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!

Pop Filter

The most amazing pop filter you maybe already had without knowing it!

This video isn’t really a ‘product’, though you can still support me finding stuff like this out through my Patreon.

Instructions for building an Airwindows SM57 mod can be found at the SM57 Mod page.


TL;DW: Super old school compressor for high definition transients. Adds no fatness, just energy.


Pyewacket is a strange beast. It’s inspired by how much I love the 60s/70s recordings out of London’s Olympic Studios, which had and used Pye compressors on many of my favorite classic and prog-rock records. Once you recognize the sound, nothing else will do: the musical event is delineated with hallucinatory intensity.

Mind you, for ten or twenty THOUSAND dollars it had damn well better hallucinate musical events on command: these are not compressors normal people can have, not anymore.

However, I’m ‘chris from airwindows’, so for me it’s not just a matter of mimicking the faceplate or even the specific behaviors of the device. I want something more original, that can get the essence of that electrifying sound. I might not play like a musical hero, but I want a compressor that can deliver that crackling voltage. And as I was listening to examples of a homebrew Pye replica, it suddenly hit me: I know how to make a compressor cut back just the body of the sound, leaving that energy and transient definition. I can also bring in the ‘brickwall filter’ behavior the Pyes have, as needed. And I have a whole life of devoted music listening off classic vinyl records to guide me. I can get the sound.

Introducing Pyewacket. Pyewacket is my compressor familiar. It may or may not have dark magic, but what it does have (demonstrated at the end of my video) is a response and tonality like no compressor you’ve heard. I can contrast it with Pressure4, and have done: where a more ’round and thick’ comp like Pressure4 brings stuff forward, Pyewacket’s soundstage sits back and the energy comes forward, from the highest treble to deep hard-kicking bass, producing a ‘retro’ sound where peak energy absolutely blows away the more thick, tubby RMS loudness. This is a compressor for a new era. We’ve been doing ‘loud and fat’ for decades now, and the loudness war is on its last legs, with automatic playback gain controls rendering it useless. You don’t have to be composing retro to use this. The only requirement is energy and information: whether as a 2-buss comp or to condition individual tracks, Pyewacket brings focus and intensity, and an incredibly clear and articulate attack transient where most compressors mangle and transform the attack beyond recognition.

And if you try really hard, yes you can kinda-sorta make it do that ‘Hole In My Shoe’ gratituous pumping thing. Rest assured, though: you probably shouldn’t.

Other people can’t do this plugin. You can’t market it in normal ways because it doesn’t do ‘BIG PHAT THICK PHWOAARRR’, you don’t switch it in and have all the music leap forward and become much bigger and in fact it might make things smaller, and an inexperienced kid with softsynths and Apple Loops might think something was amiss and be extremely uninspired. And anyone trying to tie it to the twenty thousand dollar unattainable hardware compressor would be compelled to model every little detail of the very complex and twitchy hardware unit, and that would cause that plugin to be overprocessed and it’d lose most of what made it special.

But Pyewacket is important, because it’s the sort of thing I can do when supported by Patreon. I don’t have to restrict myself to what’s going to sell to blind market forces. I can make it the essence of how Airwindows would do this sound, and I have done. As such, it is free in AU, Mac and PC VST form. If I’m poorer than you (go check on the Patreon and see, I get paid monthly) then it might be worth your while to chip in a buck a month (or more if you like).

I really, really, really like this one, and maybe you will too :)


TL;DW: Smoothest saturation or antisaturation, plus highpass.


This one started a lot! The algorithm used here has echoed through many other Airwindows plugins. It’s literally the smoothest saturation you can have in a plugin: the transfer function’s a sine. This is what’s in Channel, too: there are many ways to adapt such a simple mathematical function.

But there’s more! Because Density runs multiple stages, allowing it to bulk up the tone into an overblown, insanely fat and saturated distort-fest. And then you can highpass just the distorted stuff alone, and trim its output gain, and mix it with the unfiltered dry to produce lots of tonal possibilities. And then there’s the spatial positioning factor: saturating stuff this way brings it forward in the mix. You can also isolate midrangey elements and bring them forward using that trick.

