Omnis in XY. Does it make sense?

[Matti]

OSS4 - Technik (Jecklin-Scheibe mit vier Mikrofonen)
Zusätzlich zu den zwei Kugelmikrofone werden zwei
nach hinten gerichtete Nierenmikrofone verwendet, die
möglichst an den gleichen Orten wie die Kugeln plaziert
werden müssen (optimal zu realisieren nur mit
Mikrofonen mit seitlicher Einsprache)
Der seitliche Aufnahmewinkel wird so verbreitert. und
gleichzeitig werden mehr Raumschallanteile aufgenommen.
Diese Anordnung erlaubt eine nähere Aufstellung und
eine bessere Abdeckung von breiten Klangkörpern
(Sinfonieorchester).

Google:
OSS4 - Engineering (Jecklin disc with four microphones)
In addition to the two microphones, two ball
used rear-facing cardioid microphones, the
placed as possible to the same places as the balls
must be (at best realized only with
Microphones with side inlet)
The lateral view angle becomes wider. and
while more room sound has been recorded.
This arrangement allows for a more detailed definition and
better coverage of large ensembles
(Symphony).

Stange translation but you get the idea


Matti

[MichaelPatrick]

Matti, a photo would be wonderful. Any chances?

[Matti]

Here http://www.mdw.ac.at/I101/iea/tm/scr...03mikrofon.pdf

Matti

[boojum]

Matti - I wish the picture helped. I get the idea that the omni and the card, preferably side address, are closely adjacent. And they are between the omni and the disc?? Funny, I have the paper on my computer but referenced it only for OSS disc construction. Thanks for pointing this out.

[Matti]

-Searching...


Matti

[jacktadoussac]

stichting klankschap

I was intrigued by the above article and the techniques, and I found it's pretty good. I this case I used an XY configuration, but only one of the microphones is directed at me. Basically one mic is facing down, the other up. It sounds fairly balanced to me.

The two mics used are RE20. I put a little bit more gain on the mic not facing me. I may try to transpose that song in the future -- but I was so eager to try something new. Be gone with the horizontal plane.

[boojum]

It sounds like plain mono to me.

[Corran]

From what it sounds like from your description, were you actually at 180 degrees??

And the RE20 is a cardioid mic, not an omni. Two cardioids back-to-back at 180 degrees would be more like a single omni than some kind of stereo technique.

[Mark A. Jay]

Originally Posted by boojum

It sounds like plain mono to me.

Boojum's right.

Either the file was accidentally written with one channel assigned to both left and right, or this renders full-on mono. I can't say as I didn't record the track or write the parent file, but... I took the mp3 and converted it back to .wav, then subtracted the right channel from the left...and what I got was silence punctuated by some brief, low-level aperiodic noise bursts.

Apart from what appears to be some quantization error-based noise, the file I wrote (the delta between the channels) is essentially zero (when I say 'essentially' I mean that you cannot hear a trace of the music - if they were highly similar (but still had differences) then the delta would be non-zero and you could at least hear some of the original content).

Again, if the delta is zero then the content must be the same. If the content is the same, then it must be mono.

[Mark A. Jay]

Originally Posted by Corran

From what it sounds like from your description, were you actually at 180 degrees??

And the RE20 is a cardioid mic, not an omni. Two cardioids back-to-back at 180 degrees would be more like a single omni than some kind of stereo technique.

Two cardiods back to back would seem to approximate a figure-8 more than they would an omni. Granted, there would be lobing, and granted, what happened in such a configuration would have a lot to do with the space between them and whether or not there were any barriers / absorbers between the mics, but pointing them away from one another would not negate the principles upon which the cardoids function.

I seem to recall some folks having posted this as one of the building blocks for an M-S (the cardiods as the "S" component) when a true figure-8 is not available (and using an omni for the "M" component of the M-S).

[ofajen]

Originally Posted by Celloman

I read that Bruce Swedien calls this "fat mono". An interesting sweetener mic technique.

Mike

I recall reading that he used a pair of M49s in omni mode for this, so he was getting more HF directionality than the Earthworks would give.

Look, it's simple, it's coincident, so it sums nicely to mono. It uses omni mode, so you get flat, extended low end and no proximity effect and you get a little bit of directional cues in the HF. It is what is. Try it if you want a full-range sound and a little bit of directional cue, but a fundamentally solid mono sound.

