The No.1 Website for Pro Audio
Airwindows Interstage: Mac/Windows/Linux AU/VST
Old 18th August 2019
  #1
Airwindows
 
chrisj's Avatar
Software Airwindows Interstage: Mac/Windows/Linux AU/VST

and here's the video, uploaded later that night!


TL;DW: Interstage is a subtle and sophisticated analogifier.

Interstage.zip(341k)

So here's a first. Today, I'm more reliable than Google. I have a video, but my ability to log in to YouTube is broken. (Yesterday, it was my ability to log in to Google Docs on a doc I was invited to edit.) So, my demonstration goes unheard, because I can't sign in and upload it.

And will that stop us? NO!

Interstage is one half of an experiment. I pitted my ability to use my weird techniques (interleaved IIR, slew limiting, etc) against my new biquad filtering techniques, in the battle of the bandpasses. The biquads could give me total DC rejection and total rejection near Nyquist. My more distinctly Airwindows tricks could give me new sorts of nonlinearity and tone. Which would win?

This isn't called 'Biquadstage', so that might be a bit of a giveaway.

Interstage isn't a loudenator. It doesn't really clip (though it does go dark as you push the highs harder, in a way I've never done before). It doesn't even preserve the output peaks of heavily limited material: it'll reshape lows in such a way that the peaks might go up slightly as the deepest lows get rearranged, and it doesn't really eliminate DC offset either. So what does it actually do?

*deep breath*

Two level total of three pole IIR highpass which subtracts a pre-averaged sample and slew limits (all right, clips) against not the direct signal but the initial stage IIR lowpassed reference point used as part of the highpass. Oh, also the average it uses isn't the previous input sample, but the slew limited highpassed output.

*crickets*

Bet you wish you had the video, now and I am NOT making that up. That's literally what it does, you can see the code (that, too, is up just fine: only the video is on hold)

What does it sound like? It sounds like running through an optimal analog stage. The lows are reshaped in a characteristic way for a capacitor-coupled circuit that still allows extended lows: this doesn't suppress much if any extreme bass. It just massages it a bit. The highs run into active component electronic limits, but unlike other approaches (Channel for a bit of grit, Acceleration for ultra-clear) this is restricting treble slews based on the general amount of energy in the circuit. So it goes darker in a peculiarly analog-like way I've not done before, sounding still clear and trebly for most audio, but confining the craziest most digital-sounding treble swings into a zone that sounds like hardware. And this is without thousands of math operations of heavy processing: unlike overprocessed analog modeling, this one nails the 'energy coming out of analog circuitry' without blurring or thinning the audio at all. If you don't need what it's doing you'll hear no change at all. Only when 'excessively digital bass and treble' show up to interfere, does Interstage kick in.

I'm pretty sure Patreon is still up, so what you should do is what you should usually do: don't take my word for it or rely on my peculiar demos. This time, you can't even if you wanted to! But you can download the plugin and try it. Throw it on something, preferably something you'd like to make a little more analog-ized, and see if you can pick out the differences. Interstage does not distort or saturate, so it shouldn't change your gain staging or the density of your tracks. It simply focuses stuff a little for you, in the way that real analog circuitry inevitably does, and it's a new flavor on that theme. It's designed to work optimally at any sample rate you give it, and guaranteed not to diminish any of your tones (unless those tones require strangely artificial sonics that can't exist in nature: if so, put this on your OTHER tones and you'll have a nice contrast).

Wish me luck following up with the video when YouTube allows… and I hope you enjoy Interstage!

Last edited by chrisj; 19th August 2019 at 04:36 AM..
Old 18th August 2019
  #2
Gear Maniac
Hope will see you, hear you and listen. Like your videos, explanations and plugins.
Then I will try this next weekend)
Old 18th August 2019
  #3
Gear Nut
 

This new marketing ploy to get people to demo the actual plug using their ears only is SO retro. This kind of thing might just catch on!! Who needs YT?!?!?!!
Old 18th August 2019
  #4
Lives for gear
 

Yes, this is gorgeous. Looking forward to the other 'halves' and lead-ons from this experiment!!
Old 18th August 2019
  #5
Here for the gear
 

Can’t wait to try this one out. I’ll post my thoughts when I’m finished
Old 19th August 2019
  #6
Gear Maniac
 

YouTube fixed itself quite quickly...

