the LUFS meter! They’re available to us all, now.

This work is supported by Patreon. I’d like to see many people joining in to keep me going, at a buck or two a month so it’s easy for everyone and predictable for me. I’ll keep on giving you tools to guide you through our ever-evolving music business, and you get to keep my stuff with my blessing: it’s a Patreon, not a subscription. :)

Everything Is All Fixed Forever


Okay, maybe not, but if you’ve ever had an issue with an Airwindows ‘Patreon era’ plugin eating extra CPU…

I’ve developed a thorough denormal-numbers fix that ought to work on any CPU, any DAW, quite intelligently with very low overhead. Some (not all) DAWs needed this: if an audio region ended and the plugins began demanding lots of CPU, that means your DAW was affected. Now, if the audio region ends, you get a -250 db little hissy noise (a variant on HighGlossDither) and it’ll kick in even if you’re already seeing denormal numbers, not just for true digital black, and only one in every channel strip will activate so they aren’t cumulative and won’t build up. Lastly, if you save to 24 bit without dithering, this noise automatically reverts to digital black. Like I said, a nice Airwindows-y denormals fix, a little more sophisticated than it has to be.

And ALL THE PLUGINS are now fixed.


Every single Airwindows plugin that was released as VST (and supported by Patreon) is covered. As of right now, the primary download link for all those old posts is now updated to the new version without CPU mongering. Also, the link at the top left, where you can download ALL the plugins at once, which you might want to do for convenience purposes? All the new versions. (That would be

If for any reason you need to roll back a plugin to the previous version (I can’t see why, but just in case) you can grab all the old builds in I don’t think you’ll want to, but who knows what evil lurks in the hearts of XCode and Visual Studio, so I’m making sure the previous builds are still available.

Small print: there’s two plugins that got skipped, BitShiftGain and DC Voltage, though I did rebuild them along with the others. That’s because they’re each one line of code and don’t even define a variable to put the audio in, and I wanted to have one or two in the simplest possible form so people can see what’s the DSP and what’s the framework, when they get open-sourced. PurestGain fixes denormals. Also small print: ‘everything’ doesn’t mean VST parameter names are longer, because I found mentions of old versions of DAWs crashing and dying if you gave them longer names, and so I didn’t try to force longer names. There’s apparently a sort of laborious XML process that can do it. For now, the plugins will continue to look as they did. Lastly, if you’re on Studio One, make sure the program hasn’t stashed away old versions of the plugins to cling to, as that’s apparently a thing. Plugeat emptor! will give you the complete collection of CPU-fixed plugins, and everything CPU ought to be all fixed forever.


TL;DW: Atmosphere and texture.


Once upon a time there was the blues.

No, let’s be more specific. Once upon a time (and even now!) there was ZZ Top. Brainchild of the Reverend Billy Gibbons, swathed in funk and mystery, serving up juicy grooves from the heart of Texas.

Thing is, Rev. Billy tells some tall tales and their engineer of the day, Terry Manning, he don’t talk ATALL.

So what is a person to do when they hear these albums and the guitars slide off that vinyl like grease off a hot griddle, and you know you can’t just put up a mic on anything amp-shaped and get near that magic? You know those are dirt guitars, but the whole texture’s different. Skulduggery is afoot. And the boys ain’t tellin’.

Well, here’s what I did. It seemed to me that some of the mojo sounded like echoes and delays, but not just any old ones. You can take something as small as a dentist’s mirror, put it near the mic, and aim it until you’re reflecting another copy of your sound into the mic again: the delay is tiny but real, and the tone? Well, that’s based on how big the panel (or dentist’s mirror) is. If it’s tiny, you get only highs. If it’s a big ol’ panel, or a floor or wall, you get down into maybe the lowest bass. Any panel will do this. Billy and Terry might have been constructing lil’ forts around the amps, making a purely acoustic home for the blues. You can literally pick what range of sound you reflect, how long a delay it is (still so tiny it’s not heard as one!) and you don’t have to make it full-range: a softer reflector ignores highs, smaller panels ignore lows. If you want to juice up what your mic hears, this is one way to do it.

