bouldersoundguy
Well-known member
You can directly convert X/Y recordings to M/S signal format and do the same width adjustments. Actually you can do it on any stereo signal, but there's no guarantee that it will collapse to mono gracefully.
DrewPeterson7
Sage of the Order
Where my issues seem to be coming from (and I'm 99% sure it wasn't the mics, they weren't physically touching but they were close) was the duplicated and phase-reversed track off the figure-8 mic. I was following this:
Introduction To Mid-Sides Recording
M‑S miking is easy to set up, offers plenty of options at mixdown, and has some unique advantages over other stereo arrays. So if you've not yet tried it, why not give it a go?
...and then made a manual copy of the figure 8, panning the pair L/R, and flipping phase on one of them. Definitionally, these collapse to zero in mono, and while they were played back around the mid mic, by the time they were loud enough to be audible, I could hear pronounced hollowness as I moved my head around in the room and the out-of-phase tracks became more or less, well, out of phase.
The most logical explanation is I did something wrong, and if so I'd love to be told where I went wrong. But, these days, when I want stereo acoustic guitar, unless there's some reason where this wouldn't be appropriate (a solo acoustic performance of some sort, say), I generally just double track it. I haven't had any trouble with phase issues in multi-mic approaches when I have gone that way, though, so M-S was surprisingly, well, bad, for me.
bouldersoundguy
Well-known member
I just downloaded your tracks with your voice moving around. I put them in my DAW (Vegas Pro 18) and did the usual copy/invert deal with the CAD track. It works fine except that there's a lot of LF bleed. That is, the CAD is basically omnidirectional at lower frequencies (which is a normal trait of directional mics). So when the source (your voice) moves off to one side, the lower frequencies remain in the far side of the stereo image. The solution is to apply a high pass filter on the CAD pair (I used a submix bus) at around 200 Hz.
TalismanRich
Well-known member
Where my issues seem to be coming from (and I'm 99% sure it wasn't the mics, they weren't physically touching but they were close) was the duplicated and phase-reversed track off the figure-8 mic. I was following this:
Introduction To Mid-Sides Recording
M‑S miking is easy to set up, offers plenty of options at mixdown, and has some unique advantages over other stereo arrays. So if you've not yet tried it, why not give it a go?www.soundonsound.com
...and then made a manual copy of the figure 8, panning the pair L/R, and flipping phase on one of them. Definitionally, these collapse to zero in mono, and while they were played back around the mid mic, by the time they were loud enough to be audible, I could hear pronounced hollowness as I moved my head around in the room and the out-of-phase tracks became more or less, well, out of phase.
The most logical explanation is I did something wrong, and if so I'd love to be told where I went wrong. But, these days, when I want stereo acoustic guitar, unless there's some reason where this wouldn't be appropriate (a solo acoustic performance of some sort, say), I generally just double track it. I haven't had any trouble with phase issues in multi-mic approaches when I have gone that way, though, so M-S was surprisingly, well, bad, for me.
Listening to the side channels alone will give a horribly hollow sound for anything not on either side. That's where the mid mic comes in. But I find that you need to play with the mix to get a nicely balanced sound. Once you do I love the way a M/S sounds. It also makes a difference how the rest of your room sounds. Something with a lot of reflection can sound even more hollow.
Steenamaroo
...
+1 to that ^.
It sounds like you've just got the side channels turned up too high.
It begins to sound unnatural at a point.
The effect is, of course, more pronounced the close you are to the microphones,
as is the case with any stereo recording technique.
You can use it in close proximity, if you want, but again, there's a point at which it becomes unnatural sounding.
It sounds like you've just got the side channels turned up too high.
It begins to sound unnatural at a point.
The effect is, of course, more pronounced the close you are to the microphones,
as is the case with any stereo recording technique.
You can use it in close proximity, if you want, but again, there's a point at which it becomes unnatural sounding.
DrewPeterson7
Sage of the Order
ok, but that sort of cuts against the view that M/S offers better mono compatibility, no?
