Mid-side Issues in the digital world

[jgimbel]

I've only tried MS recording a couple times, without much success, because I think I'm doing something wrong. I've got a couple plugins that do MS encoding/decoding, but after much trying and reading, can't figure out if which of the three tracks (mid, side, and side flipped) the plugin is supposed to go on - one? All three? Just the sides? Some plugins have a button you press for mid or for side, some have one button for MS, one has a L phase /R phase/MS/ thing that I can't figure out exactly what each one is intended for, and most things either end up not changing the sound or cancelling the track out even when solod. I'm also using Cubase LE4, which doesn't just have a button to flip/unflip phase. It does have Audio > Process > Flip phase, so I believe copying the side track, doing that to it, then panning them is correct, but since I haven't really done MS before I'm not sure if I'm getting what I'm supposed to. I feel like I'm doing it right but something is just sounding weird about it.

Anyway, forget all that. I'm wondering if this kind of messy method of MS recording would work. I've got two VTB-1s, and a dual API 312 clone, that all have phase-reverse switches in them. Would it be ridiculous to use something to split the mic signal before hitting two separate channels of the preamps, and then flipping the phase on one? One thought I had was that since the phase would be flipped before being recorded, there might be ever-so-slight variations between the two signals, so it might not work correctly. But I have seen non-digital MS decoder boxes, so I'm assuming it should work. The only issue with this is the phantom power thing, because since the signal is split I'm thinking it wouldn't be to great to have the phantom power coming from one preamp up one leg. I do have a standalone phantom power box. But mic > phantom power > splitter (I have a direct box that has both 1/4" and XLR outputs that can be used together) > two preamps, one phase flipped > interface - is that just completely ridiculous?

I'm absolutely positive I'm overthinking this, but when I do it digital there's just something that's not right about it. School me!

[vikingrecording]

you're overthinking it. Now, someone may come and correct my info, but this has worked for me:
record you mid and your side like you have been.
Duplicate your side and flip the phase
group the sides together and adjust them via the faders relative to your mid mic.
adjust to taste, enjoy serious stereo!

And leave the encoding/decoding plugs out of it.

Hope this helps!

[the finger genius]

I noticed you're concerned about phantom power, what are you using as your side mic?

[cgarges]

you're overthinking it. Now, someone may come and correct my info, but this has worked for me:
record you mid and your side like you have been.
Duplicate your side and flip the phase
group the sides together and adjust them via the faders relative to your mid mic.
adjust to taste, enjoy serious stereo!

Yeah, if you do this, you don't have to worry about applying any kind of decoding. You adjust the width by keeping the two "side" faders at equal level and messing with the level of the mid fader, relative to the sides. Using a decoder scheme just allows you to take the mid and side channels (the two mics) and decode them to a pair of left and right channels. If you have a decoding scheme (via plug-ins or a box or whatever), you shouldn't need to worry about that flipped version fo the side mic. The decoder takes care of that for you.

Chris Garges
Charlotte, NC

[jgimbel]

Alright, that works! Yeah I've kind of seen that plugins will turn it into a L/R channels, but I guess I'm not completely sure what the use of this would be, unless you just planned on having the balance lean more to one side or the other.

I noticed you're concerned about phantom power, what are you using as your side mic?

Well I've only tried it once or twice, but when I did it I think I was using the MXL4000 I had at the time (I'd assume, as it was the only bi-directional mic I had). I just got an AT4050 so I'd like to give that a try as a side mic, along with maybe an AT3035 for the mid for something similar sounding, or another mic that's very different sounding, like maybe the Oktava MK319. I tend to love dark sounds and hate brighter ones, but when they're put together it can be so awesome. I remember noticing with the MXL4000 that despite being bi-directional it was noticeably asymmetrical. I haven't done anything with this pattern with the 4050 yet so I'm not sure if it has the same issue, or if this will even be anything of importance with the midside recording.

Anyway, thanks for the help!

[vikingrecording]

I love a couple 414s for my mid-side setup, also try your mid in omni! FUN!

[ashcat_lt]

There are three steps to most of these ITB M/S processes.

1) Encode. This will take a regular stereo track and encode it - using that phase flipping math problem we've all seen - so that (say) the Mid on the left and the Side on the right. If you're recording an actual M/S mic pair, you've already done this part.

2) Process. Run this "dual" track through whatever effects or processes you want to apply. Depending on what you're trying to do, and the capabilities of your DAW and effects, it might require some creative bussing to split the stereo track and apply different processing to each.

3) Decode. This does the whole math trick backwards, returning the M/S dual track back to a regular stereo track. If you bussed them out to seperate channel strips for processing, you'll usually need to bus them back to either side of a stereo channel to accomplish this step.


That said, some plugins (more and more these days) will do all three steps "behind the scenes" in one plugin.

[jgimbel]

Ah okay, so I knew I was having some issue other than just not knowing what to do. Last night I set up an M/S pair - AT3035 as mid, 4050 as side, in figure 8. Recorded each to its own mono track. Duplicated the side track, and went to Audio > Process > Reverse phase. It gave me a weird error message saying I had to restart the program, but it continued to work fine. When I solo'd the side track with the phase reversed one both panned center, no sound - canceling out. Solo one at a time and there's sound, so everything seems good. I pan the side track left, phase reversed side panned right - pretty strange but wide sound, but as far as I know that's what it's supposed to sound like. Then I un-muted the mid track. It did have a wider sound than mono, but it was HEAVILY leaning to the right side. Mute the side again, balanced. Confusing. I double checked everything to make sure I had definitely duplicated the side mic and not the mid mic, and everything checked out. Very confused. I thought maybe since I got an error from reversing the phase, it did something wonky like reversing the phase of just the left or right side or something. I restarted the program and tried everything over, no luck. I've got a couple plugins that can reverse phase, so I started over. Deleted the reversed-phase track.