Waiting to get back home [I'm abroad] and giving this a go.
Old 19th August 2019
  #7
Gear Maniac
 

Quote:
Originally Posted by chrisj View Post
and here's the video, uploaded later that night!


TL;DW: Interstage is a subtle and sophisticated analogifier.

Interstage is one half of an experiment. I pitted my ability to use my weird techniques (interleaved IIR, slew limiting, etc) against my new biquad filtering techniques, in the battle of the bandpasses. The biquads could give me total DC rejection and total rejection near Nyquist. My more distinctly Airwindows tricks could give me new sorts of nonlinearity and tone. Which would win?

This isn't called 'Biquadstage', so that might be a bit of a giveaway.
I just hope the biquad thing would find itself in the new version of Desk, at some point [BiquaDesk?...]

BTW
Can you go again through what happens when one stack 'em Interstages? additionally, I keep wandering about the inteerction between Interstage and all your tools that shape the hi-end... [Inc. All your clippers and slew rate manipulators]
Old 20th August 2019
  #8
Wow chris, I must say this might be my fave plugin from you thus far. It really does something subtle to the individual track (I tried it on a vocal and a drum Buss). It’s like it softens the clip peaks in the highs and gives off a nice rounded low end but in a subtle way. Kinda gives the track some depth too! I’m really in love with this one thanks for this!!!
Old 20th August 2019
  #9
Here for the gear
 

Looking forward to trying it out, but just as a suggestion for these more proof-of-concept plugins (really only the ones that change based on volume), it would help to have a slider that controls input and output gain, whether that's in the form of a normal gain slider or a "ceiling" control. But basically boosting by a certain amount into the plugin and then attenuating by the same amount after it.
Old 20th August 2019
  #10
Airwindows
 
chrisj's Avatar
Quote:
Originally Posted by JayJ1095 View Post
Looking forward to trying it out, but just as a suggestion for these more proof-of-concept plugins (really only the ones that change based on volume), it would help to have a slider that controls input and output gain, whether that's in the form of a normal gain slider or a "ceiling" control. But basically boosting by a certain amount into the plugin and then attenuating by the same amount after it.
True, and there are many plugins I do that cover that base. This one's a little special though: seeing as it's called 'Interstage' it mostly doesn't change so much based on volume, has no saturation function (just the unusual form of slew clip) and its role is to sit there like the circuitry in your analog board.

So in this case I am particularly unrepentant about the design this one, it's really important for it to just be there with nothing to adjust. If you gain stage particularly hot through it, you might notice it darkening a little more. It needs to react like an analog board, and part of that is having a predictable sound at various signal levels, in a subtle way.

It'll probably make more sense when I begin bringing things together into 'Console-like' sets of plugins that do the Console thing but also have more channel-strip-like features and behaviors: there have always been plugins with non-adjustable tone-forming parts to them. GuitarConditioner is another, as is TubeDesk and TransDesk, and of course the dithers. I'm particularly fond of the 'build these into your template and just have them at relevant places in your mix framing' plugins, especially when they have no 'fiddle elements' to distract you. Interstage is like the platonic ideal of that sort of thing: it's incredibly clean, but its effect will be felt if 'analogification' is what you're after.
Old 20th August 2019
  #11
Here for the gear
 

Quote:
Originally Posted by chrisj View Post
If you gain stage particularly hot through it, you might notice it darkening a little more. It needs to react like an analog board, and part of that is having a predictable sound at various signal levels, in a subtle way.
Ah, cool! I thought that was the main thing it was doing, but that's fine then.

Although it's mainly things like that [the sound getting darker] where it would be nice to be able to try out applying that part of the processing to the sound without setting up all the gain staging around it AND then figuring out how to bypass all of that to hear the difference.
Old 20th August 2019
  #12
Gear Addict
 

Perfect! This and slew2 is what my virtual synths needed to be more real.
Old 20th August 2019
  #13
Gear Maniac
 

Quote:
Originally Posted by linkstate View Post
Perfect! This and slew2 is...
This and many many more. MANY MANY MORE [from AW].

