On the thread about stereo widening syn mentioned voxengo plugs, so I checked them out (hadn’t looked there in a while) and noticed they have a free m/s encoder/decoder. I have a ribbon mic for figure 8 - so that plus one of the cheap MXL condensors made for my sort of cheapo mid-side pair, and with the decoder, well, it works!
Tried it on drum set, totally cool stereo, fully mono compatible, very realistic image of the set. Gonna be my new cool thing for the next few weeks.
Anyone used this technique? What should I know about it?
|Quote (TomS @ Feb. 26 2005,17:24)|
|Anyone used this technique? What should I know about it?|
Well, you don't even need the decoder. Just clone the figure 8 track and flip the polarity of one of them and you are set. Other than that, um yeah, M/S works, it is mono compatible and sounds really good.
Huh - shows what I know - never knew I could do that. Thanks Bubba!
OK, so what would be a good mic for the M, Bubba? The S function is being performed by my RCA 77.
Well, usually you want mics that are very similar to one another so that your M and S have the same tonality. Usually folks will use two mics with variable patterns, so two 414s for instance, one set to cardiod, the other to figure 8. For a poor man’s example, maybe a Studio Projects B3 for the S and a B1 for the M. At the end of the day, so long as it sounds good, who cares, but generally you don’t want the tonality different between the mics.
Yeah, I wondered about that; the mics I used last night sound very different. Well, I can’t afford another ribbon, guess I’ll just have to try what I have.
Thanks again Bubba!
Can’t this be done with three mics as well? One as center and two facing away from each other 180 degress as a sort of poor man’s figure 8? (Doesn’t one need to be phase flipped? I don’t remember how they should phased, or anything about spacing.)
|Quote (phoo @ Feb. 27 2005,12:39)|
|Can't this be done with three mics as well? One as center and two facing away from each other 180 degress as a sort of poor man's figure 8? (Doesn't one need to be phase flipped? I don't remember how they should phased, or anything about spacing.)|
That's exactly right - you can make a usable figure-8 mic by using 2 identical cardioid mics. If they were dynamic mics you could even just use a Y-cord (2 female XLR to 1 male XLR), with one side out of phase with the other. For condenser mics, you'd need to use 2 console inputs with the phase flipped on one.
The mic diaphragms should be right beside each other (so that they are as close as possible) and facing in opposite directions.
As one of the uninformed -
What is M/S and what does it accomoplish?
I’m always looking for a "better way"
“Mid-side” - it’s a coincident technique that gives results a lot like an X/Y arrangement, but you use a figure 8 mic (for the side,set up so that the null faces the sound source) and another mic, either omni, cardiod, or whatever, for the mid (pointed right at the sound source) and through an “MS matrix” you can get the equivalent of an X/Y arrangment. By varying the proportions of M and S mics you can actually change the effective angle between the equivalent X/Y arrangment.
The matrix is a “sum-diffrerence” matrix, and “decodes” the two signals to make a nice, mono compatible stereo signal. Much better than time based stereo like two spaced mics, IMHO. And more flexible than an X/Y arrangement, since you can effectively alter the equivalent angle just by moving a fader or whatever.
There is a bit more on M/S here:
And there is a paper by Wes dooley adn R. Streicher called MS Stereo: A powerful Technique for Working in Stereo" from 1982, posted on Dooley’s website, I think. That paper has a whole host of digrams describing the equivalent XY angles for various proportions of M and S, has a bunch of the math, and talks about the various ways of constructing a matrix. It was in the Journal of the Audio engineering society vol 30, pp 707-717. FYI.
While Archimedes is correct that you could use 3 mikes, you won’t get the ideal results that way.
One of the big advantages to MS method is it’s a solution to the old problem of phase cancellation artifacts. Unless two mikes for A/B stereo are in precisely the same spot, there will be frequencies where the two mikes will pick up the same signal at different phase points. That is, there will be a delay as the wavefront hits one mike and then the other. For large frequencies, the phase shift caused by this is negligible. But for frequencies whose wavelength is double the distance between the two diaphrams or less, you’ll get phase problems (cancellation or emphasis) whenever the two sides sum to mono, causing unnatural coloration. In other words, to avoid this problem up to 14kHz, keep the diaphrams less than 1/2" apart. (I do this by putting one capsule directly over the other – little sound will be coming from directly above or below. But with big LDC’s this is cumbersome.)
Using mid-side, you don’t have this problem (you trade it for a different one, which I’ll mention below – there is no free lunch!) Because the “side” mike is added to left and subtracted from right, when the two channels sum to mono it cancels out completely. For this reason, you want the better of the two mikes facing center. Also for this reason, while it’s best to have two similar mikes, it’s not nearly as critical as it is with LR stereo miking – where you clearly hear any tonal difference between the two mikes as an unnatural image shift or worse, unbalanced EQ in the two sides.
