#165: What is Mid-Side EQ? Enhance Your Stereo Image with Ian Stewart

Ever wondered what is mid-side EQ or how does mid-side EQ work? Is mid-side EQ essential for a professional-sounding mix? Tune in to EP 165 of the Inside The Mix podcast, where I sit down with distinguished mastering engineer and educator Ian Stewart to explore the ins and outs of mid-side EQ. Together, we unravel the techniques that can add width and clarity to your mix and transform your audio production skills.

We dive deep into the profound effects of manipulating mid and side channels to give your mix a wider stereo image and greater clarity. Ian takes us on a journey through the origins of mid-side EQ, sharing insights into how to control mid-channel loudness and boost side channels for professional stereo width. Learn how these principles can also be applied to EQ and compression for a polished, high-quality sound.

In this episode, we break down the art of mid-side recording, using directional and figure-eight microphones to craft a spacious stereo image. Ian also clears up common misconceptions around polarity and phase, helping you avoid typical pitfalls when setting up a mid-side stereo field.

Ian also provides a masterclass on how mid-side EQ impacts stereo balance and phase response, essential knowledge for mastering engineers and producers alike. We dig into linear and minimum phase EQ effects on stereo imaging and discuss advanced compression techniques that can elevate your production game.

In this episode, we cover:

• What is mid-side EQ and why you should use it
• How manipulating mid and side channels affects your mix’s width and clarity
• The history of mid-side EQ and its origins as a micing technique by Alan Blumline
• Understanding polarity, phase, and common mistakes in mid-side setups
• Using linear and minimum phase EQ for optimal stereo imaging
• Advanced compression tips using mid-side processing for a polished sound

Follow Ian Stewart for more insights: https://flotownmastering.com/

Chapters

• 0:00 Mid-Side EQ Technique Explained
• 5:07 Exploring Mid-Side Audio Techniques
• 12:03 Misconceptions of Mid-Side EQ Operations
• 17:55 Understanding Mid-Side EQ Techniques
• 29:58 Stereo Imaging With EQ Techniques
• 35:48 Linear Phase EQ Effects on Stereo
• 41:02 Understanding Mid-Side Processing
• 48:59 Advanced Compression Techniques and Tips

Transcript

Ian Stewart: 0:49

The key here is that if you're making the mid channel quieter, it's going to get wider. If you're making the mid channel louder, it's going to get narrower. Side is the opposite right? If you're making the side channel louder, it's going to get wider. If you're making the side channel quieter, it's going to get narrower right. So that's kind of the key and then you can apply that. You can superimpose that on whether it's EQ or whether it's compression or whatever. Right, whatever you're working on, you can think about how am I impacting the level and to what channel? And then that's. Then you use your little decoder ring right mid channel louder, narrower. Mid channel quieter, wider. Side channel louder, wider. Side channel quieter, narrower.

Lij Shaw: 1:33

Howdy inside the mix podcast fans. It's lidge shaw and you can follow me or find out more at recording studio rockstarscom, which is my podcast as well. Right now, however, you are listening to the inside the mix podcast, and here's your host, mark matthews hello and welcome to the inside the mix podcast.

Marc Matthews: 1:54

I'm mark matthews, your host, mark Matthews. Hello and welcome to the Inside the Mix podcast. I'm Mark Matthews, your host, musician, producer and mix and mastering engineer. You've come to the right place if you want to know more about your favourite synth music artists, music engineering and production, songwriting and the music industry. I've been writing, producing, mixing and mastering music for over 15 years and I want to share what I've learned with you. Hello, folks, and welcome back to the Inside the Mix podcast. In this episode, I have a returning guest, and it's not often I have returning guests. It's not because I don't want to, it's just because I have a lot of people on here, but I'm really excited for this one. So I've got the mastering engineer and audio educator from Western Massachusetts, ian Stewart, back on the podcast. Ian, thank you for joining me today. How are you?

Ian Stewart: 2:40

Hello, hello. Thank you so much for having me back. I'm doing really well.

Marc Matthews: 2:47

Magic. I'm good, I'm good. This is an exciting episode. I was just saying off air to Ian audience listening that this is a new it's not a new format, but it's a frontier. It's a new frontier, it's new territory for the podcast. No spoilers, you'll hear what I mean when we get to it.

Marc Matthews: 3:04

But for those of you who maybe who haven't listened to the previous episode with Ian so Ian has collaborated with fantastic independent artists as well as big names. He teaches mastering at Berklee College of Music. He's co-developed the Bass Lane Pro plugin with Tone Projects and writes for the iZotope blog as well. And if you go over to Synth Music Mastering, he actually contributed to a blog on there.

Marc Matthews: 3:29

And also, if you missed our last chat, it is episode 139 of the Inside the Mix podcast, where we dove into the future of AI-assisted music production and some sort of drum roll, some music going on there for this, because this is a topic that I frequently have conversations with or get requested from the audience to talk about. But I wanted to get an expert on the podcast to talk about this because I think I'd be doing it to service if I tried to do a solo one. So we're going to be looking at, for understanding its origins, to using it effectively in your mixes as well. So we're going to dive in. So maybe, if we start with the basics, can you tell us a bit about the origins of mid-side EQ and maybe how it's evolved over time, absolutely.

Ian Stewart: 4:27

So mid-side really originated actually as a micing technique, right.

Ian Stewart: 4:32

So if you've heard of Alan Bloomline or a Bloomline pair, right, a stereo pair of a spaced pair right Mid-side was another micing technique that he developed and it's really just another way of using two channels to capture stereo information.

