Discussion in 'Recording' started by lyrictenor1, Feb 16, 2009.
Is it possible to do MS with 3 cardioids instead of 1 cardioid and 1 figure 8?
What inputs do you have?
3 separate? Great.
Point your "S" pair at 180° to each other and invert one signal in your DAW/console. Blend with the "M".
(At least, I think so).
Yes, I have 4 inputs.
Would I still need to invert one of the side channels since they are seperate channels with seperate cardioid pick-ups (one for left and one for right)?
You would need to invert it before you add it with the other side mic.
(I would do that then treat them as normal M/S, rather than specially mix them to begin with - depending on your software ofc)
The upshot is, it's possible but less than ideal.
Cardoid dynamics aren't designed to pick up massive detail.
I think you'll find that your side facing microphones should have 1 phase inverted. But you will need to combine these two microphones into a single track. Then, that becomes your Side track. Then that would be combined with the Mono cardioid. Keeping the side microphones on separate sides is not create the phasing magic that is required with only 2 tracks. One being side. The other being middle. Then you can put it to the MS decoder in software. The previous explanation was incorrectly hypothesized. A single bidirectional microphone is a single track. It's not a pair. You want to take a pair of microphones and create a single track to create the figure of 8 microphone, electronically, in software. Otherwise, it ain't MS.
MS. as in middle side Remy Ann David
So this is my understanding:
Record the mid as one channel.
Record the two sides as two seperate channels. Then, invert one channel and then bounce the two side channels into one track.
Does that sound correct?
this is simpler then it seems.
it's just mono + stereo (albeit a strange stereo)
so it's M+S instead of MS
the two stereo cardioids already have the reversed polarity since they facing opposite directions
you don't need to mix the stereo tracks and then use MS processing
just record to 1 stereo and 1 mono track
blend to taste
No, the polarity on the R-facing cardioid needs to be inverted because it is facing in the opposite direction to the L-facing one.
If you consider the two S-mics as one, a positive pressure increase located at 270 degrees must give a positive electrical output on the L mic and also give a positive electrical output from the R mic if there were not a null in the pattern. Similar for a pressure increase from 90 degrees, but this should give a negative electrical output from both mics.
Unless you want to use a standard MS decoder, you can account for the polarities in your decoding of the signals. If the raw mic signals are M (mid), LF (left facing cardioid) and RF (right facing cardioid), for your own decoder the formula is:
L stereo channel = M + LF - RF
R stereo channel = M - LF + RF
This can be implemented on an analog mixer by using buses if you have phase inversion capabilities, or easily on a digital mixer by using 5 input channels.
To use a conventional MS decoder,
M = M
S = LF - RF
Note that here you have a straight polarity inversion of the RF mic before it combines with the LF mic to form the S channel.
Be warned, however, that using a pair of cardioids to generate the S sound field gives a different result from using a single fig-8. Stereo imaging betwen 0 and 90 degrees either side of the centre line is different, so you will need to use a different amount of (LF-RF) than if it were from a fig-8. You also need to ensure the two side-facing cardioids have their diaphragms co-planar.
with all respect, but I disagree.
maybe I'm wrong, maybe someone else could share their view
or maybe we are suggesting different things altogether.
suppose both stereo mics are omnis
but positioned in opposite directions
so, in an ideal world: Lf= - Rf
as you say (in MS terms): Lstereo = M + Lf - Rf
true, I agree
this equals to
Lstereo = M + Lf - ( - Lf) = M + 2* Lf = M + 2* Rf
if you would first reverse the polarity on one of the side pairs
the equation would be:
Lstereo = M
...which is wrong
--- for future reference: I was wrong, see replies
Now for my point of view, hopefully for the greater good of practicality:
Don't bother with MS encoding and decoding
simple record the stereo-sides as stereo
and record the mono as mono
yes it's 3 tracks instead of 2
but it's much simpler
hopefully I don't sound to cocky
English isn't my primary language
always willing to learn
Yes, your equations hold for omnis (pressure-operated microphones), and correctly show that it is not meaningful to use them for M-S. With an omni, a positive pressure pulse from any direction causes a positive output, so subtracting the signals from two microphones in the same sound field will result in cancellation (of sorts).
A figure-8 is a pressure-gradient microphone, where a positive pressure pulse at the front of the microphone causes a positive output, but the same positive pressure pulse at the rear of the microphone causes a negative output.
A cardioid is effectively a hybrid of both types, the natural phase additions of the omni and fig-8 characteristics resulting in an increase in sensitivity at the front, and a decrease (null) at the rear. However, the phase of the output is still positive all round, as with an omni.
This means that to use a pair of opposite-pointing cardioids Lf and Rf as a substitute for a fig-8, it is necessary to subtract the Rf signal from Lf signal (equivalent to adding them with the Rf phase-inverted), so that you still get a negative output for a positive pulse applied to the R side.
Matrixing issues apart, I have serious doubts about using cardioids to form the S channel of an M-S setup, as the sound of M-S depends so heavily on off-axis microphone response. Many cardioids do not show up well off-axis, both as departure from a flat frequency response and colouration of sound.
Ah, I guess I'm beginning to understand what I should be reading more about.
a cardioid is the summation of an omni and a fig8
so, practically speaking:
if you want to adjust the mono component of a stereo recording you're better of using a decca tree like assembly with omnis or sub-cardioids (a spaced pair + one seperate center mic)
still learning and loving it
isn't it true that the summing of two cardioids opposite direction, one reversed polarity, equals theoretically in one omni (neglecting the polar frequency response)
Yes, summing two perfect cardiods will make an omni. If the response of the cardiod is 1+sin(theta), then the sum with them 180 degrees out of phase is 1+sin(theta) + 1+sin(theta+180). We want to make sure this is 2, so 1+sin(theta) +1+sin(theta+180) = 2 => sin(theta)+sin(theta+180) =0 which we know is true because sin(theta+180) = -sin(theta)
The difference of two cardiods can also be calculated, and you will find it makes a figure 8. So mathematically it is the same, but most cardiod microphones are not perfect cardiods (as was mentioned earlier especially at higher frequencies) and your MS will not have as clear a stereo image in the upper mids and above as with a traditional MS with a figure 8.
Oops, I didn't read GG's statement right. The answer is no. Summing two cardiods in opposite directions same polarity (sum) is an omni. reverse polarity (difference) is a figure 8.
I haven't written out the equations, but my first reaction for the result of combining opposite-facing cardioids is:
In-phase: omni (but elliptical)
That's funny that we wrote the same thing at the same time.
Must be an EE thing.
thank you very much all
As an EE, i did some doodling with Excel last night and plotted out the opposite-facing cardioid patterns with in-phase and out-of-phase summing. They showed omni and fig-8 respectively, as we (both) had said.
How close to theory it would work with two real-world microphones or how good it would sound, I wouldn't like to say.
The direction the microphone is facing does not change it's "polarity".
Separate names with a comma.