some kinda phasing prob with M+S mic’ing
So I was doing some tracking the other night.
I was tracking piano parts for a new piece and to get a natural stereo field and natural sound I chose to use the mid-side technique using a pair of Studio Projects B3 mics. There were several different segments recorded, but note that all segments were recorded in the same sitting, no changes to ANYTHING were made as far as the source or input pathway or settings within N. Everything came out really well with the exception of the first segment which is messed up.
For those of you not familiar with M+S mic’ing, you basically have (typcally) a cardiod mic pointed more or less directly at your source, and a figure-eight mic 90 degrees off-axis to the cardiod mic (so that the top and bottom of the ‘8’ are looking right and left of your source). Then, for those of us that don’t have a schmancy piece of gear with an M+S encoder, you put the cardiod mic to one channel, and you split the figure-eight to two channels, pan them hard right and hard left and reverse the phase on one of those channels. You generally adjust the cardiod channel fader to get the right level blend, and then as you bring up the fadders for the figure-eight mic it brings in your stereo field.
So I’m using the track mixer in N to handle my M+S processing. I cloned the mono track recording of the figure-eight mic, but reversed the phase on the cloned track (using the “phase shift” buttons on that track’s eq pane is that where you do phase reversal in n-Track?)
Out of the 5 segments recorded, the last 4 sound great. They are behaving as they should. The first recorded segment, however, sounds panned hard-left when monitoring off the stereo master, and in fact the VU meter on the master channel confirms this. BTW, the recording was done using v4.5, mastering is being done on a separate system running v5.06. If I solo the cardioid track, it is on center where it should be. If I solo the figure eight tracks the one panned left is on the left, the one on the right is on the right. Put them all together and boom there goes the right channel. If I put the phase-shift to normal on the right figure-eight track the master channel shows that all is well (audio is coming out of both channels, and the sound is appropriate as it should sound like it is all on-center and it does.
I’m stumped. Again, segments 2, 3, 4, and 5 sound great, and all 5 segements were recorded in one session, no changes made that would affect this. I also checked to make sure the offset is the same for all three tracks and it is. I have even tried opening up the two .wav files in a new song, cloning the figure-eight to two tracks etc., etc. Same problem. So it seems the problem resides with the .wav files, and that something happened during tracking. Dunno, just wondering if others have experienced anything like this. I tried setting the all the channels to stereo and experimenting with phase-shifting only one side of the right OR left figure-eight channel trying to isolate the issue. No luck…
There’s an easier way to mix MS, just use the free “Image” vst from http://mda-vst.com, and set it in “MS->LR” mode. Very cool and simple plugin. Then it only uses one track.
Also, in n-Track, you don’t have to clone a track to invert one side. Just check the “convert mono track to stereo” box in Properties, and then invert one side (in the EQ – yes, that’s where you do it).
Don’t know what your problem is, definitely sounds like something must have changed after the 1st segment. Post a 20 second clip stereo wave file with M on left and S on right and maybe one of us can figure out what’s up.
“If I put the phase-shift to normal on the right figure-eight track the master channel shows that all is well (audio is coming out of both channels” - this our problem,
if you are not sure how it happens ‘GOOGLE’ for ‘phase cancellation’ -and or- ‘remove lyrics (karaoke)’ -
then download the free plugin linked below as it may help you without re-recording
Out of my element here - because, frankly, I can’t imagine reversing the phase on a stereo track for any purpose other than to remove most of the “center” channel from a stereo field (as in Karaoke) - but I think the point is that segment 1 behaves differently than the subsequent 4 segments and that IS weird.
sweet_beats - the only thing that occurs to me is scientific method. Do it again and see if you can replicate the result. I suspect that you will find that even though you recall doing the exact same thing with all five segments you did, in fact, do something different with segment 1. I think the fact that it is ‘segment 1’ supports this hypothesis. That after dealing with the first segment you thought of some sort of short-cut or change to the process that, you thought at the time, wouldn’t change the result but it did, in fact, change it.