Ian Stewart: 4:50

And basically the original micing technique was you would have one mic that was either an Omni or sometimes a Cardioid. Early days it was more typically an Omni, but as time kind of went on, more and more people would start using a Cardioid and that would be your mid mic. I'm doing air quotes there and so that you would point that at basically whatever you wanted the center of your stereo image to be. So you point that there. And then you'd have another mic and that would be your side channel and for that you would use a figure eight mic and it would basically be oriented 90 degrees to to whatever wherever your mid mic was facing, to wherever your mid-mic was facing right. So if you imagine kind of the front and back of a figure eight mic, you've got right. You've got two, these kind of two spheres that come off either side, so you would orient that across and 90 degrees to that mid-mic. And then you have to come up with there's this whole matrixing thing, right. So if you've ever dived into mid-side stuff, you've probably heard about mid-side matrices or seen matrixing plugins that can do some of this stuff. But basically what's happening is you're adding the signals together in different ways, and the thing that even allowed this to work, that allowed Alan Blumlein to come up with this technique, is this kind of particularity about figure eight mics. Is this kind of particularity about figure eight mics? So the, if we think of it as a front and back, right of where that figure eight is, the polarity of those two are opposite to one another. Okay, so polarity related to phase, not exactly the same. And we're going to I'm going to be a little bit of a stickler when we talk about this that we have to. We have to differentiate between polarity and phase. Yeah, um, right, so polarity can be. It's a binary thing, it can be one way or the other.

Ian Stewart: 6:42

So if you imagine a snare drum hit, um, if you record that, and when you play it back through your speakers, it makes the cones kind of come out towards you first and then move back in. That would be one polarity. If you flip the polarity, then the cones would move away from you first and then come out right, and so the two sides of a figure eight mic have that characteristic right On one side it's going to make the signal come out first. On the other side it's going to make it go away from you first. Again, if you're thinking about this in terms of kind of how your speaker cones are moving, or something like that from you first. If you're, again, if you're thinking about this in terms of kind of how your speaker cones are moving, or something like that, um, and so, to get the left channel, you take the mid mic and you take the side mic and you add those signals together and that gives you left, yep, okay. And then you take the mid mic and subtract the side mic and that gives you the right channel.

Ian Stewart: 7:36

And so, kind of the way that you can think about subtraction in audio is just flipping the polarity, yeah, right. And and what's happening here is because the polarities of the two sides of the figure eight mic are opposite. When you add them with the mid mic, if you're, depending on how you flip them, they cancel each other out, right? So it's reinforcing either information from the right or the left, depending on how you're manipulating that polarity in relation to the mid mic, which is kind of capturing all of it. So that's how this all started right and then people started to realize, oh, like we can perform kind of these same operations on a left-right stereo signal, we can transform that into a mid-side signal, do some operations on it, do some EQ or compression or whatever, and then take it back to left-right stereo. But at its core it was originally a micing technique and it's just another way of representing stereo using two channels. So yeah, I don't know, does that, did that make sense? Can I elaborate on any of that?

Marc Matthews: 8:43

No, I think, I think that's, I think that's good and it does make sense and it echoes something that I did when I was doing my master's degree, and it's all coming back to me now when you're talking about polarity and then when you're talking about the, the figure eight mics and how you're positioning them, because I remember going through all the various different micing techniques and and um, all the different microphone types as well.

Marc Matthews: 9:03

So it's all really interesting stuff. And what you mentioned there about polarity and phase um, so you mentioned that you wanted to discern the difference between the two why do? You think they're often maybe confused? Might be between the two. Why do you think they're often maybe confused? Might be the wrong word, but why do you think they may be misrepresented or confused maybe?

Ian Stewart: 9:20

Yeah, I think just intertangled a little bit right. I think part of it is that because phase, one of the ways that a phase change can show up is 180 degrees phase shift and that, if you don't have any other information about the audio signal looks exactly like a polarity flip. Right of course, yeah, but the thing about phase is that it can be anywhere in between, right, it could be 17 degrees or 90 degrees or 43 degrees or 128 degrees or whatever, anywhere up to 360 degrees.

Ian Stewart: 9:52

So we basically use the number of degrees that are in a circle to measure phase rotation or phase shift. And the thing about phase is that it can be specific to frequency. Right, the same amount of timing shift is a different amount of phase shift depending on the frequency. So it's not. You know, polarity is just basically flipping the sign of the signal or the direction of it, whether it's going up first and then down, or down first and then, up.

Ian Stewart: 10:23

Delay is just moving the whole thing in time. Phase is kind of doing a combination of those things and it's dependent on the degree of rotation. And so if you simplify the idea of phase, yeah, it looks a lot like polarity, but in fact it's. It's dependent on the degree of rotation. And so if you simplify the idea of phase, yeah, it looks a lot like polarity but in fact it's, and in practice, right when we get into talking about EQ and how EQs actually work under the hood and what's going on, phase is much more nuanced than just simple polarity.

Marc Matthews: 10:49

Yeah, I could see where the way you described it there makes sense and how it could be intertangled. Like you mentioned there, when you do with the phase and you flip it that 180, then it would look like, or perceived to be, the polarity. So that does make sense in terms of what you described there. So I think we're quite cool now to maybe jump on to miking techniques. So we mentioned there about miking, so we're not going to move on to micing techniques, but maybe the transition into what we now see mid-side as today, or rather hear mid-side as today, and how we use it.

Ian Stewart: 11:20

Right, well, so yeah, I mean, I think people saw this for a long time, it was the micing technique, and then it kind of, you know, someone put it together like, hey, we can go the other direction too. If we can take a mid signal and a side signal and do these kind of mathematical operations to get left right, we could also just go back the other way. And actually the way we do that, if we're starting with just a left right stereo signal, if we want to get mid, we add left and right together, we take left plus right, and that gives us our mid channel. Yeah, right. And now that's something that we can work on again with EQ, with compression, with whatever it is. For side, we just take left minus right.

Ian Stewart: 12:04

And there's a little detail that I'm kind of glossing over, which is that there are some level shifts happening. So technically, when you add two signals together, they get louder. So oftentimes part of a mid-side matrix is you'll do a gain change as well, you'll compensate for the gain change so that things don't end up clipping or whatever, right. So you do left plus right and then divided by two, which really just means turn it down 6 dB, and so you kind of got to do these gain structure tweaks almost so that again you don't end up with clipping and so that you get equal level going through. But yeah, you know, basically left plus right, that gives you your mid channel, left minus right.