Cheers,

Otto

[Corran]

Originally Posted by Mark A. Jay

Two cardiods back to back would seem to approximate a figure-8 more than they would an omni. Granted, there would be lobing, and granted, what happened in such a configuration would have a lot to do with the space between them and whether or not there were any barriers / absorbers between the mics, but pointing them away from one another would not negate the principles upon which the cardoids function.

If you flip the phase of one of the cardioids, yes you get a figure-8 pattern. Both of them in phase makes omni. This is the principle that multi-pattern dual-diaphragm mics are based on.

[Mark A. Jay]

Originally Posted by ofajen

I recall reading that he used a pair of M49s in omni mode for this, so he was getting more HF directionality than the Earthworks would give.

Look, it's simple, it's coincident, so it sums nicely to mono. It uses omni mode, so you get flat, extended low end and no proximity effect and you get a little bit of directional cues in the HF. It is what is. Try it if you want a full-range sound and a little bit of directional cue, but a fundamentally solid mono sound.

Cheers,

Otto

I'd like to qualify my comments in that I have not (personally) heard a recording using the technique you describe as 'fat mono'. My comments about the sound were based on the mp3 that was posted and what I saw when I did the subtraction.

Because the delta between the channels is zero (apart for the aperiodic quantization noise), like I said, I suspect that one channel was accidentally written to both left and right channels in the file - there should be some difference observable in the signals even if they are in an X-Y configuration but "mostly mono"; they can't be identical over the entire audible spectrum because they do not occupy the same point in space (let alone differences in frequency response due to tolerances between the mics and such).

The delta file indicates that there is no difference between the left and right channels.

[ofajen]

Originally Posted by Mark A. Jay

I'd like to qualify my comments in that I have not (personally) heard a recording using the technique you describe as 'fat mono'. My comments about the sound were based on the mp3 that was posted and what I saw when I did the subtraction.

Because the delta between the channels is zero (apart for the aperiodic quantization noise), like I said, I suspect that one channel was accidentally written to both left and right channels in the file - there should be some difference observable in the signals even if they are in an X-Y configuration but "mostly mono"; they can't be identical over the entire audible spectrum because they do not occupy the same point in space (let alone differences in frequency response due to tolerances between the mics and such).

The delta file indicates that there is no difference between the left and right channels.

I think someone else attributed "fat mono" to Swedien. I've tried it, and it sounds about as I described. I usually don't bother. I'll just track mono if I want a mono sound and use spaced omnis if I want a lush stereo sound, but it is there as an option. Might be more useful with a pair of M49s, but I don't have any.

Cheers,

Otto

[Mark A. Jay]

Originally Posted by Corran

If you flip the phase of one of the cardioids, yes you get a figure-8 pattern. Both of them in phase makes omni. This is the principle that multi-pattern dual-diaphragm mics are based on.

My understanding (and I could be wrong) of M-S is that the phase from the figure-8 mic is to be reversed (because there is but one transducer for the "S" component) and then hard-panned to the other channel; this would make both channels seem to have a positive phase for sound being incident upon the diaphragm (real and virtual). The omni is then panned to center, correct?

Two cardiods placed back to back would seem to result in each diaphragm 'seeing' a native positive pressure as sound strikes each respective diaphragm; inverting the phase of one cardoid would make the phase equal between mics.

Thought experiment: if you picture two cardiod mics in free space, facing apart (180 degrees), and picture a pressure perturbation that is propagating from right to left, then the perturbation that strikes the microphone on the right cardoid will result in a positive pressure (and output), and as it passes to the left cardiod it will result in a negative pressure perturbation and output (because the daiphragms are 180 degrees apart).

Likewise, a wave incident from the left will produce a positive output from the left cardoid, and a negative output from the right cardoid as it passes the right cardoid.

So, it would seem to me that you would not want a phase inversion of one of the cardoids for a 'faux S-component', because doing so would make the relative phase between the mics the same, that is, applying a 180 degree shift to one channel makes that mic seem like it is flipped 180 degrees. In a figure-8 mic, an incident wave from one side produces a positive pressure on the side upon which it is incident, and a negative on the other as it passes by...does it not? In M-S, the signal from the "S" mic is flipped by 180 degrees (and panned hard to the other channel) to achieve what the cardoids achieve when facing back to back.