Record five segments by the exact same process and compare - I’d be very surprised if one of them turned out differently than the others.
learjeff…if I had a nickel for every useful tip you’ve offered…
Thanks for the link to the freeware VST pack…looks really neat. Downloading now…
Also, THANK YOU for the pointer for using a single stereo track for the ‘S’ mic…that does indeed simplify things.
I will try to get a snippet posted. Wish I could upload to the forum but…I’m having some domain issues with my website and so it may be a bit before I can upload there and post a link.
DR Jackrabbit, I’m confused. You say “this is our problem” but I don’t understand what the problem is with my statement you quoted. I’m saying that if the three mono tracks (the cardioid mid-channel, and the two mono tracks for the side mic) are all set phase-normal, then the playback sounds like everything is dead-center in the stereo field, which is confirmed by looking at the master meter, and I’m saying that is the way it should be if all three channels are phase-normal. We are really talking about two mono tracks, and no matter what way you slice it, if you split a mono track to stereo and pan the two identical channels to hard right and hard left it is still going to sound as though it is coming from center unless you start messing with the phase or the timeline relationship between the two (i.e. minute delay). So the split mono side-mic tracks are still going to sound mono, and you mix that with the mid-mic mono channel panned center and you get…mono! So what I’m saying is that it behaves properly when all tracks are phase-normal, and I’m also saying that when each track is soloed individually they behave properly as well regardless of the phase settings. It is when I phase-invert one of the cloned side-mic tracks (in this case the one panned hard-right) that the problem presents itself in some sort of wierd right-channel pase-cancellation. I really want to understand what you are saying…please help me out.
BillClarke, When you duplicate a mono track and pan the two hard-left and hard-right, the phase inversion on one of those tracks doesn’t cancel the middle of the stereo field as the two tracks are not coexisting in the same realm of the stereo field (they are polarized). The result is a perceived widening of the stereo image with reduced definition of the placement of individual sounds in the stereo field. The cardioid mic captures your source directly, while the figure-eight captures the room since its axis is oriented perpendicular to the cardioid mic. When you duplicate the figure-eight track keep in mind you are duplicating a mono track of made up of stereo information. does that make sense? The figure-eight mic is capturing sonic information from the right and from the left of your source using its two diaphragms facing back-to-back from each other, so duplicating the track, panning the two opposite each other and phase-inverting one track is not same with a figure-eight mic as it is with a uni-directional mic. Yes you are phase-inverting two identical mono tracks, but those tracks contain stereo information and by doubling the track, placing those tracks at opposite ends of the stereo field, inverting one side and then blending with the center-image cardioid, you very acurately recreate the natural stereo field of the room. It is based on the fact that a figure-eight mic is generating a ‘+’ signal on one side and a coincident ‘-’ on the other. By inverting the phase on the ‘-’ side you are able to sum the ‘-’ information with the mid-mic, as well as summing the ‘+’ information from the left-channel with the mid-mic. You are extracting the stereo information from the mono track when blending with the center channel. I’m not explaining it very well…try going here for more info if needed.
Scientific method aside, notwithstanding your excellent logic, I setup the physical and virtual connections from my mics through to the record tracks in n-Track, and made no other changes each each of the many times I pressed ‘record’ for the whole night. I did many, many takes, keeping the 5 mentioned above. I don’t believe any settings were changed. Not saying it isn’t possible, just saying that is why I am so baffled. I’ll be happy to admit my failing memory when it is exposed.