Ian Stewart: 12:42

Now he gives you your side channel, and and this is like I want to highlight another important here, another important point here, which is that a lot of times people will say mid and sides and put an S on the end of sides. It's not sides, it's the side channel. There's one channel and that represents your side information. Right, it's not and, and, but, point being, it allows us to start thinking about, you know, is our, is the content closer to the center or is it closer to the edges of the stereo field? And now we can EQ or compress or whatever it separately, depending on where that pan information is coming from. And there's a point where that gets a little complicated again and it gets misrepresented a lot.

Ian Stewart: 13:33

But I'm we'll, we'll come on to that and we've got some, some examples to help like think about that a little bit and hear it but but yeah, by by doing that right now we can again operate on different kind of portions of the stereo image, whereas, right, if you just had an eq that you eq'd the left channel on the right channel separately, you might kind of get this imbalance right. You could get a weird tonal shift. Sometimes that's the right way to approach stuff, but often it's not. Whereas with mid-side we can think about do we want to impact things that are further out on the stereo image or closer to the center and balance how we EQ those things?

Marc Matthews: 14:13

Yeah, it's interesting you mentioned that about when you add the left and the right together, we get the mid channel. That's correct right and then to get the sides. It's the left and right subtracted, Is that?

Ian Stewart: 14:23

right. Did I get that right? Yeah, yeah, exactly. Left minus right. Or again, the other way to think about that is take the left channel, flip the polarity of the right channel and add them together.

Marc Matthews: 14:34

Of course, yeah, yeah Right.

Ian Stewart: 14:35

Because, again, channel, flip the polarity of the right channel and add them together. Of course, yeah, yeah, right. Because again, in in audio subtraction is just flipping the polarity and and because it's a it's a bi-directional signal, it goes above zero and below zero. So how do you subtract? Right, it's just flipping the polarity and then summing them together yeah yeah, so with that I mean mid-side eq.

Marc Matthews: 14:51

I see a lot of I say, I say a of tutorial. I hear it referenced a lot online and people say they're using it and whatnot. What are the common sort of maybe not misconceptions, but errors? User error let's say that you encounter.

Ian Stewart: 15:05

Right. I think there's one big one right and it seems like almost and I've kind of alluded to it here, but I've tried to at least foreshadow that there's a little bit more going on right. But the big thing that you usually hear is people will say well, you can use the mid channel to get stuff that's right in the center of your stereo image right and affect that. And you can use your side channel or the sides, which, again, that's really not right. It's one channel, side channel is one channel, but you can use the side channel to change stuff that's at the edges of your stereo image. And that's kind of as far as people usually explain it, and it leaves this impression.

Ian Stewart: 15:44

I think, that if you've got a hard pan guitar off to the left and you want to make it brighter without impacting the vocal, you can just EQ that one way.

Ian Stewart: 15:53

Or if you want to make the vocal brighter without impacting the vocal, you can just EQ that one way.

Ian Stewart: 15:54

Or if you want to make the vocal brighter without impacting the guitar, you can just EQ the vocal that's in the middle one way and it's not going to affect the things that are out at the edges. And the fact is that that's an oversimplification. It is more nuanced and there's kind of there's crossover between those things and, moreover, because mid-side right, how we construct a stereo image from a mid-channel and a side channel is, like, very, very dependent on the polarity relationship of the negative side to the mid. If we start changing the phase, which again is related to polarity, we start changing the phase of those channels independent of each other. It impacts the stereo imaging. So that was a lot right. But those are to me those are kind of the things that often get glossed over but are really important and like audible, and again we've got some examples to highlight this stuff. But yeah, it's more nuanced than just center and edges and you do have to think about the phase response of whatever you're doing.

Marc Matthews: 17:10

Yeah. So just to recap that you mentioned, it's not so the, the idea that you mentioned about the if you've got a guitar that's hard left or hard right, thinking that you could use, you could use mid-side eq just to eq that guitar and not the vocal because it's on the left or the right. That is not that. That's an not an oversimplification, but that's a misconception. Would that be right?

Ian Stewart: 17:31

oh, yeah, a little bit and and again. There there's nuance to all of this and I want to try and ease into it and not just hit everyone totally over the head at first. But yeah, yeah, and we'll see. So I think maybe, I don't know, do you have other things that you want to maybe kind of talk through a little bit before we get into some of our audio examples?

Marc Matthews: 17:52

I think we should dive into the audio example. Should we do that? Yeah, I think that'd be good. This is the frontier I was talking about at the beginning of the episode audience listening. This is the first time we've done one of these technical conversations and we are going to be using audio. So yeah, let's do it. Let's do it.

Ian Stewart: 18:09

Yeah, so we've got three files that I want to play at first, right, and they kind of serve to illustrate this misconception a little bit, right? So we're going to start. We've just got a little guitar loop, just a little acoustic guitar loop, and it's going to pan from the left channel all the way over to the right channel. So at this point, if you're listening on a phone right or something that's more or less mono, headphones are great. This is definitely a time to put in some decent headphones. Or, you know, go sit in the sweet spot of you know, if you want to listen on speakers there, you'll get some of the same effects, right. But we're going to start and again, this guitar is just going to pan from hard left across center over to hard right, and then we're going to go and listen to that in the mid channel and the side channel and see what happens cool.

Marc Matthews: 18:57

So this is audio clip number one.

Ian Stewart: 19:00

Then let's give this a go Right so you can hear that starts over on the left side, pans slowly across, goes through center, ends up on the right. So the next thing we'll do is we'll do a mid-side encode right. So we're going to take this left-right stereo file and encode it to mid-side and then we can listen to just the mid channel and just the side channel. And so what I think a lot of people imagine happens is that. Well, actually, let me, let me ask you, and maybe you can, maybe you have an idea of what happens, but maybe you can pretend that you're your average person who hasn't thought too deeply about this and and try and guess what happens or what they. I don't know what you get, where I'm going with this Right.