What’s the tradeoff? Well, when you’re much closer to one speaker than the other, you’ll get the same kind of problem (coloration due to phase cancellation for wavelengths less than double the distance between diaphrams). So, even with MS, you want to get those capsules as close together as possible.
Using 3 mikes, I think you’re even more likely to get phase cancellation artifacts, and you’re not likely to get the ideal sum-to-mono behavior as a result. Also, any difference in coloration or response between the two side mikes will be as noticeable as imperfectly matched mikes for LR. In other words, you’re likely to get all the disadvantages of both methods, and the full advantages of neither. But that’s just my prediction – I haven’t tried it. Practically, it might work out to be a compromise of sorts.
MDA also has a free mid-side plugin called “Image”. It’s a very nice one, in that you can also use it to widen any stereo track that’s already stereo. (It won’t have any effect on a mono track or stereo track with the same signal in both sides.) You can use it in MS->LR or LR->MS or even LR->LR mode. Now, what good is LR->LR mode?
One of the advantages to MS mode I didn’t mention above is that you can control the stereo width by adjusting the gain on the Side channel. With the MDA plug, the input signal gets converted as necessary to MS mode. Then the plugin’s controls apply. Finally, it gets converted as necessary to the output mode. So, regardless of input and output modes, inside the plug it’s always MS mode, and you can control the signal width, from 0% stereo to 200% stereo (where the S channel has twice the gain of the M channel, widening the signal). Just note that if you hyper-widen the signal, it will get a lot quieter when summed to mono.
Finally, I use MS technique for stereo imaging effects like pitch shift doubling and mild chorus. Sometimes we use these FX for their coloration effects. But when I use them more for their imaging effects and don’t want the color, I use MS technique by adding the effect to left side and subtracting it to right – e.g., using the FX on an aux bus and following it with MDA Image in LR->MS mode. This means that for anyone near the center of the listening area, the effect will be purely stereo imaging and not coloration. And if summed to mono, the imaging effect disappears entirely. Of course, folks much closer to one speaker or the other do get the coloration effect, so I only do this when the subject can tolerate it, like Rhodes or electric guitar, but not voice or acoustic guitar.
|Quote (learjeff @ Feb. 28 2005,07:12)|
|What's the tradeoff? Well, when you're much closer to one speaker than the other, you'll get the same kind of problem (coloration due to phase cancellation for wavelengths less than double the distance between diaphrams). So, even with MS, you want to get those capsules as close together as possible.|
Lerjeff, I am not sure that I understand this (and I am sure that my futher comments will amply demonstrate this!). There are timing issues involved in using two mics in a M/S set up, since it is simply not possible to have the two mics coincide perfectly, but these have more to do with vertical than horozontal perceptions, don't they? So the coloration problem you mention would essentially relate to the vertical and not horozontal. That is, unless there is a M/S method that doesn't have the mics placed one above the other. But vertical imaging differences are much more tolerable than horozontal ones, aren't they?
Or have I missed the point completely? How obvious would the coloration you mention be?
First, Tom: vertical imaging isn’t done using differential delays like left/right imaging is done. It’s more through EQ, and much more subtle than left-right imaging techniques using amplitude and differential delays. (Ditto for “behind you” images.) By “differential delay” I just mean a delay in one ear versus the other depending on the angle to the source.
So, forget about pointing mikes up and down or above or below each other to have anything to do with vertical imaging. It doesn’t work that way, simply because we don’t have ears on our chin & forehead!
Assuming that one mike is directly above the other, we don’t get any timing problems from sounds coming from center or either side. We would get problems from sounds coming from very high or low angles, but generally that’s very small compared with what’s coming from the same level as the mike.
But that’s something to keep in mind: vertically centering mikes doesn’t work well in situations where sound can come in at a steep vertical angle. E.g., don’t use mid-side right on stage, with sound coming from floor monitors below, cymbals, or PA mains above, and amps directly in front. At least, don’t do that with LDC mikes where you can’t get the capsules nice and close together.
However, I might be overemphasizing the problems due to coloration. Note that on that web page there are a number of mike setups where the mikes are a distance apart. Personally, I’ve had better success with one-point techniques – but I like a good mono mix and I’m willing to sacrifice some separation for that.