Again, I didn't spend more than a few minutes playing with this, had a bunch of mics and stands out anyway so set up a mic config, tracked, duplicated and flipped one, bussed them together, and then slid sliders around while listening for a few minutes... but to me, it seemed like I had the choice between either "sounds pretty much mono," or "stereo, but very unstable with audible phase issues as I move my head around.
It would seem like better mono compatibility would mostly be a product of the side channels nulling and leaving just the mid audible. Which makes it debatably better, since the stereo mix is going to have an entire additional microphone's worth of sound in it that's simply not there in the mono.
In any event, in a rock context, I've found simply doubly tracking with different mics and positions works better for ME, for the sounds I'm after, than any single-performance stereo technique I've tried, but I was just legitimately curious if somehow I'd made a mistake setting up a M/S array and mix, because it sounded, well, unusable.
Steenamaroo
...
As I say, there are no phase issues.
There can't be phase issues as all the sound is arriving at both microphones at the same time.
It sounds like maybe you just aren't a fan of mid/side, and that's fair enough.
It's probably best suited at a reasonable distance, maybe capturing a small group or something like that.
That's not to say you can't use it for guitar or whatever but, as I mentioned earlier, the closer you are to the setup the more unnatural the thing is likely to become.
You're 100% right about why it collapses perfectly to mono - The sides cancel out leaving the cardioid centre microphone,
so mid side in mono is literally just a cardioid mic in the middle.
There can't be phase issues as all the sound is arriving at both microphones at the same time.
It sounds like maybe you just aren't a fan of mid/side, and that's fair enough.
It's probably best suited at a reasonable distance, maybe capturing a small group or something like that.
That's not to say you can't use it for guitar or whatever but, as I mentioned earlier, the closer you are to the setup the more unnatural the thing is likely to become.
You're 100% right about why it collapses perfectly to mono - The sides cancel out leaving the cardioid centre microphone,
so mid side in mono is literally just a cardioid mic in the middle.
rob aylestone
Moderator
We never found out what quesne was recording. All the sound files are a single person speaking. He said originally he was recording his daughter. Was this her playing a guitar, or the two of them singing? We never actually heard any music, so weve been talking about recording one person’s voice. I have out a few thoughts together making assumptions so forgive me if there are no multiple participants, or guitars or anything stereo.
M/S is a technique that is at its best at a distance, in excellent sounding spaces, but where at the recording session, you dont know how wide you need it to be? It has saved my bacon in churches loads of times where unexpected things happen, or rehearsals are too short to do a recording and have time to check width. All my experiments in small rooms with it are not so good. The snag being on something like a piano or a guitar, you end up with the front facing mic aimed at the sound source and the side mic missing the main sound source and facing empty space or walls. Once you start recording a delayed sound reflected back from a wall, it messes with the matrixing, creating the weird effect that makes the apparent position of things move far more than they really are. In a choir recording, for example, the angle between furthest left and furthest right singer can be as wide as maybe 120 degrees, simply because of where they are, and where the mics have to be. The usual coincident techniques produce a hole in the middle where those singers are in the two mic’s nulls. Move them together and you lose the edges. More distance produces too much reverb to signal and robs clarity. M/S works really well. Many of my recordings feature those people with very distinct voices, and annoyingly, M/S reveals their exact position. In my studio, I dont consider a guitar to me an instrument with width, even when the player bangs the wood for percussion. One end is tinny and fretty, the other deep and boomy. Two mics lets you capture both sounds and blend them later, with, if you want, some panning. I firmly believe this is not ‘stereo’, just a nice technique. M/S doesn't do the same thing.
Maybe here, all the issues are just down to it not working the way we expect. The main mic still has the old problem, sound hole or neck? The actual width of the instrument or voice is minimal. Especially the voice. It has zero width. Voices just have placement in the stereo field.