And then there’s the negative values: if you UNsaturate, you get a thinned out lean tone and it drops back instead of pushing forward. And you can blend that too.

Density’s one of the better utility plugins. It’s there to reshape tones in myriad ways, mostly having to do with fatness or thinness, also having to do with upfrontness or recedingness. It can also give articulation to sounds that are murky, or simply produce the hugest fattest roaring wall of grunge you ever heard.

This sort of thing is supported by my Patreon, so if this or the many other plugins I’ve produced prove useful to you, please chip in. I’ve got a lot more work to do before my plugin library is ported, and even when that’s done, I’ll still be creating new plugins, with your support :)

Lowpass, Highpass

TL;DW: Lowpass gives rich textures, Highpass timewarps your tone.

Lowpass / Highpass

These are mixing EQs, not mastering ones (though I’m not the boss of you). They’re complementary: the one is the inverse of the other. However, because of their peculiarities that makes them behave quite differently. What they have in common is they’re interleaved IIR filters, something people don’t normally do. The experiment here has to do with my discovery that digital audio only exists in sets of samples (never just as the isolated sample: the waveform isn’t there, the sample value is only a signpost that the audio is to weave its way around)

They’ve also got a very unusual parameter, soft/hard or loose/tight, which controls how the IIR filters are fed audio. When you offset it, you get a situation where the cutoff is higher at louder volumes, or at quieter volumes. This is on a sample-by-sample basis so it’s a tone-character modification, subtle but interesting. Loose/tight is just the best way I could describe what’s happening there.

Lowpass gives you a treble rolloff (some have joked that I make dozens of treble rolloffs! Yes, but they all sound different) and what’s immediately obvious is, the stuff right up by Nyquist on the threshold of hearing is not rolled off with the rest of the treble. Also, if you only want to cut extreme treble, you can do it with just adjusting the soft/hard control away from the center position. At deeper cut settings, the soft/hard control gives you two different textures (both of which keep a hint of ‘air’ right up top). The dry/wet control allows you to blend your result. Lowpass gives you big sounds with various colorations and a sparkly gloss that comes from your underlying sound: it’s a big-ifying filter that might suit huge synth pads or orchestral tracks.

Highpass, the inverse of this, gives totally different impressions. The same filter-offset behavior turns into ‘loose/tight’ and the extreme treble gets stepped on, rather than retained. This makes Highpass take on ‘classic’ tonalities, particularly with the offset on ‘loose’, which gives a tubey and softened texture. If you run it full-wet, you’ll get a radical ‘analogification’, wiping out all extreme lows and the highest highs, and sounding like some small vintage radio at high filter settings. It’s a small-ifying filter that’s also a time warp (with offset on ‘tight’, you have a transistor radio instead, still retro-sounding!) and all you have to do is dial in your boost area and then balance it with dry/wet to get intense texture shaping that normal EQs can’t come close to delivering.

Again, these are not mastering EQs unless you face really unusual mastering requirements. They’re mixing tools, and they really do act like different animals so they’re each contained in their own plug. They’ve been around for ages but the revision to VST form has brought them a new level of tonal sophistication plus the very useful dry/wet controls that take them out of ‘experiment-land’ (they have always been building blocks for plugins such as Guitar Conditioner) and makes them stand alone as useful mix tools.

Lowpass for buttering up your textures and making them glossier and deeper, and Highpass for giving tracks that retro analog voicing… and Patreon because (not unlike the DAW Reaper) I’m giving you stuff very freely and trusting that it will earn me a place in people’s sound engineering worlds. I began Lowpass and Highpass in 2007, inventing the basic concepts (and still no other EQ does anything like it). In 2016, returning to them, I’ve made them even more useful. Imagine what I’ll do in 2026, or what I could do with a serious income from the Patreon. I think I can earn my keep :)


TL;DW: Mastering grade very low Q tilt EQ.


This is a new plugin, not a port from an existing Audio Unit. It’s based on a variation on the Average concept. Turns out, it’s the tail end of the ‘averaged samples block’ that causes the cancellation node. ToneSlant implements a much bigger sample block (100 taps) but linearly fades the samples off between the first and last sample in the block.