Duplicated the side track again, put a plugin called "stereo tools" on it. It's got a million little buttons in a tiny interface, but for each the left and right channel it's got a "+" sign, click it and it turns to a "-", so I assumed that's phase reversing. So I click around a bit with the other track solod, and indeed, if I put them both to "-" everything cancels out, so the phase has been reversed on this track, yes? I pan, and same issue. HEAVILY leaning toward the side the reversed phase track was panned to (this time to the right).

So okay, maybe that plugin just has a little too much going on and I don't use it enough to get it right, so I go to a simpler plugin that is literally just a phase reverse. Turn it on, phase is reversed, off, phase isn't. Tried it, same issue.

So then I thought hm, maybe it's something to do with the effect being pre or post panning, so when I pan it to the side it's only effecting one channel. Damned if I know, I thought it could be a possibility. So I took the phase-reversed side track panned center, and exported the track. Then I imported it back in, set it up panned opposite the normal side mid - SAME ISSUE.

What am I doing wrong! I figure since it's the same thing happening each of these ways, they are all definitely doing the same thing, not anything I'm doing wrong with the plugins, or an error happening when the program got an error from reversing the phase that way. It must just be something I'm doing wrong, yes? What is it! I feel so stupid!

[the finger genius]

It's really hard to tell without being involved from step one, but here are a couple guesses.

1. Have you done any test to make sure that both sides of your figure 8 mic are voiced the same way? (Same approximate levels, frequency response, etc....)


2. Can you describe your mic position? The mid mic should be pointed at the source with the side mic pointed at a 90 degree angle from the source. The two mic capsules should be as close together as possible.

[jgimbel]

I tried my best to describe it from step one. Yes, my mic position is a typical mid-side setup, mid pointed at the source, side mic at a 90 degree angle, capsule as close as possible. I promise you the mic position is not the issue, it's exactly what I've seen and read a million times.

When I say it's HEAVILY leaning toward one side, I'm not talking about something as subtle as a slight difference in frequency response. My previous multi-pattern mic, the MXL 4000, was darker on the back than on the front. When I got the 4050 I tested it on a few sources, and moved the mic around, because I was curious to see if it'd have the same difference in freq response. They're amazingly similar, as little of a difference as I could ask for. Also cool was how similar the response is between the different patterns, very, very close.

Regardless, this is not a subtle thing like that, though I could see that being a big deal in an extreme case where the sides were REALLY different. But nope, not here. I don't know if I should post the sounds up here or something..the mid mic, the side mic, then the mid with side, duplicated/phase reversed side, panned. There aren't any issues with the mics though, everything is working fine going in. I know any time people ask for advice folks often suggest it's something much simpler than what they're suggesting, but I'm not a novice with mic position, and something like a difference in frequency response wouldn't slip past me. I may not be a pro, but I'm not exactly an amateur!

[the finger genius]

Not trying to insult your intelligence, it just sounds like one of the side channels (the unflipped i guess?) is canceling out with the mid channel. Although then, I guess you would notice that when just playing both mono tracks without any processing. Hell, I don't know. If you want to post the two tracks I'm happy to test in my DAW with my usual M/S process.

[Recycled_Brains]

Is it possible that the M mic's XLR cable, or maybe the input/output of the pre that it's plugged into, is in oposite polarity of the side mic's equivalent?

That's all I got.

[millzners]

I'd just like to point out you're using the exact same mics and setup that I often use. I absolutely love the sound and the flexibility this arrangement gives me even on something simple like a guitar cab.

I record the AT3035 as mid, the AT4050 in figure 8 for side. Then I do the usual: split the AT4050 by sending it to two different tracks, invert the phase of one, pan all the way left and then right, then send those two into a single track. Maybe you're doubling the original side channel with two tracks b/c of how you have it split out? That happened to me once, gave a similar effect of one side being twice as loud as it should...

I absolutely love this setup... I don't know if this is helpful, but I got excited someone else is doing the same thing. This weekend I may try an SM57 for a 2nd mid on a guitar cab, just for a 3rd mic to throw in the mix...

[dave watkins]

yeah mid side is pretty much the bee's knee's i use it all the time, but what is also good about tracking in mid side (and not using an encoder/decoder) is that sometimes you'll get to the mix process, and think that the mid side tracks have tooo much stereo spread and kind of over take the mix. you have the option to lessen that effect by turning down, or turning off the side mic all together if necessary, and sometimes the crazy off axis signal from the figure 8 by itself can be exactly what you need. oh and i definitely second the mentality that you can use any polar pattern for the mid mic, i guess cardioid is what's considered proper, but the only thing that really matters is in the end is how it sounds so... do what you like.

hope you get it worked out

[jgimbel]

millzners, that's so awesome to hear you're using the same mics, that makes me really excited and anxious to try to get it to work!

dave, yeah it's such an interesting thing to be able to have the certain kind of control over the stereo field. Just changing the panning of a stereo pair doesn't seem to have that same effect as actually widening things by adding more of the mics that were picking up those actual sides. I feel like I'm going to LOVE this technique if I can get it to work.

I'm going to record a sample today and I'll upload the mid mic, side mic, and then mid with side, duplicated/phase reversed side, panned. I'm interested to find out if this issue will happen for other people using my same tracks, or if it's something going on within my program.