So, in the classic M-S, if you didn't flip the phase of the "S" mic and simply panned it to the other side, then the phase would be incorrect. Or...have I missed something? You have to forgive me, but my methods of choice are XY, ORTF, anbd binaural, so maybe I'm missing something here...

[jacktadoussac]

I took some photos of the way the mikes are arranged. One is on a Latch Lake overhead stand, the other is on a normal floor stand. The mike suspended on the overhead is pointing down, at 45 degrees, the one on the floor stand points up at 45 degrees. There is a 90 degree angle between the mikes when you look downward at a 45 degree plane.


I think it's doing something and is more than mono. There are separate left and right channels, each mapped to separate mike cables.


NOTE:
There could be a glitch in this that to some updates from TC Electronic. I don't like blaming third parties, but in this case it is well deserved. My computer no longer recognizes the control panel -- after I opted to update the software. There's a chance that internally the control panel that precedes Cubase is blending Mic 1 and Mic 2 into a mono feed that is separately identified as Input 1 and 2. (***This is only speculation)

[Corran]

You are confusing yourself with all the M-S stuff.

In a dual-diaphragm mic, an omni pattern is created by summing two cardioid-patterned capsules equally that are facing in opposite directions. The fig8 pattern is created by the rear element being phase-inverted. All other patterns are a variation on the levels between the elements.

The easiest way to figure this out is to imagine a sound coming from 90-degrees off-axis to both elements. They would both pick up the sound the same. In a figure-8 mic, the side is a null, because one cardioid element gets it "in-phase" and one gets it "out-of-phase," since their polarities are reversed, which cancels out the sound. As opposed to both being in-phase, which would make it not null (omni).

As for M-S, if you were using two cardioids back-to-back to get a faux-fig8 signal, you would flip the right-facing mic's phase, sum both mics to mono, and THEN double THAT track, and pan hard L/R and flip the phase again on the right side. Or you could have 4 tracks and do it manually (left mic panned both ways, the L channel normal and the R channel phase-reversed, and the opposite for the right mic). Then add the M track of course.

Also note in your thought experiment that a sound originating from one side would actually not be picked up by the cardioid element facing the other way, due to the pattern (in a perfect acoustic with no reflections, obviously).

[jacktadoussac]

First of all, the only thing I did has take a pair of cardioid mikes which I sometimes use in an X-Y arrangement, parallel to the floor -- and arrange these in a plane that is angled 45 degrees to the floor.

Presumably, this will pick up more of the room and pick up less of a stereo image.

I don't really care what it is called, and which theoretical model it corresponds to. It may be preferable to simply record with one mike in mono but I think there is a bit of extra depth attained by playing with the angles -- and going away from the rule that stereo mikes should be parallel to the floor.

[Corran]

(I was responding to Mark, not you. I erroneously thought you might have been recording with back-to-back mics. In your pictures you are using simple X-Y but semi-vertically. However the mono-ness of your sample implies possibly that your setup is being summed to one channel rather than true stereo.)

[Mark A. Jay]

Originally Posted by Corran

As for M-S, if you were using two cardioids back-to-back to get a faux-fig8 signal, you would flip the right-facing mic's phase, sum both mics to mono, and THEN double THAT track, and pan hard L/R and flip the phase again on the right side. Or you could have 4 tracks and do it manually (left mic panned both ways, the L channel normal and the R channel phase-reversed, and the opposite for the right mic). Then add the M track of course.

So...

"you would flip the right-facing mic's phase, sum both mics to mono, and THEN double THAT track, and pan hard L/R and flip the phase again on the right side"

1) Flip right-facing phase (call this 'b'); -1*(b) = -b
2) Sum both mics (we'll call the other mic 'a'), thus; a+(-b), or equivalently, [a-b].
3) Create a copy of the track from (2).
4) Pan this signal to the left, left channel gets [a-b]
5) Invert the sum [a-b]; -1*(a-b) = -a+b
6) Pan this hard to the right channel.