M/S is a technique that is at its best at a distance, in excellent sounding spaces, but where at the recording session, you dont know how wide you need it to be? It has saved my bacon in churches loads of times where unexpected things happen, or rehearsals are too short to do a recording and have time to check width. All my experiments in small rooms with it are not so good. The snag being on something like a piano or a guitar, you end up with the front facing mic aimed at the sound source and the side mic missing the main sound source and facing empty space or walls. Once you start recording a delayed sound reflected back from a wall, it messes with the matrixing, creating the weird effect that makes the apparent position of things move far more than they really are. In a choir recording, for example, the angle between furthest left and furthest right singer can be as wide as maybe 120 degrees, simply because of where they are, and where the mics have to be. The usual coincident techniques produce a hole in the middle where those singers are in the two mic’s nulls. Move them together and you lose the edges. More distance produces too much reverb to signal and robs clarity. M/S works really well. Many of my recordings feature those people with very distinct voices, and annoyingly, M/S reveals their exact position. In my studio, I dont consider a guitar to me an instrument with width, even when the player bangs the wood for percussion. One end is tinny and fretty, the other deep and boomy. Two mics lets you capture both sounds and blend them later, with, if you want, some panning. I firmly believe this is not ‘stereo’, just a nice technique. M/S doesn't do the same thing.
Maybe here, all the issues are just down to it not working the way we expect. The main mic still has the old problem, sound hole or neck? The actual width of the instrument or voice is minimal. Especially the voice. It has zero width. Voices just have placement in the stereo field.
DrewPeterson7
Sage of the Order
Maybe I'm misspeaking, but when you take something, duplicate it, and flip polarity, isn't it now perfectly 180 degrees out of phase with itself, so it cancels to a null? Maybe I'm using the wrong term here, but that's what I'm trying to say - in the mix, you have a pair of "side" tracks that sum to zero when recording something M/S with a figure-8 mic capturing the sides.
If so, then yeah, I'm not a fan - it either basically sounds like a mono recording because your side tracks are low enough to be essentially inaudible, or they're high enough to be heard and if you move with respect to the speakers you're listening on you get these weird hollow sounds as the side tracks progressively move in and out of cancelation. I can't really understand why someone would want that effect. Probably sounds great with headphones, but play that back into a n acoustic space, and it's a hot mess.
Unless there's something I'm not understanding here, of course.
bouldersoundguy
Well-known member
That's not how it works. It's not moving "in and out of cancelation" in the sense of phase induced comb filtering. The only cancelation is between the two versions of the side channel, which is 180° out at all frequencies. The cancelation is the same at all frequencies rather than more or less depending on frequency. When done correctly, the stereo image is purely amplitude based.
It's important to distinguish between phase and polarity, two related but very different phenomena. Phase, in the specific sense, is a product of a time difference between two signals that have content in common. Polarity is the relationship between the positive and negative aspects of the waveform, whether that's pressure, displacement, voltage or whatever. Phase discrepancies will cancel different amounts (and add different amounts) at different frequencies, depending on the amount of delay. Polarity inversion will cancel equally at all frequencies, thus no "phasey" sound. If there's a "phasey" or hollow sound, it's probably from room reflections or some other source. Also, if the mics themselves don't have essentially identical phase response, that could be a problem.
The first image illustrates polarity. All the differences are restricted to the amplitude domain (vertical dimension on the timeline). If they are mixed together at the same level they would cancel out 100% at all frequencies.
The second image illustrates delay. All the differences are restricted to the time domain (horizontal dimension on the timeline). The result is that most of the frequencies of the right channel will be out of phase with the left channel, to a degree depending on the amount of delay and the frequency in question...
The second image illustrates delay. All the differences are restricted to the time domain (horizontal dimension on the timeline). The result is that most of the frequencies of the right channel will be out of phase with the left channel, to a degree depending on the amount of delay and the frequency in question...