It feels like we were handed a real electronic capacitor, and we can choose where to solder [patch] it in the [virtual electronic] chain.

Wow Chris, just wow... Highly appreciated !
Old 20th August 2019
  #14
Airwindows
 
chrisj's Avatar
Quote:
Originally Posted by JayJ1095 View Post
try out applying that part of the processing to the sound without setting up all the gain staging around it AND then figuring out how to bypass all of that to hear the difference.
We have some philosophical differences there. The way I see it, all the time you spend to do that is time spent not thinking musically. If you needed to hang a red flower across your monitor to get the right vibe for mixing a particular track, I'd be totally down with it but would not endorse you dropping out of the zone to test whether the red flower was REALLY helping you: that would be counterproductive.

I can make stuff that doesn't harm your sound. I can make stuff with subtle effect, and stuff that's measurable and does predictable things (that said, I'm going to need to look into a possible misbehavior at high sample rates with this plugin, I'm hearing about problems).

But much of what I'm doing is working at various thresholds of significance, and by design I'm trying to make the subtler side of things more appealing and inspiring. If 90% is fine you can just get GarageBand and call it a day: we're not about that.

You should not be spending more time second-guessing yourself than you spend mixing. That can be in separate sessions, where you hash out which products, plugins, instruments are trustworthy. Once you start creating, that's not the time to stop and test everything. I get that many products (not always mine) build in deceptiveness such as 'marketing boosts', but over time the truth will become apparent as you'll naturally compensate for such tricks in normal mixing, and then develop dissatisfaction with stuff that's not working the way you want.
Old 20th August 2019
  #15
Here for the gear
 

Quote:
Originally Posted by chrisj View Post
dropping out of the zone to test whether the red flower was REALLY helping you: that would be counterproductive.
I'm not sure if it really is philosophical differences, just that I'm wanting that extra control for a couple of reasons:

First, in order to better learn what it is that the plugin does. Now, of course, that's kind of difficult for something that's designed to be subtle, but I think it's worth putting in some time to learn what your tools actually can do.

Then, once you know the plugin well, it's a case of being able to recreate whatever effect(s) you found while learning the plugin. Say I have a track where after all the other plugins, it's peaking at -18dB and (after testing the plugin out) I want to add Interstage and have it darken the sound roughly to the extent it does when things are peaking at -2dB. That's the kind of situation where I don't want to have to stop mixing for a bit and insert gain plugins before and after as that seems counterproductive to me.
Old 20th August 2019
  #16
Airwindows
 
chrisj's Avatar
I've got an update: redownload it and give it a try. What I changed was, it's going to use the SAME threshold for 'how far it can slew in a sample' for every sample rate, rather than scaling it. What I had, was the threshold getting way smaller if your sample rate was higher (on the rationale that it'd be the same, since you were covering the 'same distance' to get the same slew)

I had folks at high sample rates running into issues, so here, try the same thing with one change: not altering the 'threshold' for the slew by sample rate. It's always the same threshold now. No other changes, and no changes if you are at 44.1K.

Interstage.zip(341k)

Since the former should have been exactly right and wasn't, and since this is the same at whichever rate you use, this is more likely to be exactly right (and work properly at higher rates: it's still very much active, in fact now it's identical to what's in the 44.1K version).
Old 20th August 2019
  #17
Gear Maniac
 

What would be the optimal gain staging before the highs start to clap down ?
Old 21st August 2019
  #18
Gear Nut
 

Does it act differently by sample rate? It seems 88.2/96 ver kicks in easily than 44.1/48. I think channel6 was same. And what's the different from Channel6's slew clipping(besides threshold.)?

Excuse me if I'm asking pointless question.
Old 22nd August 2019
  #19
Gear Maniac
Quote:
Originally Posted by kojei View Post
Does it act differently by sample rate? It seems 88.2/96 ver kicks in easily than 44.1/48. I think channel6 was same. And what's the different from Channel6's slew clipping(besides threshold.)?