Another factor to keep in mind is how prominent the miked signal will be in the final mix. If it’s ambience or background and won’t be very strong, you can tolerate a lot more potential for phase issues, just because there’ll be enough of the signal coming directly (e.g., off the board) and the coloration from the ambience mike won’t be enough to matter. It’s more important to get the mikes in phase with the direct signals and worry about that timing discrepancy, than the timing problem between the two mikes.
Especially for live recordings, folks are used to a bit of coloration. The venue will certainly add some color just due to the room. I’ve made recordings I liked where there were phase problems all over, due to using lots of mikes, almost haphazardly located. Sure, it might not please a purist, but gee those cymbals sure splashed all over the soundstage!
Um…you mean to say you don’t have ears on your chin and forehead…?
Right. So, where exactly is the coloration one hears when one is listening closer to one speaker rather than another coming from? What is it due to?
Well, not the last time I looked … or was that Picasso in the rest room really a mirror?
That coloration would only happen for mid-side miking if the capsules weren’t directly above and below each other, or if you have signals coming in from too high or low an angle – i.e., it’s not a problem for ideal mid-side miking. However, it is a problem when using mid-side technique with effects like chorus, where you take the variably time-delayed signal and add it to left while subtracting it from the right.
Let’s look at that in more detail. There are two simple ways to use a basically mono chorus effect to produce a nice stereo chorus image. One way is to put the unaltered signal in the left channel and the altered signal in the right channel (or vice versa). This gives a lovely swirling stereo image. Of course, if you’re sitting right by the left speaker, mostly what you hear is the dry signal. If you’re sitting by the right speaker, you’re hearing mostly just the “wet” signal. If the chorus settings are relatively mild, this will sound pretty much indistinguishable from the dry signal. If you’re sitting where you can hear both speakers, you get that lovely swirling image. If the two channels get summed to mono (or if you’re far from both speakers, or go through a door, etc.), you’ll hear the wet and dry summed together, giving a colored sound (and the color warbles as we’re used to with a mono chorus sound).
The other way is the mid-side technique. In this case, we put dry in the middle, add wet to the left, and subtract wet from the right. Folks sitting in the middle will get a lovely swirling image. Not quite the same as with the LR method above, but quite similar. Not quite as intense, either. Of course, if the sides get summed, the effect cancels itself out and all we get is the dry. However, if we sit close to either speaker, we’ll hear phase cancellations, coloring the sound (and this coloration warbles as we’re used to with a mono chorus effect).
Sometimes when we use a chorus we very much want the coloration. Other times, we mostly want the image. However, you can’t get the image effect without some risk of getting the color. By choosing between LR and MS, we can choose whether we get the color as we move from center, or whether we get the color when summed to mono (e.g., moving away from the speakers). I find it’s best to use an effect like chorus when it’s OK to get some coloration, like on a Rhodes or electric guitar. And this is true regardless of whether I’m using it with LR or MS technique.
I hope this sheds a bit of light on the subject! To sum up, there’s no inherent coloring when using MS mike technique. That only comes into play when using a stereo imaging technique like delay, pitch shift, or chorus on a mono signal.
Recently Phoo mentioned using comb filter followed by MDA Image. If he used it in LR->MS mode, then that would be using comb filter in MS technique, something I haven’t thought of trying yet – I’ve only done it in LR, the normal way the plugin does it. I’ll have to do some thinking and some experimentation to figure out what that means. One of the advantages to comb filter is that it naturally sums to zero. That is, you get frequencies spread left and right, but all the original frequencies are there. Sum them to mono and they’re all still there. At least, that’s my theory based on an oversimplified model of how they work, but I have more to learn about them and how they work under the covers, and it might turn out to be false. Ah, the list of subjects to investigate never ends!
Final minor point, MS technique was a way we used in the bad old days of hardware to get a stereo effect using a mono reverb. Just split the reverb return to two mixer channels, pan them apart, and invert one channel (doesn’t your mixer have an “invert” button? Gee, mine didn’t either, but the “cue out” sends happened to be inverted phase!) Bingo, nice stereo reverb, which cancels out in mono. However, this kind of reverb image was rather artificial – I much prefer setting “early reflections” different in left and right with n-Track reverb for a much more natural sounding image.
Got it now! That is an excellent small essay there, Learjeff!
Fascinating that the perceived effect of something like chorus can change in those ways. Never considered it before.
Again, many thanks!
Ok, for the slow-man now:
-I have a couple of mics same model, with multipattern (cardioid and eight)
Can i connect them directly to my mixer and record the stereo wave file?
Because i need a M-S processor, isnt?
Well, what is this M-S processor, a hardware unit that i need add to the chain before the mixer, or a software plugin that i can apply to the stereo wav file?
Sorry the slownes understanding.