Ian Stewart: 19:58

But, um, if we think about, okay, if we take this guitar and it starts off hard left and it goes through center, ends up hard right, if we listen to that and in just the mid channel, what's that going to sound like? Like, when it's panned hard left, are we going to hear it in the mid channel, or is it not until it gets to the center that we're going to start hearing it? What, like what would you think would happen? I think we're going to hear it in the mid channel, yeah, yeah, and and audience listening. I'm going to ask you I'm, I'm, I'm holding off a little bit here.

Ian Stewart: 20:27

I'm going to ask you to try and imagine what you think you're going to hear. Right, and if the common wisdom is that the mid channel is stuffed down, the center and side channel is stuff out of the edges. Like, what do we think we're going to hear? Is it going to fade from silence up to guitar as we get to the center and then back out to silence as we get further off to the right, or are we going to hear something else? So at this point let's go ahead and play it.

Marc Matthews: 20:50

This is number two right. I like this. This is number two. Yeah, let's do it.

Ian Stewart: 21:10

So, yeah, we hear it right away. Right, we hear it the whole time through and really all that happens is it gets a little bit louder when it gets to where it was panned, to the center again. This is this exact same guitar that pans from hard left to hard right.

Marc Matthews: 21:23

Just listening to the mid-channel representation of it yeah right, so it's just that it so I was gonna say is that because what you mentioned earlier about the, the summing of the, the, then, as it moves to the middle, we're getting that sum of the stereo and the mid channel we're summing it and that's why we're getting that slight bump, that perceived increase in loudness there.

Ian Stewart: 21:44

Exactly, that's exactly it, right? So if mid is left plus right we've said that a few times right, mid is left plus right, so it starts off and it's just in the left channel right. So we get the left channel, we hear that and that's our mid channel. And then, as it pans across and we get more and more level in the right channel, at some point we get to the center and we've got equal level in left and right and now those are adding together. It's making it louder. Some of this is obfuscated a little bit by the pan law in Pro Tools.

Ian Stewart: 22:15

When I did this, so the pan law almost makes it a little less obvious, but you can still hear it. It gets a little bit louder when it's in the center and both left and right are adding together, and then again as it gets further off to the right it kind of softens down a little bit again.

Ian Stewart: 22:32

Because you're just getting that one channel. Yeah, I love this. That one channel, yeah, I love this, this is great. Now, before we play the next one, this is the thing I want to highlight.

Ian Stewart: 22:37

If you think I want to EQ the vocal or some cymbals or the snare drum or something that's down the center, and you put a high shelf on it and you think, oh, I'm not affecting stuff out at the edges because I'm using the mid-channel Wrong, right, we can hear that stuff that's hard panned shows up in the mid channel and the kind of I don't want to say danger, but sure, danger, let's go with danger. The danger here is confirmation bias. If you think I'm just impacting stuff in the center, you might not even be listening for stuff out at the edges of the stereo field and you don't hear that you're changing it. But then when you send it off to your client and assuming you're working for a client or the rest of the world hears it and they don't know what you did right, they're not seeing the EQ moves, they don't care about what you did engineering wise, they just hear it as yeah, everything's a little bit brighter, yeah, um, so so you gotta, you gotta be aware of that and know that yes, if you're EQing the mid channel, that's impacting the whole stereo image and the thing about so the the, the kind of offset is that, yes, it's, it's impacting the stuff that's right down the middle a little bit more, but it's still going to make stuff out at the edges brighter or duller or whatever your EQ move is that you're doing.

Marc Matthews: 23:51

Yeah, this is this is really really good.

Ian Stewart: 23:53

This is really good.

Marc Matthews: 23:54

These examples are really highlighting this. This is fantastic, yeah, right.

Ian Stewart: 23:57

So the next thing yeah, so let's play that third file mark. So this is just if we listen to it in the side channel, Right.

Marc Matthews: 24:05

Here we go. Yeah, thank you.

Ian Stewart: 24:40

Cool, yes, Interesting, Right, but as soon as we get panned even the tiniest bit away from dead center, it shows back up in the side channel, Right? So this, this tells us practically, if we think about this. This tells us our side channel. If we're EQing our side channel, yeah, we're getting stuff predominantly that's hard panned, but really we're getting at least a little bit of everything until it's dead center. So really, if you do want to affect things you know you want to change stuff that's hard pan but you don't want to impact the vocal your way to do this is going to be to EQ the sides. I did it, I just did it. Eq the side channel, it's not sides.

Ian Stewart: 25:22

Ian it's side right, You're going to EQ the side channel and by doing that you can avoid impacting stuff that's right down the center. But still, even stuff that's just a little bit off center, just panned out, you know, 5%, 10%, 15%, you're going to have a little bit of an impact on that. So, again, you want to listen out for those things when you do them.

Marc Matthews: 25:41

Yeah, yeah. So this is great because it's just highlighting the point that it's not as straightforward as the like we mentioned earlier about if you want to affect the mid or the side channel. It's not as straightforward as okay, I'll just use this mid-side EQ and I'll affect the side and the mid. You really have to think about the whole stereo field and how you're affecting that and that confirmation bias. What you mentioned earlier I think is really important because if you are just EQing the mid channel, that confirmation bias, as you mentioned not that you're doing it actively, but you are effectively, you're thinking psychologically, you're gonna be disregarding the sides because you think you're only affecting the mid. So that's really really important, I think, for the audience listening yeah, really important, I think for the audience listening yeah, really important, Like you.

Ian Stewart: 26:27

Really, you absolutely have. If you're doing mid-side EQ, no matter what you think you're trying to do, you absolutely have to listen across the entire stereo field for effects or side effects of what you are doing.

Ian Stewart: 26:41

And it's like, you know, we've all been there, right, like we've all sat there and cranked a knob and, like you know, finesse that EQ to perfection and then we realize it's in bypass or we're on the wrong channel, and yet we were absolutely positive. We heard a change, right, Like I do that, I'll do that, I'll be sitting here at my desk, I'm working, you know, I've got my main EQ and I'm like, oh yeah, that's perfect. Yeah, half a DB, three quarters, yep, that's great. Oh, it's not even patched in, right, and yet I'm completely convinced that I'm hearing something. So it's so easy to do when, when you think you're doing something and that's what you're listening for, that's what you'll hear. So it's just it's really about for me, it's about understanding that you are impacting more than just one specific part of the stereo field and then being able to know what that sounds like and listen for it accordingly.