What that does, is produce an extremely transparent ’tilt EQ’ with a controllable corner point. And, it’s implemented in such a way that you can set it to null out (at extreme high boosts) and then bring in only the brightest highs (shown in the video). Quirky, maybe, but it broadens the ToneSlant toolkit. So, the main uses are:

-Extremely low Q treble rolloff above a set point (with a fixed dB/oct)

-That, plus you bring in the dry signal by not putting the Highs to an extreme (it’s like a positive/negative wet/dry/wet, but that doesn’t fit on the label, especially for VST)

-Total cancellation with Taps at 1 and Highs at +1, and then you put Taps to just barely greater than 1 and you can have a very natural high-shelf controllable with the Highs control

These come out of the algorithm: the reason it performs so well sonically is both the extremely low Q and the simplicity of the algorithm. Not everything I make belongs in a mastering studio, but ToneSlant is peculiarly suited to that use: it’s like those specialty EQs that have very few parts and impart no color to the sound. You should be careful applying ToneSlant, because it ‘hides’ and tries not to be apparent as EQ. Use it as such, when you need perhaps a strong ’tilt’ but don’t want the result to sound equalized.

It’s AU, Mac and PC VST and free: just another reason to keep me around through supporting my Patreon! If you look at the last decade or so of Airwindows, I think it’s fair to say I think of new plugins, and that will continue. It’s just that now, you support Airwindows because you want me to make more, not because I’m preventing you from having it. I don’t think there is a future in preventing people from having digital things. It seems impractical and sure to cause problems. Therefore, my future is ‘be worth keeping around’, so enjoy ToneSlant as a token of my good will :)


TL;DW: A lowpass filter you probably don’t have in your DAW, with a distinctive tone!


This plugin is an exploration of a fascinating filter type that, I think, is really underrated. It’s a straight-up extremely simple averaging filter. If you set it to integer values, it exactly averages that number of taps on the input audio (adjacent samples). If you pick in-between values it interpolates, causing the ‘frequency’ of the filter to smoothly blend between the values.

There’s a reason you don’t see this filter used for lowpasses: it’s not technically correct. In fact it whacks a great big cancellation node into the high frequencies, and the tone (while pleasing) is very obviously affected by this. It’s a bit of a flangey quality.

BUT, we know better than to be limited by technical concepts, right? After all, the classic Scream Tracker resonant lowpass is known to be broken, and that has a real character to it. And even the Roland Supersaw is known to have some quirks that make it what it is…

Average has a fantastic tonality once you accept the response quirks. You can dial in the notches to suppress unwanted content, and (in a post-video revision) it’s got a dry-wet control so you can moderate the effect—and, typically for Airwindows, if this dry/wet is set to full wet, the calculations for doing that drop out of the plugin entirely so you’re not doing an unnecessary multiply. And it’s AU, Mac and PC VST and free (because this whole project is supported by Patreon and continuing to grow month by month into a very awesome public service for all plugin users!)

Hope you like it :)


TL;DW: Limit or wipe out the highest frequencies, near the Nyquist frequency.


Sometimes I just get lucky :)

This plugin is just the anti-aliasing technology I was experimenting with in 2010, applied to Slew, my simple slew clipper. I thought it was going to make the slightly grungy Slew smoother (Slew is a clipper, and can be expected to have clipper-like qualities). That anti-aliasing code was a bit odd. It’s possible I made a mistake somewhere.

If so, I’m keeping it, and now what I invented is out for VST also, and free to all.

Slew2 doesn’t act anything like Slew. What it does, is it puts a particular behavior on the extreme highs. Set to an intermediate value, it’s an acceleration limiter. Cranked all the way up… well, check out the video. It behaves like one of the nodes in Average (also coming to VST) with a 100% cancellation, but the point of total cancellation is also the Nyquist frequency. Slew2 produces a very natural-sounding brickwall filter exactly at that frequency, with the response falling off faster and faster until it’s totally gone when you hit the Nyquist limit (digital sampling theory: at 44.1K (CD quality) it is 22.05K where the treble cuts off).