If you’re playing with super-short echoes, you’re reinforcing the lows. Unless it’s out of phase, flipped upside down in the DAW, in which case you’re cancelling them! And then, supposing you have one delay that’s in phase and one that’s out, and you calibrate them just right, and then you’re neither reinforcing or cancelling the lows, instead you’re just thickening the texture of whatever you’ve got… all the little detail doubled, tripled, dripping down the mix, but the body of the thing basically the same and no sustain, just a couple of delay taps in real close…

I’m not Billy and Terry. Since I’m Chris, I’ll fess up: that’s exactly what I did, and you can have it in Hombre. It’s two calibrated delay taps, which you can tweak a little, and if you bring them in you’ll thicken and diffuse your tones without altering where the lows sit, or adding much in the way of extra sustain. It’ll be punchy and get out of the way like reverb won’t, but it’ll be fatter and juicier than the dry signal. This is my interpretation of the ZZ Top secret sauce, or at least one of ’em, implemented in software rather than acoustics.

I’ll never know how close I came, because them Texas boys don’t tell tales out of school. But Hombre is my humble offering for a simple plugin that brings a little mojo to what would otherwise be a dry voice or guitar… and it won’t muddy things up, just grease ’em a little.

If you like me being out there thinking up ideas like this and taking on the great mysteries of the audio world, please support my Patreon, just a dollar or two per person so it doesn’t get too much like riverboat gambling and high rolling. I’ll keep on being a thinkin’ fool, and putting out cute little tricks like this one. Hope you find it handy: it might be the easiest way to throw in two tight quick echoes, one in phase and one out, because I’m not aware of anyone else facilitatin’ specifically that. Well, now there is!

Thankee. (chrisj will become un-Texan in three, two, one…)


TL;DW: A more complex, multi-tap chorus.


Here we can fill out the Airwindows palette of modulation plugins a bit… like Chorus, this is using my special slightly dark interpolation with a little pre-sparkle to get an adaptable, rich chorusing effect. But ChorusEnsemble uses a bank of chorus taps to get a more complex, textured sound that’s farther from the original. You can set it wrongly, so don’t assume all the settings are appropriate: that said, a little care should give you nice lush chorusing that’ll work great on pads and backgrounds. The reason I allow for the ‘ugly’ settings is, who’s to say you might not have a use for them, and if you find that use you’ll have a tonal element that other people don’t have on tap (generally, it’s so hard to sell plugins that can sound wrong and broken that people will tend to shun that and limit you to ‘nice’).

Whether you like setting ChorusEnsemble ‘nice’ or ‘naughty’ (‘nutty’?) I hope you enjoy it. I’m making strides on fixing the denormalization bug some plugins have on some DAWs, and I’ll be posting about that as well. This work is supported through Patreon, and not through charging you directly for the plugin (or holding ’em hostage and taking ’em away again if you don’t pay). If you like seeing people act the way I do, the only way to really encourage it in this world of commercial plugins is to throw money, which makes it a more interesting story to hear about. The high-earning Patreons are the ones that get attention in a sort of feedback loop, which those of us who are guitar players should be familiar with. Both those kinds of feedback loops are desirable and delightful :)


TL;DW: Classic 2-buss compressor.


Here’s another one of the bestsellers. Ironically, version 4 is way better than any of the previous versions, and free! (Patreon supported)

Logical’s a compressor. It was designed from the start to work on the 2-buss in the most demanding conditions: people are really picky about their 2-buss compressors, and you can’t mess around. The tone has to be spot-on and it’s got to be transparent and able to let the music through. Additionally, when we’re talking about ‘tone’ and something called ‘Logical’ you can see that it’s going to be in the SSL style: there’s a sparkliness which requires some extra coding attention.

You can approach compression duties from several directions with Logical. It has three distinct stages, and will entirely bypass stages it’s not using. It’ll go from 1/1 compression, up to 2/1 using just a single stage (for the utmost transparency): keep it below 2/1 ratio and use the threshold control to bring in the compression. This is a traditional 2-buss natural-sounding compression. From 2/1 to 4/1 ratio, you can get various behaviors and the two stages in use still sound very clean: the speed control will give you different kinds of ‘swing’ and spring-back out of the compression.

Then as you pass 4/1 ratio and go off to a max of 16/1 (approximate, but that’s the basic idea) there’s a tone change, and as you get into crazy high ratios, Logical goes a little bonkers. This was NOT available in previous Logicals. The issue was, if you rely heavily on that final compression stage, things can get messy. You can push Logical until it’s nasty and so full of energy it’s forcing you to use the makeup gain to PAD the output, just to handle all the madness.