Excuse me if I'm asking pointless question.
Have you downloaded the patched version? It shouldn't do the "engaging faster at higher sample rates" thing, now.
Old 23rd August 2019
  #20
Gear Nut
 

Quote:
Originally Posted by SleepCircle View Post
Have you downloaded the patched version? It shouldn't do the "engaging faster at higher sample rates" thing, now.
Oops, somehow I had missed last few posts. Patched version works better than previous version but it still works differently at higher sample rate. It seems higher sample rate version works more broad frequencies than lower oner?I also checked with musical source and the difference was bigger with higher sample rates than lower one.

Attached pics are -6dBFS sine sweep with interstage and shows different results.

Channel6's slew clipping is same behavior. So, I wanted to know which one is ideal.(and the difference between chnannel6's slew clipping and interstage one...)

edit: every sample rate sounds good though
Attached Thumbnails
Airwindows Interstage: Mac/Windows/Linux AU/VST-441khz_sweep.jpg   Airwindows Interstage: Mac/Windows/Linux AU/VST-882khz_sweep.jpg   Airwindows Interstage: Mac/Windows/Linux AU/VST-thd_vs_frequency.jpg  
Old 23rd August 2019
  #21
Gear Maniac
Huh! That is a baffling pattern; how interesting.
Old 23rd August 2019
  #22
Airwindows
 
chrisj's Avatar
Don't know what to tell you, other than none of it is harmonic distortion at all, sine sweeps won't help and will be hugely dependent on signal level, and the whole thing is a kind of hard clip that's meant to produce the sorts of artifacts you see, just like a soft saturation will produce literal and more predictable harmonics.

Instead, I'll follow up with questions of my own, and an answer. Interstage has a different slew clip than Channel because it's operating from a stored reference point that's itself lowpassed, so it clips strictly from a treble-dulled signal and not from the real signal. You don't hear that signal directly: it's part of the highpassing (which Channel also does…) but not heard directly, and it's a fixed frequency that is determined according to the sample rate, not just a fixed proportion of the sample rate (it should be relatively stable and not go up in pitch when you go higher in sample rate)

Do you like the sound of the way Interstage does it, better than the sound of how Channel6 does it? Channel6 scales the slew clipping like the original 'buggy' Interstage did, so it'll kick in faster at higher rates (it shouldn't but this seems to be what happens). But Channel6 uses direct calculation of the current and previous sample, and NOT the more sophisticated 'smoothed' reference position.

There is no way to tell whether any of them are 'ideal': you're not even seeing a sensible measurement when doing a THD plot like that, it's not even the same kind of distortion, there are no right or wrong answers other than: does one sound better to you? They can't sound the same at different sample rates: slew is measured using sample rate and trying to make it 'the same' made it hugely different for reasons I can't currently understand. Technically, you should stop seeing any effect at really high sample rates because the slew between samples will become tiny, but that doesn't happen. So we're all confused together.

I've got an Interstage that doesn't start applying MORE effect with high sample rates. I could give you a Channel7 which likewise doesn't get slew-clip-happy at high sample rates and is otherwise the same… or I could give you a Channel7 which works with the stuff I introduced in Interstage, without too much trouble (sort of supplanting Interstage in the process, but Interstage will continue not to have saturation) or even do that and let you crossfade between Density and Spiral algorithms at the cost of adding yet another slider to it (I'd been asked for this and it led to the discovery of Mojo and Dyno, but not to a 'Spiral-ness' control which I'd been trying to get. But a simple crossfade is do-able)

Which sounds best? It looks like a Channel7 is in the cards, because these things directly affect it or even count as long needed bugfixes for the use of Channel at high sample rates.
Old 23rd August 2019
  #23
Gear Maniac
It's very strange that the slew clipping gets changed by the sample-rate, even when you had something that was supposed to compensate for that. What was the math for it, again? Maybe instead of changing the threshold depending on the sample rate, you could modify how you calculate the slew rate over-all…?
Old 24th August 2019
  #24
Gear Nut
 

Thanks for the answer.

Honestly, I'm still in the phase to experiment your slew thing for my mixing chain. I was really interested in saturation into slew-clipping chain.. So, I can't verdict it now. Checking higher sample rates is part of my experiment to check the artifacts of aliasing from saturation though I just found different slew clipping behavior during that experiment.