People are using this for de-essers, and it’s pretty much ideal for any sound that must not be overwhelming in the super highs. The peculiar character of the rolloff means it’s suitable for almost anything because it won’t affect lower treble, and the fact that it’s really a clipper means you can use the slider to set a threshold where, the harder you push Slew2, the more it’ll refuse to let more brightness through. I’ve not tried using it as an acceleration limiter on a mastering lathe, but it could probably do that: which also means if you’re going for classic analog tone, you can just toss this on the 2-buss and use just enough to take excess brightness away. Slew2 is one of the ‘secret weapon’ plugins I was specifically asked to port to VST, and I’m happy to bring it to you all.

Still doing the Patreon, without which I will be one poor and hungry DSP nerd. Please support it and encourage others to support it, because I have a lot of plugins still to come. It’s already 11 plugins in just a few months, and there are many more, not even counting new inventions in the works.

Dimensional Kicks/Hats/Snares for Patrons

I’ve been asked to do something that’s only available to my patrons, but I’m not willing to do that with my plugin software. Therefore, I’m doing it with samples, because I can!

These are collections of kicks, hats and snares, for EDM or sample augmentation. They’re tracker-style, not ‘simulation’ style, meaning they use no round-robining or dynamic variation (many do have several intensity levels that work together, though).

You can watch the video for over an hour of explanation or just download the zip and make some music :) the samples ending in 16 are 16 bit wavs suitable for use in traditional trackers like MilkyTracker (XM can be used somewhat in Unity, and they’re reduced to 16 bit with the Naturalize dither so they’ll adapt to any tracker need). The samples ending in 24 are 24 bit wavs and what I’m using in Renoise, or you can use them in other DAWs anywhere you need a good EDM-style sample.

In the video I explain what’s special about these: such as the ‘crystal’ hat samples that are done with a harp mic for a special dark tonality, or how the Jomox Kicks are all sampled with a DI (from a modded Jomox MBase II) plus run through a guitar amp with a modded Airwindows SM57 miking it, and lastly with a room mic that also picks up the trigger keyboard’s keys being hit with a drumstick for extra air on the transient. This is a pretty killer sample set for any electronic purposes, even before you get frisky with Renoise or whatever. Initially I was going to release these in dribs and drabs along with songs to demonstrate them, but I’m working so hard on plugins I have no time to make music. Therefore, I’m releasing the whole entire collection, because I know I’ll only end up making more.

Join the Patreon at least once for at least 1$ and you can download this there: it’s not on the Airwindows website, just on Patreon. :)


TL;DW: With this buss compressor, sub-lows get bigger the harder you drive it.


Compressors are tricky little beasts: they tend to eat low-end, they react differently at low intensities than when you slam ’em, and each sort of compressor has a ‘voice’ that defines how it acts.

There’s a crude way to deal with the low-end problem, which is to put in a crossover: either to split into multiband compression, or simply highpass the sense circuit so the compressor can’t compress extreme lows and just lets them through. I’ve been asked to do this and have actually tried it on PurestSquish, but I was never satisfied with the results. It seemed like just half-breaking the compressor and I felt there had to be a better way…

Thunder is that better way. Instead of passing through the bass, it transforms it in a way similar to FathomFive, and uses that live, active signal to modulate what the compressor sees, where in turn the compressor’s output is modulated by the intense lows. It’s a little complicated, so you can watch the video or simply download it and try it. Thunder is free (bear in mind that my only compensation is the Patreon I’ve created. To do this onward into the future, that Patreon must succeed)

If you’re auditioning Thunder, a word of warning. You’re going to have to monitor the extreme lows. This plugin can prepare music for the hugest sound systems or the finest audiophile playback, but if you can’t hear what it’s doing you may get in trouble with it. Decent headphones ought to suffice (not earbuds!). If you have subwoofers, this will test them. You can also use the mix-check plugin SubsOnly to test how you’re doing, that will spotlight the subs for you.

