Airwindows Flutter2: Free Mac/Windows/Linux/Pi CLAP/AU/VST3/VST2/LV2/Rack
by chrisj   22nd July 2024

You made me laugh with the intro. And I enjoy it when ya go all mad scientist! This is a cool piece of gizmology.

Dear Chris, just a question about ToTape7, if you would not mind answering such a thing at the moment. The whole NAB/IEC curve system is more elusive to me than I'd like to admit, but I am assuming it requires the recording going into the tape machine having a treble boost; the pre-emphasis, and the playback having certain treble frequencies attenuated; the post-emphasis. Should it then be so that the channels and busses of a mix would have a tape plugin set at pre-emphasis and post-emphasis values respectively, or is the pre-emphasis/post-emphasis thing supposed to be happening within a single instance of the plugin itself? Apologies if this question extends beyond the scope of this plugin.

That bass example is other-worldly. I've never noticed this effect whilst using other plugins with flutter, but I can't understand what makes your version different. Whilst fluttering, is the relationship between the speed and pitch of the audio somehow unique to this plugin?

I love ToTape6 (especially the soften control) so I'm excited for 7, giving us increased control sounds exciting. I think wow would also be a huge inclusion, sometimes I miss it from ToTape6.

Will the separate left/right flutter become a part of your ToTape series? My assumption was that multitrack tape flutters in a uniform manner between tracks, but I've never actually used analogue tape.

Have you considered the next ToTape as a multi-channel plugin? Given the nature of multitrack tape, I'm surprised I've never seen a 24/48+ input tape plugin with separate saturation stages for each input. If things like flutter are also unique per channel then that too could be implemented this way, but with a single plugin (with a single set of controls) being fed by multiple tracks. It would also mean a single instance of oversampling for an entire projects worth of tape processing. Having as many outputs would mean it could also be used in a "traditional" style at the beginning of each track that it outputs to, feeding your Console plugins (or an analogue console, or just a standard DAW mix).

I guess the same question would apply to your Channel plugin for mainly the same reasons, but also because neither analogue tape nor Channel have a summing stage. Therefore we could still have efficient global Airwindows processing but leave the summing to our DAW, the benefit of which becomes apparent when printing stems or multitracks of a finished production; I enjoyed Console but I need DAW summing lest my multitracks sound different than the approved 2-track bounce. Multiple ins/outs on ToTape and Channel would offer profound flexibility and useability for producers, composers of music for picture, and mix engineers printing for immersive audio.

To my ears this is MUCH more 'correct'-sounding than flutter1

Thank you very much for this, Chris (And all your other plugins)

Flutter is multi-facetted and happens, if you want to get really dirty with it, in 3 axes.



Axis-1 is the tape-speed, that's longitudinal flutter, which affects the pitch of a sound.

This fluttering happens BOTH during recording AND during playback.

You can imagine how it would sound if you recorded a sound and put your finger on the tape to stop it briefly during recording. When you played back this sound it would have a short pitch-jump where you stopped it during recording (It's technically called a speed-ramp)

If you use your finger to speed up the tape while playing it back, then this pitch-jump would sound different than if you hadn't touched the tape during play-back.

So from this you know that recording adds its own flutter, and playback also adds its own flutter.

For multi-track tapes this means that if you record a completely stabile sine-wave onto a number of individual tracks, recording it to only 1 track at a time (Like if you wanted to record your own voice multiple times to create a layered choral piece that sound like multiple people singing the same phrase), then you get a channel independent flutter per track.

If you record the sine-wave to all tracks at the same time, then you get the same flutter on all tracks.

It's 2 very different (Well, technically different, as in reality it's very subtle nuances of difference) sounds, and depends on what kind of tape-recording method you want to simulate (All tracks recorded at the same time, or each track recorded on its own one after another)



Axis-2 is channel-bleed, which happens as each tape-lane moves laterally in and out over the other lanes play-heads.

Technically this also happens both during recording and play-back, and simply means that part of one tape-channel bleeds into another channel.

This could mean that if you record a dry sound on track 1 and track 2, and during play-back feed track 2 to a reverb, that track 1 occasionally would also get fed into the reverb (The amount depending on how far the track would wobble in and out of its lane, which usually is only very little of course. Unless you touch the tape physically to literally push it around, which can be used for odd effects although it rarely sounds very good to do this. Yes I've tried it )



Axis-3 is head-pressure, vertical flutter, which is usually mostly directly connected with longitudinal flutter, meaning that as the tape-puller pulls harder on the tape (Speeding it up) this also increases the tape-pressure against the play-head, and when the trailing tape is too fast this leads to less pressure against the play-head.

Imagine yanking a dog on a leash; (You obviously shouldn't, but some dog-people do it all the time. I have a near-by neighbor who does it, resulting in some painful squeals coming from the dog every time) when you first pull there is a lot of resistance (The weight of the dog) but then as the dog jumps at you the rope slacks and becomes loose until you pick up the slack and the rope is once again stretched out.

This tends to affect the volume of a track on tape, and sometimes also the sound's spectrum.

Incidentally, this is also the axis where the chemical compound variations on the tape has their effect (Volume and spectrum variations. Think of it like noise on a film-roll that varies from frame to frame)



So those are the 3 types of flutter you get.