Thank you all for your thought provoking ideas and suggestions. I really appreciate it. I’m going to mess around with it some more, and I’ll tray to get some audio data posted.
i have just simulated your problem -
lets look at it like this -
track one is left panned figure of eight mic -
track two is centre cardiod mic -
track three is right panned figure of eight mic - phase reversed -
when you playback to a 2 track stereo output this is what happens -
track one becomes the left track of the stereo pair -
track three becomes the right track of the stereo pair -
track two is added to the stereo pair, being a mono signal equal ammounts of track two is shared between the left and right stereo pair - but heres the tricky bit - the mono signal is IN PHASE with the left track and OUT OF PHASE with the right track - two signals 180 degrees out of phase cancel each other out - hense only left track has audio -
what they dont tell you about M/S recording is that the figure of eight mic is in fact a stereo mic not a mono mic and as it has two independant sides to it (and two output leads), it picks up two differing signals which are not so prone to cancellation by the centre microphone -
a dual capsule mic is still mono it just sums the input from both capsules together -
i cant see in your original post what mics you used ? -
You basically reiterated what I said in my last post…
I concur. BUT, a mono signal split and panned hard-right and hard-left with one channel phase inverted does NOT cancel. As you bring the two channels progressively to a center pan, then yes they diminish until they completely cancel each other out when both panned center, but when placed opposite each other on the stereo spectrum and phase-inverted both channels are quite audible.
BillClarke et al, okay…this is really embarrasing though…I have to admit that I figured out the problem…not sure how it occured BUT nevertheless my failing memory IS EXPOSED!!
I discovered that for some reason the cardioid mic didn’t get tracked on segment 1…what I recorded was two tracks of the figure-eight mic so naturally the right channel would cancel out if I have three channels of the same file L-C-R and I phase invert the right channel. Discovered late last night…my confession is complete. Still not sure how I did it though…no plugin will fix THAT problem. Two separate .wav files (named differently by ‘N’ and everything) were created of the same material on two separate tracks. Haven’t taken the time to tease out the logic on that. I’ll try to recreate the error so I can avoid doing it again!
Jackrabbit I’m using a pair of Studio Projects B3 mics.
Dr. JR, you’re assuming that the two mikes are pointed in roughly the same direction (or that they’re omni).
With MS technique, this is NOT the case. The S mike points to the side, and gets the signals whose wavefronts are at a 90 degree angle to the mid channel (M) mike. OK, since it’s not a “laser beam” mike, the signal is max at 90 degrees (and minus-max at 180 degrees, since it’s fig-8).
The bottom line is that sweet-beats describes the precise mixing matrix required for MS technique. You should try it yourself and learn what it’s good for – it’s an excellent technique for a number of reasons. (Reasons I’ve posted here numerous times, and which I also believe I’ve posted on the n-Track wiki.) Like any techique, it has its advantages and drawbacks, so it’s not a cure-all one-size-fits-all panacea to stereo miking. However, it’s by far my favorite for miking an acoustic guitar. The karaoke thing (center channel cancellation) does NOT happen if you use the right kind of mikes and point them 45 to 90 degrees apart (with M mike pointing dead center).
Sweetbeats, when recording MS, you only need to record two channels. You can toss the 3rd track in all cases, since it’s simply the 180 degree inversion of the L. If you do this, and for all segments, invert the right side on the track facing away from the S mike, you should be golden.
You must have changed something between the first segment and the others. Hey, been there, done that, and forgotten about it too!
And in the future, just record one stereo track for the two mikes. Put the M mike on the left and S on the right. Then use MDA Image in “MS -> LR” mode.
BTW, folks, you do NOT need a fig-8 mike to use Mid Side. You do to capture a true soundstage, but not to simply add stereo ambience when miking a single instrument. You can also get away with using cardioids when miking a duo where one guy is the lead and the other is mostly backup, point the M mike at the main man, and whichever side the other guy is on, point the S mike that way. If it were a trio and you did this, the guy on the other side would be under-represented. In some cases this can be a good thing.
|what they dont tell you about M/S recording is that the figure of eight mic is in fact a stereo mic not a mono mic and as it has two independant sides to it (and two output leads), it picks up two differing signals which are not so prone to cancellation by the centre microphone -|
False. A figure-8 mike is a mono mike with a pickup pattern in a figure-8 pattern, and where a signal arriving one loop is negated before being added to the signal from the other side. Many fig-8 mikes have two capsules – but most of these are switchable as fig-8, cardioid, and omni. However, for any capsule that can detect sound from either side with the proper phase relationship (e.g., a PZM mike mounted on a panel floating in the air – don’t do this at home kids), only one capsule is needed.