The range of adjustment starts with gentle ‘glue’ compression as the subsonics begin to open up the mix. Then, a bottom octave will appear anchoring everything. Push beyond that and you start to get into more compressed 2-buss, with the extreme lows taking on a punchier, thumpier character. Add more Thunder and the bass gets frisky and aggressive and dominates. Then, when you go even beyond that, we’re talking about ridiculous monstrous mega-bass, and more or less wrecking the sound of everything else (it’s useful to be able to go to weird broken settings in case you want to step back just a bit and have a really extreme effect). This is not a multiband compressor (there’s only one stereo comp in the plugin) but when driven really hard it can go even more bonkers than a multiband compressor.

Remember, the low Thunder settings are useful too as a glue comp, for a buss compressor that’s extremely transparent and true to the tone of the mix! The middle settings are just as functional. All of these settings have their own usefulness. Choose wisely (or unwisely, if you prefer).


TL;DW: Minimalist 2-buss safety clipper at -0.2dB with powerful anti-glare processing.


It occurred to me I hadn’t brought any clippers to VST yet. Also, I’ve been getting requests for the latest ADClip in VST format.

Thing is, ADClip is part of the for-pay collection that was formerly sold through Kagi. It’s got up to ADClip6 and grew all kinds of features, like the ability to monitor clips only, to store clipped-off energy and re-introduce it into the sound for added loudness, to monitor those things, and so on.

There’s also an alternate clipper in that collection, OneCornerClip, which makes the leading edge of the wave blend progressively into the clip at a rate you specify. That one can be set to a wide range of tonal effects under heavy clipping, and that too is part of the Kagi collection.

Those are stretch goals on the Patreon. If I start skipping ahead and doing stretch goals for nothing, what good are they? It’s not even fair to run around producing free versions of the Kagi plugins right away: firstly, they can still be bought in Audio Unit form and secondly, people spent money to get sounds you can’t otherwise have. I can’t do ADClip (or OneCornerClip).

However… I had freebie clippers that hadn’t yet seen VST format.

Here’s what I’ve done. I started with Clip2 and the original Clip freebie, and I used the new plugin as a way to experiment with ideas about the recurrence of mathematical constants. For instance, a common value used in reverb allpass filters is 0.618… which can be continued into a mathematical constant, the Golden Ratio (0.618033988749894848204586). It’s common for these things to turn up in disparate situations, so I look for them. And in the code where OneCornerClip rapidly converges onto full clipping with the broadest resulting variation of clip onset tonality, I found it zeroing in on cos(x) == x (0.739085133215160641655312) and selecting that optimal value for the guts of the thing I ended up with ClipOnly.

If you watch the video (one of the least necessary to watch for instructions: it’s a clipper at -0.2dB) you’ll see a lot about the other factors that went into ClipOnly. I show ADClip and its many controls, but then make an impassioned case for designs like ClipOnly, a known quantity with a predictable sound and no controls or adjustments whatsoever. Under normal circumstances there should be nothing you’d want or need to adjust, since the clip level combined with the anti-glare prevents any ‘digital reconstruction overs’ in normal use. It doesn’t do anything tricky to ‘enhance loudness’ beyond what it already is, unlike recent ADClip versions. It does have a very sophisticated enter/exit clip behavior, but to put it bluntly you don’t need to be thinking about that when mixing into a safety clipper (or tracking into such a mix).

In fact, you should not be thinking of that OR me while tracking your music and mixing it, you should be thinking about the sounds you’re making, and any distraction into the lands of ‘oh gee Chris is so clever, ClipOnly was well worth the money, look at all these knobs’ is hurting you even as it rewards me, the plugin developer, with attention.

And THIS is why I program things like ClipOnly, where possible. It might be the most forgiving safety clipper yet, as far as avoiding glare and harshness (I’m measuring extreme highs as much as 12 db down from the max under heavy load, and usually clippers by their nature exaggerate ‘crunch’ even when perfectly implemented). But ClipOnly is also part of a concept that I can run with now that I use Patreon: I get to make stuff that’s not showing off, either with controls or concepts or fancy faceplates, and just delivers the sonic goods.

I hope you like ClipOnly, but I also hope you quickly stop noticing anything about it, so you can focus on your mixing. You should be able to work into it, even hit it with heavy load and severe clipping for effect, without fussing or being distracted. The best silver bullet plugins are not the trickiest :)

Older Posts


handsewn bespoke digital audio

Kinds Of Things