To use Chris' plugin to simulate a multi-track recorder, each MONO-track should have its own flutter-plugin with a MONO-output plugin after it (Voxengo has a free plugin called MSED, which you can use to kill the side-channel from Chris' flutter-plugin)

If you do this, then you will get the physically correct effect of phase-interaction (On any given frequency) with the listener's physical environment (Which means you will get an actual stereo-effect, albeit completely dependent on the listener's own environment, so not really a very predictable effect as each listener obviously have their own environment) out of a mono-track during play-back when the flutter is set neither too fast or too slow (The speed has to hit whatever speed will resonate with a listener's physical environment ) )

If you, on the other hand, want to simulate this resonating-effect (So simulating putting a listener into a specific virtual environment) then you should NOT kill the side-channel of Chris' plugin (But technically this would be like putting a reverb on a sound and then playing the sound in the real room the reverb simulates, so it becomes an exaggerated effect and no longer a real simulation, which can still be used to create interesting sounds of course. I'm sure we all use virtual reverbs all the time on our music. I certainly do )



So all this is of course academic in normal music-use, where just using one flutter plugin on each track will be quite good enough to create a much more interesting sound than the clinical digital precision allows for on its own (And for that use this version of Chris' flutter-plugin sounds really amazing and very realistic to what I have experienced with tape in terms of longitudinal flutter, which no doubt is the most interesting and most sought-after of the 3 types I do have a plugin myself, however, which simulates the vertical axis, called stereo-shimmer. The last axis, the lateral one, is the most difficult (Or at least the most tedious to set up in a DAW, and not really worth the effort for normal music-creation where maximum physical correctness of a tape-mechanism is not needed) to simulate, as it involves cross-feeding randomly/semi-randomly between audio-tracks)

But you asked, and since I have mangled quite a few tapes in my time (I apologize to each and every tape, and thank them for their self-less sacrifice in the name of science and learning and simple curiosity ) perhaps you find my geeky/nerdy answer useful (If not, then perhaps somebody, coming via a search-engine to this thread some day, will)

Tape is fun, digital is easier, they're both good and they're both bad

Not at all, in fact I thank you for reminding me as I absolutely should be including this in the upcoming ToTape7 if I mean to make it that expansive and customizable

It's very much the LATTER. The recording process is wrapped up in one instance of the plugin, including the time between recording and playback, so the NAB/IEC pre and de-emphasis must be applied before the saturation, head bump, AND in fact also the flutter algorithm. In so doing, it'll tend to saturate highs more easily. I probably also ought to look into exaggerating the effect a bit to take into account how bias frequency makes extreme highs saturate more easily, but I'll probably want to abstract that somewhat because it's too easy for a simple frequency effect to turn into just aliasing.

Bear in mind for this I'm GOING to be wanting to go into abstractions more like synthesis and analogies more like 'synth knobs that do recognizable things' rather than 'exactly imitating real hardware and interface but pretending it's not a plugin'. It's for that reason that Flutter2's controls are labeled like they're an LFO, even though it's not strictly just an LFO as it's not just a traditional oscillator.

You've even reminded me of something to look into and also update: I have a plugin already out, called Dubly. It's not part of any ToTape plugin… yet. But it's an abstraction of exactly what you describe, so try playing with Dubly and see what you make of it. I've already made a plugin like that, I just need to bring it up to speed with what I'm doing lately, see if there are things I'd change. In particular, I think I'd be wanting to allow more user control over the frequencies at which it operates and the depth of effect, expressed in more synthlike terms

Four

I love your thinking. Indeed it's true that you're getting TWO flutters each time you use tape, the record one and the playback one. And some of your axes are quite interesting, and it's probably worth me keeping note of your point that flutter alters head pressure which has known effects on the sound and that modulation is tied to flutter. But consider this, it's affected how I got to Flutter2 and I'm keeping it for the new ToTape7.

Record and play back a two-track stereo tape. You've got one flutter on record, another on playback, and you're not recording the stereo tracks on individual passes, so the record pass is 'the same' flutter.

Now slice the tape up the middle, and do it again. Now you've got one pass, but entirely independent flutters, not linked at all. So far so good.

But if the tape was capable of being just a one-dimensional stretch we wouldn't have flutter at all! It can also stretch ACROSS the tape, it's just not (or technically it's narrowing slightly as it's stretched), and flutter is a phenomenon of areas of increased drag and let-go, stretch and snap-back. That drag surface is not an idealized point.

If your tape is in fact two inches wide, and your tracks are on the extreme edges of the tape, the flutter IS going to be different, perhaps entirely different, over even one pass. Not as much as I'm making it, not as much as a tape slit up the middle, but still. By definition it can't be one-dimensional any more than a speaker can be pistonic. THAT is why my Flutter2 sounds more realistic to you. The assumption that tape flutter must be identical per-pass is an awful mistake, and I'm discarding it to get the result sounding better.

Consider that if you take, for instance, a stick, and compare bending the stick to stretching it. You'll find that stretching is quite difficult, but bending the stick is way easier. Similarly, stretching and releasing a tape IS what flutter literally is, but producing a difference between sides of the tape is far easier than stretching the tape as a whole. The flutter equivalent of this would be very brief equalization-of-tension moments where one side of the tape is speeding up and the other being dragged: causing the free stretched area of the tape to move laterally a tiny amount. You could confirm this by seeing if the free area of the tape is jittering along its width axis at all, because ANY motion along that axis can only be producing pitch differences between channels, at the very high frequency we experience flutter.

five - the tape azimuth may also be fluttering, so different channels phase differences (and eq roll-off) are 'fluttering' in amount as the tape wiggles across the head.. in varying amounts depending how far from the middle they are...

Absolutely, 100%, that's the mechanism by which the channels produce different pitches. You can be certain that there is some azimuth flutter in there, and that's literally the only thing it does. Exactly right

yeah but it also rolls off the highs during the 'out of azimuth' phase as the short wavelengths become shorter than the effective head gap.. (traveling at an angle to the headgap makes the headgap seem longer)

Right back at you Chris

Yes you're correct of course, I wasn't considering those effects but tape is indeed a flexible material that can stretch and contract differently at different location-points. Good point
(Never leave tape in a hot car! )

I'm looking very much forward to ToTape7