As always, thank you for you concise and informative input. I’m always relieved when I post a topic and see your replies.
I downloaded the MDA VST plugin pack and quickly checked them out. VERY cool, especially for FREE. Astounding really…the vintage echo actually does a nice job of emulating vintage echo degradation and overdrive. I am an avid fan of vintage dub…I own a Roland DC-30 that doesn’t get much use, but with the MDA pack it might be getting less!
M+S mic’ing IS really misunderstood, but pretty easy to try it out and see, as with many things. I LOVE your mention of the PZM mic in the ‘S’ position. I own a couple Radio Shack PZM mics…might have to try that too!
I’ve found M+S mic’ing to bring an intimate and natural ambiance to a recording that you CANNOT achieve with X-Y or A-B techniques. The recording that started this thread was of me playing our upright “grand” piano…it is in our living room. We recently acquired it, and it is old and neglected, but suprisingly well in tune (if you consider being about 90 cents flat accross the keyboard in tune ). The hammer linkages rattle and click, and I recorded it using the M+S method and on playback it sounds like you are in the room with this old piano. It is really intimate. Drums too are something I’ve used it for. If you are looking for that “rip-your-face-off” tom fill that careens from one side of the sound-stage to other, then M+S mic’ing might not be for you. But if you are looking for a sound of the kit as a unified instrument and you want it presented as though you were sitting in the room as the kit was recorded, NOW you are talking M+S mic’ing. Its a great tool to have in the bag, and with plugins like MDA Image it is even easier to do.
The point about the PZM is I don’t know whether it would work. PZM’s are also known as “boundary” mikes; they work on the principal that they’re attached to a wall at the edge of a soundspace. Floating one in the air? I don’t know how well it would work, or what size to use, etc., etc. But it might be fun to try and who gives a hoot for the theory if it sounds good?
The main thing to know about M-S is it involves a different phase problem check. When done the normal way with mike capsules close together, this is unlikely. However, MS is also fun to use with FX, and the following check is especially important in that case.
We all know that when we’re nearly done mixing, it’s important to check our mixes in mono. [Frankly, I find that mixes work best if we frequently adjust levels in mono or at least cross-check them, and a mix that sounds great in mono will sound fantastic in stereo (assuming reasonable stereo technique). But the reverse is not at all true; stereo allows us to use image for separation so often instruments will blur together in mono that are nicely defined in stereo. And of course, any stereo mix we make will often be heard in mono, either through a door, or in a low-fi MP3, or when too far from speakers, etc. Sorry, back to the point …]
One of the main reasons we double check in mono is to discover if we’ve made a mistake with phase-related techniques, including poor mike spacing or overuse of ‘clone and drag’ effects. If we have, the result sounds thin in mono.
Well, when using MS techniques, we need to add another check to our list: monitoring left and right side alone. This is because, if we’ve goofed with our use of MS, or overused it on an instrument, it will NOT cause a problem when mixed to mono because S always cancels out in mono. However, if there ARE phasing problems in the S channel, it will cause the classic phasing problems for anyone too close to the right or left speaker!
If there were phasing problems in the M channel, you’d hear them while mixing, so it’s really only the S channel, or the S and M channels together, that we need to worry about here.
As an example, I made a stereo soundfont from my Rhodes piano. I added an effect using the pitch shift from CoolEdit96. For some unknown reason, the result sounds more like subtle stereo chorus than merely pitch shifting should. In stereo, it sounds lovely, exactly the effect I wanted in the first place. In mono, the effect is gone completely – the reason for using MS. However, when listening to either channel alone (or when significantly closer to one side), you can hear the phasing effect of a subtle mono chorus effect, a little like a phaser sounds.