Marc Matthews: 27:28

Most definitely. So we've got some other audio examples here. So should we rattle through those as well, because this is great stuff.

Ian Stewart: 27:35

I'm loving this, yeah, so let me give maybe a little bit of a preface for this, right. So before, I kind of alluded to the fact that phase is going to be important here and phase is of course, different than polarity but if the polarity of the side signal in relation to the mid signal is part of how we reconstruct our stereo image and the timing between those and the polarity relationship is like it is at its core, how we get left and right out of mid-side, if we change the phase response of either the mid-channel or the side channel independent of one another, that's going to start impacting stereo placement.

Marc Matthews: 28:22

Yeah.

Ian Stewart: 28:22

And and one of the really the way I got into thinking about this and the thing that started me down this whole path was it goes back to the some of the idea behind Baseline Pro, honestly, which for a long time you would kind of hear the common wisdom about monoing the low end was just put a high pass filter on the side channel filter on the side channel.

Marc Matthews: 28:44

Yeah, right, yeah.

Ian Stewart: 28:44

Mono your low end, put a high pass on the side channel, crank it up to a hundred Hertz or 400 Hertz or whatever, I don't know. Some you know you hear insane numbers about it, but the the issue there is that high pass filters or low pass filters, any, any sort of pass filter like that, though, has a pretty severe phase shift associated with it. So a six DB per octave high-pass filter um has a 90 degree. So it goes from zero degrees when you're in the pass band where nothing's impacted, right, and then as you go through that crossover or that turnover frequency, like you know, your cutoff frequency, um, and below that the phase shift increases to 90 degrees. Yeah, that's 6 dB per octave.

Ian Stewart: 29:27

At 12 dB per octave your phase shift goes up to 180 degrees. At 18, you go through 270, 270 degrees, and so for every order of magnitude that your filter steepness increases, you get an additional 90 degrees of phase shift. And what can happen if you start using really steep filters? You'll get through a full 360 degrees of phase and more right, you'll just keep having these phase shifts as you go through that crossover region. So if we pause and think about that for a moment in relation to stereo placement. In relation to stereo placement here, mark pop quiz Okay, so what's the difference between a hard-panned left signal and a hard-panned right signal if we're thinking in terms of mid-side?

Marc Matthews: 30:19

The difference between the two?

Ian Stewart: 30:21

Yeah, how would we describe what the difference between a hard-panned left and a hard-panned right is if we're looking at it from a mid-side?

Marc Matthews: 30:30

perspective, just where they're placed in the stereo field. Am I being too simplified?

Ian Stewart: 30:36

A little yeah, yeah, how about? If we think about the polarity thing?

Marc Matthews: 30:41

There's a hint, they're opposite. Yeah, they're opposite.

Ian Stewart: 30:45

Right. Yeah, yeah, it's opposite polarity that's. Or if we superimpose that on phase, it's 180 degrees of phase. That is the difference between left and right. So if we do something to the side channel that shifts its phase 180 degrees, now that's if something that was on the right is now on the left and vice versa. So you literally flip the stereo image right, and even when it's less than that, even if it's 90 degrees, it doesn't flip it but it starts to smear it out. So something that was at one very well-defined point right, hard panned exactly there, now gets kind of spread out across an area that's over. You know, if something started on the right now, it gets spread out across an area that's over.

Ian Stewart: 31:26

You know, something started on the right now it gets spread out a little bit right okay so the phase response of our eqs affects both stereo positioning and width on a kind of microscopic level. Right, it can smear stuff across an area rather than having a defined point, which, let's be honest, sometimes that's cool, right. Sometimes you want to make something a little wider, you start with a mono signal and you want to make it a little wider. That's a way you can get there. That's really cool. But other times, if you want something positioned exactly in one spot, you can end up, if you do some weird mid-side stuff, if you're not careful, it can end up kind of spread out and a little diffuse. Wow. So, yeah, right, so the phase response of our equalizer is really important and has a big impact on the sound that we get out of it.

Ian Stewart: 32:18

Yeah, so that's where these, these next um audio examples go. So we're going to take same same guitar loop, we're going to start and we're just going to pan it off a hard left right and just listen to it without any processing. Okay, uh. And then we're going to apply a um, uh, which, which is this Sorry, I've just got to check my my notes.

Marc Matthews: 32:45

here I'm going to say number five.

Ian Stewart: 32:47

Yeah, yeah. So that's five, right, let's, let's start there. Yeah, let's go ahead and start there. Let's do that, right. So that's, that's our baseline, okay, um, and then what we're gonna do is oh, I see, I've got, I'm looking at this, I was, I'm looking at my notes and I've got my images in here too. Um, so then we're gonna do a four and a half dB boost at 2k and we're going to do it in the mid channel. Yep.

Ian Stewart: 33:32

And for the first one we're going to use I think sorry, just double checking we're going to use linear phase. Okay, Okay, right, so we'll start with linear phase. So this is a linear phase. Eq mid channel boost four and a half dB at 2K.

Marc Matthews: 33:50

And this is number six.

Ian Stewart: 33:51

This is number six. Yeah, audio file number six.

Marc Matthews: 33:53

Yeah, let's do it. Here we go.

Ian Stewart: 34:07

Right, right. So that's maybe what you would expect if you're EQing the mid channel. So, of note, it's gotten a little narrower. At 2k. Right, we've kind of pulled 2k towards the center because we're boosting the mid channel, that pulls it towards the center. But the rest of it, the low end and the real high end, are still kind of off far left.

Marc Matthews: 34:34

Yeah.

Ian Stewart: 34:35

Okay, now next up, we've got number seven. This is the same, exact same boost. So I think I did these with um. If I'm remembering correctly, I did them with Pro-Q 3, which you can just switch between minimum phase and linear phase. So I took the exact same EQ setting and just switched it to minimum phase, and now let's take a listen to that. So this is number seven. Here we go. Number seven, yeah.