My current go-to saturation is eq pre/de-emphasised Pafnuty into Channel6(drive less than 50%) at 44k/48k. I wanted to use channel6 as DC reducer and slew clipping(Neve threshold). I don't know how slewclipping contribute to this chain since I obviously don't use it very hotter(usually VU=-18 to -24).

To my understanding, slew clipping works well for big transient peaks like contains in drum's OH channels and it's good thing to maintain dynamic range for me.

But still I can't verdict about the changing sonic-wise from slew clipping.

Channel7 is welcome but I don't rush you for this because Channel6 at 44k/48k works for me right now.

Anyway, I'm still wanting to learn the uses of slew clipping...

Quote:
Originally Posted by chrisj View Post
Don't know what to tell you, other than none of it is harmonic distortion at all, sine sweeps won't help and will be hugely dependent on signal level, and the whole thing is a kind of hard clip that's meant to produce the sorts of artifacts you see, just like a soft saturation will produce literal and more predictable harmonics.

Instead, I'll follow up with questions of my own, and an answer. Interstage has a different slew clip than Channel because it's operating from a stored reference point that's itself lowpassed, so it clips strictly from a treble-dulled signal and not from the real signal. You don't hear that signal directly: it's part of the highpassing (which Channel also does…) but not heard directly, and it's a fixed frequency that is determined according to the sample rate, not just a fixed proportion of the sample rate (it should be relatively stable and not go up in pitch when you go higher in sample rate)

Do you like the sound of the way Interstage does it, better than the sound of how Channel6 does it? Channel6 scales the slew clipping like the original 'buggy' Interstage did, so it'll kick in faster at higher rates (it shouldn't but this seems to be what happens). But Channel6 uses direct calculation of the current and previous sample, and NOT the more sophisticated 'smoothed' reference position.

There is no way to tell whether any of them are 'ideal': you're not even seeing a sensible measurement when doing a THD plot like that, it's not even the same kind of distortion, there are no right or wrong answers other than: does one sound better to you? They can't sound the same at different sample rates: slew is measured using sample rate and trying to make it 'the same' made it hugely different for reasons I can't currently understand. Technically, you should stop seeing any effect at really high sample rates because the slew between samples will become tiny, but that doesn't happen. So we're all confused together.

I've got an Interstage that doesn't start applying MORE effect with high sample rates. I could give you a Channel7 which likewise doesn't get slew-clip-happy at high sample rates and is otherwise the same… or I could give you a Channel7 which works with the stuff I introduced in Interstage, without too much trouble (sort of supplanting Interstage in the process, but Interstage will continue not to have saturation) or even do that and let you crossfade between Density and Spiral algorithms at the cost of adding yet another slider to it (I'd been asked for this and it led to the discovery of Mojo and Dyno, but not to a 'Spiral-ness' control which I'd been trying to get. But a simple crossfade is do-able)

Which sounds best? It looks like a Channel7 is in the cards, because these things directly affect it or even count as long needed bugfixes for the use of Channel at high sample rates.
Old 28th August 2019
  #25
Gear Nut
 

K I'm officially in love with/addicted to Interstage, this baby is great!
Old 31st August 2019
  #26
Gear Addict
 

Really nice sounding even by Airwindows standards.
Old 21st October 2019
  #27
Gear Addict
 
candyflip's Avatar
In regards to gain-staging, how hard are people hitting Interstage?
Old 21st October 2019
  #28
Lives for gear
 

Quote:
Originally Posted by candyflip View Post
In regards to gain-staging, how hard are people hitting Interstage?
I found I like the Interstage effect to be stronger so I try to hit it somewhere between -12 and -6 db. Depending on the source. More than that gets into more obvious distortion to me.
Old 21st October 2019
  #29
Gear Addict
 
candyflip's Avatar
Quote:
Originally Posted by profvonsok View Post
I found I like the Interstage effect to be stronger so I try to hit it somewhere between -12 and -6 db. Depending on the source. More than that gets into more obvious distortion to me.
Good to know, thanks! Only just getting round to this one...there are so many AW plugs, Chris is a monster! Looking forward to trying it out.
Old 4 weeks ago
  #30
Lives for gear
@ chrisj where would you insert Interstage : inside or outside the (Purest)Console chain ? ?
Topic:
Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Forum Jump
Forum Jump