This time, and in honor of Logical going free VST format, it’s not set up for only good behavior. This time, it’s your responsibility to not blow up your outputs by thoughtlessly cranking the ratio. Consider it an audio chainsaw made of silk and glorious victories. Not every top-selling plugin got this much better when I revisited it. I’m very pleased with how Logical4 came out, and I hope to see it talked about a whole bunch. This one’s worth a lot of ‘did you hear?’.

Also, in the video, you get to see my reaction to a classic silly mistake: bypassing it and then thinking I’d unbypassed it again. It was actually only about forty-five seconds of earnestly explaining the nothing that I was accomplishing, but I saved you the Moment Of Realization so you can enjoy my discomfiture. My excuse: firstly, it’s funny and we’ve all done that, and secondly, I’ve just released Logical 4 free for AU and Mac and PC VST. So I think I can be forgiven a little foolishness with the bypass button :)

Please help my Patreon grow in numbers: more than individual high pledges, I’d like to see lots of people discover what I’m doing. If there are lots of people all of whom are pledging only a small, affordable amount each month to keep me doing this work, it brings me stability and lets me do stuff I care about. You’ll be seeing more neat things out of me, and I remain at the level where I’m doing the top-seller plugins one a month, and if I reach $700 again (from smaller individual patrons this time) I’ll be picking stuff off the bottom of that list too. We did get to StarChild, so don’t underestimate the coolness of the ‘less top selling’ plugins that were part of my Kagi storefront. :)


TL;DW: Mono chorus, also works as a vibrato.


Here’s the start of some modulation plugins: Chorus will give you a nice basic mono chorusing effect, sweeping one moving delay tap against the dry signal. What makes it unusual is a pile of odd Airwindows things to adjust it this way and that.

The sweep is done with a peculiar Airwindows interpolation which both applies a bit of an averaging function, and also a subtle lift to highs to compensate for that. The idea is to have the moving part be totally fluid, analog-like, even though it’s digital.

To drive that, there’s a treble boost going into this smoothing delay tap, and it’s based on the Airwindows Energy algorithms: that means this plugin interacts with the raw sample rate in two different ways. It’s designed to make it so, whatever the sample rate, the sounds project through to the most extreme highs without edge or dullness. That means Chorus won’t sound exactly the same at 44.1K, 96K and so on: it will try to deliver the most finely tuned treble for each sample rate, whatever that is, and the tonality might be different. Also, since it’s running a delay buffer, it’s giving you full use of the buffer at all times: the broader settings might be more useful at high sample rates.

In general, this should be pretty approachable. Hope you like the tone: there’s more where that came from. The continuing process of bringing these out to you, is supported by Patreon which makes all this possible :)


TL;DW: Unpredictable distorty noise!


And then sometimes there’s a plugin that just makes you go ‘wut’…

DustBunny was an accident. I was doing something and put out a plugin (possibly a freebie) and there was a bug and I was in a hurry and put it out without checking… and quickly learned something horrible was wrong. Initial reports were along the lines of ‘oops’ and ‘yikes’, and when I checked, sure enough, the plugin erupted in terrible scrunch, and I in turn erupted in apologies and scrambled to fix the problem (which wasn’t that hard).

But even before I’d got the fix out (and simple oversight problems, I sometimes fix within hours), more reports were coming in: hold on, don’t fix it, it’s cool!

That wasn’t what I had in mind, but the solution was obvious: DustBunny was born. This twisted little plugin just applies a weird accidental scrunch based on where the bunny control is set. Please don’t use it on the 2-buss, or in mastering :) but more seriously, this is born to sit inside some kind of weird plugin matrix device as part of a nefarious sound design idea. It would’ve been perfect in one of the parallel effect chains used on the latest DOOM soundtrack, mangling a sinewave. If you hit it with high levels it gets kind of jumpy, so you might want to give it more restrained levels. If you don’t like running a gain trim in front of it, run something more amusing like a delay or a flanger to pad your signal a touch.

DustBunny is kind of like a joke, except for it’s real and does produce an unusual, distinctive effect (or 1000, as all the settings are kind of unique). I’m still here to post things like this because of my Patreon, which has been slowly but steadily growing since last year. The more the Patreon grows, the more cool things I can do and bring you the happy results.

Or, sometimes, just bring you odd little presents like DustBunny. Hope you like it! Next comes my mono Chorus, and then it’s Logical, and then one of the ‘unsung’ plugins, which doesn’t exist anywhere else and sold very few copies in its day. But that would be telling :)

Older Posts


handsewn bespoke digital audio

Kinds Of Things