Ian Stewart: 35:12

Right, and then number eight just goes from left, panned with the EQ out, Then after like a half a bar, or a bar and a half or something, goes to linear phase and then to minimum phase, so we can hear them all back to back. So let Mark, let's maybe play number eight as well, and then we can talk about this a little. Cool, Right, so it's. I mean, I guess the first thing to say is on a mono source like this, maybe that's kind of cool, Maybe the fact that it spreads the image out and kind of smears it a little bit is a cool thing, and you want to utilize that. So in a mix that could be a really cool thing. But I think I hope you'll hear, especially and honestly, I will say, if you are listening on speakers, depending on your room and a lot of things, this may be more or less obvious In headphones. I find this like this isn't subtle, Like the EQ, part of it is the same, but the imaging changes drastically. Um, and and so it, especially in a mastering context or even on a bus, right, If you're applying something to your master bus, your two bus, or or an instrumental bus or whatever um, the difference between linear phase and minimum phase is big, especially if you're using something like a high pass filter, especially right, Um so. So you really, like, you got to think about that and and what I usually say and I this is kind of what I explained to my students at Berkeley, though is like it's not that one is right and the other is wrong, it's just that they are very different, and so it's really worth having an EQ where you can audition those differences without having to try and replicate settings and whatever.

Ian Stewart: 37:06

Right, If you can just switch between minimum phase and linear phase and listen to the differences. So Pro-Q 3 is great to do that. Right, it's just a drop down at the bottom. You can switch that. The Ozone EQ actually has a mixed phase mode, so if you put it in digital mode, but then there's a little window that pops out on any given EQ filter you put in and you can slide the phase anywhere between completely linear phase and completely minimum phase.

Ian Stewart: 37:33

So being able to do that and really you know hear how the phase relationship or the phase setting of that channel impacts the overall stereo image and you know move stuff or smear stuff or you know changes, it is really important. And again, sometimes you may want one, sometimes you may want the other, but it's one of those things that once you know to listen to that and you know what it sounds like, you're going to hear it everywhere because we've been doing it for so long. But we now have the tools to kind of be a little more nuanced and precise with it, and so, yeah, you know we can think about what the change we really want to impart is and do just that amazing.

Marc Matthews: 38:14

I love this. So just to recap then, for for the audience listening, you mentioned there about we've got these different slopes, different order slopes. You've got 60b, 12db, 18, 24, so forth. So when they're using these slopes, you're at theseoffs, we're having different. It's affecting their polarity, was that correct? Have I got that right? And then it's smearing the phase.

Ian Stewart: 38:33

Yeah, it's affecting the phase. It's affecting the phase shift right. So polarity is one or the other. It's zero or 180,. Right, it's a binary thing, but it's affecting the phase and the steeper a slope of a filter, the greater the phase shift. But like these filters that we just listened to, those were bell filters in the mid range right, that's not even a high pass, that's just a bell filter.

Ian Stewart: 38:53

And this is another thing that I will people see people say is like, well, yeah, sure, for a high pass, um, or a really steep shelf, maybe it makes a difference, but for bell filters it doesn't really matter. But like that was a, I'm pretty sure I could go back and look at my settings. That's a cue of one.

Ian Stewart: 39:09

That's not a particularly sharp bell filter at 2k and we can hear that there's really a very appreciable differences, difference in the stereo imaging. Um, right, so it really, whether it's a shelf or a high pass, or a low pass, or a bell, they, they all have some phase shift associated with them. And honestly, I mean, the phase shift with a with a bell filter is subtle. It's it's not a lot. It's not a lot of degrees in terms of, you know, measuring phase shift in degrees, but it's enough that it starts to smear the image.

Ian Stewart: 39:41

Um and the interesting thing I mean the interesting thing about a bell filter, if you actually look at the phase plot, basically. So if you do a boost, I'm not going to remember which way goes which, but as you're coming from below that frequency, the phase will go one way.

Ian Stewart: 39:54

So it'll trend, say positive, it'll go to plus five degrees and then at the center frequency of that bell it'll curve back down. It'll be right at zero degrees there and then as you extend above it it'll go swing below. So it'll curve back down. It'll be right at zero degrees there and then as you extend above it it'll go swing below. So it'll go negative five degrees and then come back to zero, right. So like, if we think about that again in terms of stereo relationships and mid side, it's pushing things in opposite directions and that's why you get some of that smearing, right. It's kind of pushing the relationship in different directions around the center frequency. So yeah, really, it's in different directions around the center frequency. So yeah, really it's. You know, any sort of EQ that you're doing, whether it's a high pass or a shelf or a bell or a whatever. It's really worth auditioning both minimum phase and linear phase to understand you know how that's going to change things?

Marc Matthews: 40:43

Yeah, excellent stuff. So you mentioned that about linear phase EQ. So I mean, I don't know if you'd be able to go in into the great detail. How is that EQ? How is that set up differently to sort of negate that linear phase, that phase shift element of the EQ moves that we're doing?

Ian Stewart: 41:02

Yeah, this is one of those things that that's like it's, uh, it's very tough to just talk about and so much of it comes out of like the math of how it works.

Ian Stewart: 41:11

Yeah I thought, but but essentially essentially um with minimum phase. Part of what's happening is signals different, different frequencies are getting delayed slightly different amounts. Yeah, yeah, and that's actually what causes the effect of EQ, right, the fact that different frequencies have different amount of phase shift and get delayed slightly different amounts and then add back together with the original signal and kind of add or subtract in different ways. That's kind of what creates EQ curves and shapes, and so in linear phase, linear phase is something that we can only really do in digital um, at least with respect to EQ um, other than one maybe very esoteric design that I don't think has ever been implemented, um. But but what it's basically doing is it's compensating for the that delay, right? So digital we can. So that's why a linear phase EQ has some latency to it, whereas a minimum phase EQ can be almost zero latency or completely zero latency, right. So linear phase EQs have some latency associated with them and that's so that they can basically sync stuff back up.