Fortunately, with the sound of a Rhodes I don’t mind this a bit. I’d like it fine on an electric guitar too. But it would sound terrible (IMHO) on an acoustic piano or guitar, where I might be using pitch shifting to generate a stereo image from a mono signal but preserving the natural sound (what pitch shifting is best for, but alas the software methods I’ve found just don’t quite do).
I was a bit surprised the first time I noticed the phasey effect, but after a moment’s reflection the reason was clear. So I had to amend my belief that MS avoids phase problems, to that it shifts it from phase problems on sum-to-mono, to phase problems in either side. Of course, with good miking technique, you won’t have these anyway; it’s generally only when you try to color outside the lines or use it to create a stereo effect from a mono signal that you need to worry.
About the MDA effects: I find they’re great when they’re relatively simple mathematically. For more sophisticated effects involving complex DSP, I suspect they simply coded the standard algorithms and didn’t fuss about avoiding infinitesimals and other practical issues (stuff I can name but don’t really understand) and as a result the plugin works but doesn’t really sound great. For example, I don’t care for their pitch shifting and chorus effects.
However, I do like their “combo” speaker sim. It may not be a fantastic speaker sim, but for free it’s fabulous, and along with a few other free tube sims can add life back to a direct recording of electric guitar. Sure, it’s as good as a POD or Johnson J-Station, but heck, it’s FREE. There are a number of other nifty little utilities I forget about until I need it.
And back to MS miking: it’s now my favorite way to record myself on acoustic guitar. With other stereo techniques the mikes were either too close to the guitar (causing problems if I moved a muscle) or too far away to get a good image. And balancing the two was always a problem. I now use an SM57 for the Mid mike, and an SP B3 in either fig-8 or cardioid for the Side mike. I’d like to get a C1 and try that for the Mid mike; no doubt I’d get a crisper tone that way. But I like the warmth of the SM57. I use a pair of SP VTB-1 mike preamps, great addition to any home studio and very affordable!
I’ll read your fabulous pontification on M+S mic’ing a couple more times through to glean more, but I think I avoid phasing problems pretty well in that I’m using two of the same model of mic (SP B3’s, albeit not “matched”), and I am placing the S mic in figure-8 upside-down directly above the M mic so that they make one long cylinder, and placing them as close together as possible. I’d like to experiment with other combinations though. The PZM’s, I was picturing trying them placed back-to-back with a 2’ x 2’ (or so) baffle sandwiched between them…57 as a M mic…no scoffing here. I have three of them and I’ve found them to be very versatile in settings when others would raise eyebrows. Had a great drumset recording with two 57’s in an X-Y pattern over the kit and a D112 out front of the kick in a church sanctuary. The toms came out huge (no suprise there) but with the room sound the overall capture was very balanced and smooth without the harshness that I’ve struggled with when using condenser mics in conjunction with a kick mic to capture a kit.
I have recorded acoustic guitar with a C1. I like the B-series much better, but I was also not using M+S techniques at that time. I do like my C1 though, and it may have been a poor combo of the particular guitar and the mic.
One thing I believe more and more; that there really are very few hard and fast rules unless you believe you can only do what others say, or what has been done before you. We were having a problem with snare dynamics in our church. Fairly large sanctuary. The drums are enclosed in a plexi-glass “booth” with a sorber panel roof. Low-end set of mics on the kit. Allen & Heath GL3300 console and a Presonus ACP-88 among others at FOH. I’d love to get the snare and kick on their own compressor insert but most of the ACP-88 channels are needed elswhere so we are having to stick with using one channel on the ACP-88 as a soft-limiter insert on the subgroup with the drums. Because of challenges with the room dynamics we have struggled with the balance of getting the snare to present in the mix without being uncomfortable, especially when shifting from harder back-beat material to brushes and such, so I went with my gut one night and put that cheap mic under the snare and it totally solved the problem. We were laughing it sounded so good in the room and totally ate it when PFL’ed in the cans. You have to just try things.