TalismanRich
Well-known member
That is exactly what the OP was experiencing initially. His two side channels were inverted polarity but were summing in mono (stereo width of the pan control was 0) hence they were turning into a complete null, leaving only the mid microphone. The error was that the side channels were NOT panning left and right until he adjusted the pan control to give 100% width. At that point it worked.
rob aylestone
Moderator
The best way to think about M/S is to image you have the mid mono mic signal, and it's right in the middle, between your two speakers. The side signal gets blended in and the side faders essentially push it left and right, away from centre. The amount, is the level of the audio in the side channel.
Steenamaroo
...
This ^^
Apparently it's not important, since no one does it, not even big name equipment manufacturers or software developers.
Labelling the 180 degree polarity flip button "phase" is about as correct as naming the volume slider "pan" and just saying, 'egh, it's fine.
Everyone knows what we mean. You're a pedant.'
It sounds like you understand it just fine.
From your descriptions I'd guess there isn't much of a stereo image to capture, so you're having to increase the sides way too far to hear any appreciable effect.
...or you simply don't like the sound of M/S recording.
Doing a solo voice test in a well treated room, for example, isn't likely to have much interesting information in the sides.
Doing the same in a very live room would be a very different story, or with three singers side by side.
I don't usually bother with the "er, well, achsully it's polarity, not phase" but it is important in this context because one of the main benefits of M/S is that it's impossible to have phase issues.
Phase issues are the result of sound arriving at different microphones at different times and are with reference to a particular frequency.
With two mics unequal distances from the source you can move one of them back and forth slightly, altering the frequency at which they're out of phase.
With polarity (180 degree, anyway), it's either in or it's out.
The reason you hear anything at all with your wide panned opposite polarity tracks from your fig-8 mic is just because of the differences introduced from your speakers, their positioning, the room, the asymmetry and uneven density of your head (not an insult
), etc.As you move around in your room, the degree of perceived cancellation will change, so that's the phenomenon you're hearing,
but I'd argue that the sides shouldn't be high enough for that to be an obvious or distracting thing,
which is what makes me think your test recording didn't have a lot of left/right information to hear.
That image ^ is an example of waves out of phase - Not opposite polarity.
Notice they start at different times. That's because they arrived at the microphone (in our world) at different times. I.E. One mic was farther away than the other.
There's theoretical perfect cancellation there except for right at the start and right at the end, when one track is flying solo.
Consistent cancellation in this example only works because it's an unchanging frequency.
Once you do it with real audio you'll have cancellation at that ^ frequency but the majority of the audio will be fine. (look up comb filtering?)
Sorry folks, pre-coffee rant over!
TalismanRich
Well-known member
... but a very good rant. We often use terms interchangeably when they aren't. We might know and understand what's going on and what we really mean but when trying to explain things to someone else, using the wrong term can cause confusion.This ^
Sorry folks, pre-coffee rant over!
DrewPeterson7
Sage of the Order
Quoting this and not the rest on phase, since that I'm comfortable with - it's the same basic appeal of something like two matched mics in an X-Y position, where with their capsules essentially touching there's no pahse cancellation going on, yet you're still able to pan them hard L and R and create some stereo space.
But, reading this, it sounds like this is just not an approach well suited to tracking a single instrument, so trying it on a single acoustic guitar was probably never going to yield any sort of stereo spread unless the figure-8 signal was lout enough to be audibly "hollow."
Considering I mostly work solo recording instrumental rock music, this is probably an approach I won't ever really have cause to use... but it's at least good to know I was executing it correctly, just on an inappropriate source!
bouldersoundguy
Well-known member
An additional source of phase shift is the innate phase response of devices in the signal path, including mics. That is generally a case of group delay, where different parts of the audio spectrum are delayed different amounts. Most common types of eq do this.
Steenamaroo
...
Yeah, that all sounds pretty reasonable!
It's not something I ever really use, for that reason - mostly I'm working with single instruments or sources in close proximity.
I'd definitely consider it for an ensemble or a performance in a very nice live environment, like a church or concert hall or something.
Steenamaroo
...
Let's not...
Edit : They should have called them 'linear polarity' eq just to be dicks.