Ian Stewart: 42:23

Another way to think about it that's a little bit weird is there is actually there is a way you can get linear phase out of an analog EQ, and that's if you run the signal through it forward and then you run it back through it backwards. The thing there is that you get double the EQ boost, so you have to. If you think you want to do a 2dB boost, you just have to do 1 dB initially. You go through forward once and then backwards, and so the phase shift. It goes one way when you're going forward and then the other way when you're going backward and you get a linear phase response. But that's kind of equivalent to what's happening, right. I don't know that it's strictly convolution that's used, but it's basically doing this and it's using these short time delays and latency to be able to do that.

Marc Matthews: 43:12

Interesting this. When you mentioned there about delay in my head I was like it took me back to when I was doing my degree and I remember when we were using a program called Max MSP and we were creating EQ, creating eqs using delay lines yeah, yeah, exactly, yeah, yeah.

Ian Stewart: 43:29

That is literally all eq is. It's, it's delay, and a little bits of feedback and summing things together yeah yeah, and basically it, like all eqs start with very simple amount, types of delay and recombining and shapes, and then you can build upon them and iterate them to get bells or to get high passes or low passes or shelves or whatever. Or now we have these more esoteric shapes band shelves and flat tilts and stuff like that.

Marc Matthews: 43:57

But they all start with very simple, very simple principles yeah, yeah, that's amazing because, yeah, like I said that, it just in my head.

Marc Matthews: 44:05

I was like ah, light bulb moment I remember. Now I remember the delay lines, yeah, I do.

Marc Matthews: 44:10

It's all coming back to me now. It's been so long it's probably been almost 10 years that I was having these discussions and going through all this stuff. But, ian, this is amazing. I'm absolutely loving this. I'm going to have to do this more when we're having these audio examples and these technical deep dives.

Ian Stewart: 44:24

I mean, I'm really glad you're up for it because you know, this is one of these things that this is like my favorite topic, honestly my favorite audio topic. I just because there is really cool stuff that you can do with it, and I think it just it's. Once you really understand it, it almost unlocks more possibilities, right? It seemed at first it might seem like it's limiting, like I'm saying there are all these constraints and all these things, but once you understand what's happening, then you can do things like invent Baselane, pro or whatever, right, like it's the fact that I thought so deeply about this stuff for a while that allowed me to come up with some of these things and ways to manipulate stuff. And yeah, you can do some really cool stuff, stuff. And yeah, you can do some really cool stuff. And so I mean just I think one. I mean, unless you have other other tangents, you want to go down or whatever, I'm happy to chat as long as you want. But one thing that I I also a way I present this to my students and I think that's a helpful way to think about mid-side in general is, rather than think of it as affecting different areas of the stereo field, if you think of it as X specific width manipulation, that starts to make more sense, right? So if you're using an EQ, it's like frequency specific width manipulation.

Ian Stewart: 45:45

And especially if you do it with linear phase, right, if you boost the mid channel, if you boost a frequency in the mid channel, you're going to make that frequency range narrower, right, because it's going to pull stuff in from the sides. You've got more mid information. So that seems to a mid side thinking device. It's there's no thinking going on, whatever, it's just a matrix. But it seems like it's more towards the center, right? If the level of the mid channel is stronger compared to the sides, that seems like it's coming more from the center. If you boost a frequency range out in the in the side channel, that seems like it's coming from one side or the other. That's going to pull it more towards the sides. So it's kind of more like frequency specific width control than really EQing just the center or the edges. And then if you apply that to other things like compression, right, we've been talking a lot about EQ, but mid-side compression is pretty popular too, right? The way for me to think about mid-side compression is as level specific width control. So if.

Ian Stewart: 46:42

I'm compressing the mid channel, that means I'm making I'm turning down loud things right. If I'm compressing I'm turning down loud stuff, that means loud events that are based in kind of the middle of the stereo image are going to make the whole stereo image wider, right? So if I get a really loud snare hit that's right in the middle and there's a whole bunch of reverb and I compress the mid, it's going to kind of emphasize the reverb out on the edges.

Ian Stewart: 47:10

Right If I compress the side right, it's going to emphasize the center, so it kind of with with with compression, it kind of goes the an opposite direction, right, but by compressing the side you're making things. When it gets louder out at the edges, you're making things narrower. Um, so yeah, if you think of it in terms of these methods of width control that can either be dynamic or frequency based or whatever um, that to me starts to make more sense Once you start getting to things like saturation and whatever. It breaks down a little bit and weird things can happen there. But very cool things can also happen there.

Marc Matthews: 47:47

Yeah. So just to reiterate, then, what you said there about we've gone back to EQ If we've got that mid channel and then we're doing that EQ on that mid channel and we're boosting in the mids, it's pulling in information from the side. Was that the way you described it? It's pulling in information for the side and then it's making it narrower. Is that correct?

Ian Stewart: 48:04

Yeah, it's not so much that it's pulling in information from the sides, right, the information has to be in the mid channel to start with. Yeah, yeah, but assuming you're operating on a more or less broadband signal, right, that's roughly even frequency distribution from, you know, 20 Hertz to 20 K or something, right, when you boost the mid channel, that now the mid channel is louder compared to the side channel in that frequency range, which means the image gets narrower there. Yeah, right, so it's not so much that it's actually taking information from the side channel and bringing it in, it's just changing the panning of it and making the image narrower at that frequency. And again, if you've got a bell filter right, there's a slope to it and the width follows that slope.

Marc Matthews: 48:57

Yeah, that does make sense. And I realized I said sides as well, so I it's so easy to do. I mean, I did it earlier too, so yeah, yeah I am going to police myself now going forward and make sure I say side and then what you said there about eq. I appreciate we're not really delving depth deep into eq today. But no, sorry, compression compression, yeah, yeah compression is effectively kind of doing the opposite.

Ian Stewart: 49:20

Yeah. Well just because, right, you're turning down loud stuff when you compress sort of yeah, yeah, depending on the attack time right, so you can make.

Ian Stewart: 49:28

If you have a long enough attack time, you can make the initial transient stand out more and then the tail of it. You bring that down and it feels wider. There's super powerful stuff you can do here and you can start to unlock. But yeah, again, you just got to think. I mean, the key here is that if you're making the mid-channel quieter, it's going to get wider. If you're making the mid-channel louder, it's going to get narrower. I like that.

Ian Stewart: 49:54

Side is the opposite. Right, If you're making the side channel louder, it's going to get wider. If you're making the side channel quieter, it's going to get narrower. Fantastic, I love that. You can think about how am I impacting the level and to what channel, and then you use your little decoder ring right Mid-channel louder, narrower, mid-channel quieter, wider Side-channel louder, wider Side-channel, quieter, narrower. I should make like a bracelet that has that or something. Yeah, definitely.

Marc Matthews: 50:36

Yeah, most definitely, I think. So I think that's great Immediately when you mentioned there about compression and you said about you could have the slower attack time so you could have that initial transient hit and then you could just have this slightly wider and make it sound wider, and I can hear how that would work on something like a snare or something you get this whack or a drum buss or like rooms, right, a room pair and you put that in MS and do some compression on that where you can get some punch from it and not have the initial hit feel too roomy and big, but then, as you know, that dies down.

Ian Stewart: 51:12

You bring the side channel up in level and get more room around it. Yeah, I mean, there's like there's super powerful, cool stuff you can do once you really start to. I think it's one of those things that feels very easy to just kind of like fiddle with and maybe you get it and maybe you don't, and maybe you're happy with what happens and maybe you're just like this sucks and delete this plugin right, yeah, yeah.

Ian Stewart: 51:30

And once you really understand what's going on, my feeling is it allows you to be much more directed and focused with how you apply it and like you can. Now you can, rather than just fiddling things. You can have a goal in mind and be like yeah, ms is the tool, this is how I'm gonna do it. Here we go, boom, that's what I wanted. Awesome, yeah, right. So definitely.

Marc Matthews: 51:51

I know I'm gonna be doing over the next, over the weekend, then I I'm gonna be trying all this. I really hunkered down in the studio just thinking right. This is the goal now, Specifically, obviously with the EQ, but the compression side of things as well. I'm 100% going to be getting my teeth into that more.

Ian Stewart: 52:08

Yeah, I mean definitely. I really strongly suggest for you, for anyone that hasn't done this, grab an EQ that can flip between linear and minimum phase and just start playing with that and really listening. If you don't have a room that has really great stereo imaging in it, headphones are by far the easiest way to hear this at first. But if you've got a setup that has really well-defined stereo imaging and kind of not a lot of early reflections and first reflections, you can definitely hear the stuff in speakers too. Lot of early reflections, um and first reflections you can definitely hear the stuff in speakers too. But play with that and just start to understand what it sounds like and what the different options are, because they're they're both valuable. They are absolutely both valuable, um and very different most definitely.

Marc Matthews: 52:53

Yeah, uh ian, this has been fantastic, excellent stuff. Thank you so much for jumping on and sharing your wisdom on this today my pleasure, thank you.

Ian Stewart: 53:00

Thank you for fantastic, you know, having a platform where, where we could do this, this is I. I'm excited just to have this, to be able to say, you know, when someone asks me about it, rather than rant at them for an hour, say here go listen to this plus their audio examples yeah, yeah, most definitely and uh, as I mentioned, I cannot remember if I said this at the beginning or I said it off air, but it is a topic of conversation that's uh regularly.

Marc Matthews: 53:25

I'm asked about is I think I did say this on air is the use case for mid-side?

Marc Matthews: 53:31

yeah you compression, whatever it may be. So hopefully well, I'm sure now this is giving the audience some really good information in terms of the decisions they make going forward and how they explore it as well, and so just a platform to kick on and just really like properly get your teeth into it. So it's been amazing. Ian, this is a good opportunity now for you, maybe if you want to share where the audience can find out more about you, if they want to look you up. Possibly.

Ian Stewart: 53:58

you mentioned mastering, maybe they want to send some music your way for mastering, or yeah, absolutely um, I really, you know, I generally refer people to my, to my website that's flowtownmasteringcom dot com, f-l-o-t-o-w-n. So sort of like bowtown, but not, um, yeah, and, and there you can, you know, I've got links to blogs and, uh, there's an intake form or a contact form and you can read about kind of what I do and here's some examples and stuff. Uh, but then also, you know, like, if, uh, you know you like pictures of fall foliage and comets and Northern lights. My Instagram is Ian Stewart music.

Ian Stewart: 54:32

Um, you know, I post very little music related stuff. Someone sent me a vinyl, vinyl of a project that I worked on the other day, so I posted that because that was kind of cool. It's not that often that we get physical representations of things that we work on anymore, with the dominance of streaming, so I posted that. But a lot of times I'm posting pictures of, you know, me and my dogs on a hike. Or the Northern Lights the other night were pretty incredible. Or last night we went up to the top of a mountain nearby and looked at the comet and uh, so, yeah, that's a fun way to get in touch too, but that and and flowtownmasteringcom are are the best brilliant.

Marc Matthews: 55:08

I did see that on instagram. I saw that the uh the vinyl. It's cool, isn't it, when things like that happen.

Ian Stewart: 55:13

Yeah, I actually still haven't listened to it. I might do that afternoon. Unwrap it and go take a listen.

Marc Matthews: 55:18

Fantastic, Ian. This has been an absolute pleasure. It's been great having you on the podcast again. Likewise.

Marc Matthews: 55:24

Audience listening. Obviously, you've listened to this episode, but do go check out episode 139 as well if you haven't listened to that one yet, where we talk about AI and AI as I mentioned at the beginning assisted music production as well. So that's a really, I think just the episodes around 139 and that sort of 138, 140. There's a lot of AI-based music production stuff going on there, so do go check those out. I'm waffling now, ian. It's been an absolute pleasure and I hope to catch up with you again soon.

Ian Stewart: 55:52

Absolutely. Thank you so much, Mark.