This yields:

Left channel: [a-b]
Right channel [-a+b]

What we have (above) is a difference signal in each channel, inverted in phase with respect to one another. In the classic M-S approach, we have one signal that is inverted in phase and placed in the other channel; there is no difference information (but were you to subtract the left channel from the phantom (phase-altered) right channel, the difference would just be a 6 dB gain bump, that is [a] - [-a] = 2[a].

That is, in the figure-8 mic in M-S case, its phase from one channel is inverted and hard panned to the other channel, right? In this case we get:

Left channel: [a]
Right channel [-a]

Were we to leave the cardiods' phases alone it would seem, at first glance that because they are the complement of one another in terms of phase (facing away from one another), that basically the left mic (a) is essentially the same as mic b with a phase inversion, and vice versa. That is, 'a' = '-b', and 'b' = '-a'. From a phase (and mathematical) perspective this seems to make sense, because:

If 'a' = '-b', then 'b' = '-a'. Thus, if I leave the phase of the cardiods alone I get:

Left channel: [a]
Right channel: [b]
- but -
since b = -a, then...

Left channel: [a]
Right channel [-a]

And this is what happens in the classic M-S when you invert the phase of channel 'a' and make it channel 'b'.

I have to think about this "Or you could have 4 tracks and do it manually (left mic panned both(?) ways, the L channel normal and the R channel phase-reversed, and the opposite for the right mic)"

[Washington]

Originally Posted by jacktadoussac

I took some photos of the way the mikes are arranged.

It could be interesting, provided you really do not move at all during the take (a shift of an inch at a distance of roughly ten inches would definitely be quite large once reproduced hard-panned by speaker 10 feet apart).

Still, unless I am mistaken, the RE20 is a cardio. We were talking omnis here.

[Tom McC]

Originally Posted by Mark A. Jay

My understanding (and I could be wrong) of M-S is that the phase from the figure-8 mic is to be reversed (because there is but one transducer for the "S" component) and then hard-panned to the other channel; this would make both channels seem to have a positive phase for sound being incident upon the diaphragm (real and virtual). The omni is then panned to center, correct?
So, it would seem to me that you would not want a phase inversion of one of the cardoids for a 'faux S-component', because doing so would make the relative phase between the mics the same, that is, applying a 180 degree shift to one channel makes that mic seem like it is flipped 180 degrees. In a figure-8 mic, an incident wave from one side produces a positive pressure on the side upon which it is incident, and a negative on the other as it passes by...does it not? In M-S, the signal from the "S" mic is flipped by 180 degrees (and panned hard to the other channel) to achieve what the cardoids achieve when facing back to back

Mark, you overanalyse this and draw the wrong conclusion.

Just consider simple geometry: a cardioid is built up from a circular "+ blob" and a Fig-8 shape with + and - blobs. So adding two coincident, opposite-pointing cardioids causes the circular constituents to reinforce each other and the Fig-8 constituents to cancel each other out. That is, omni.

Subtracting the signals (i.e. one mic phase-reversed) causes the circular blobs to cancel out and the Fig-8 blobs to reinforce That is Fig-8 (pointing in same direction as the 'proper phase' cardioid).

[7rojo7]

I like single capsule omnis better than dual capsule omnis, I like single capsule cardioids with acoustic baffling better than fixed plate charge capsules and I like 'well designed' (some are not symmetrical) ribbons for fig8, I like dynamics in whatever flavor for close mic'ing loud sources
I trust (so it's emotional as well) the physical factor more than the electrical factor, even though I'll admit that electricity is physical.
I like simplicity
Corran is 100% correct in describing how dual capsule microphones work.
This is why I like what I like, although fig8 is more precise with a dual capsule mic and I don't own a symmetrical ribbon, for most purposes I'll baffle the back of it to get hyper cardioid from an LD condensor.

[Corran]

Originally Posted by Mark A. Jay

stuff

I don't really follow the point of your math. If you are trying to prove a point I don't see what you are saying.

[Tom McC]

Mark, some more comment for completeness:

It actually pertains for EVERY "V = a + bcosX" directional mic pattern - e.g. hypercardioid, cardioid, subcardioid, Fig-8 etc.- that the summed signal from two coincident, opposite-facing mics will generate a perfect omni pattern, and their difference signal will generate a perfect Fig-8 dipole. The only changes through the series will be the strength of the resultant mic signal:

pattern equation => sum signal difference signal
omni 1.0 + 0.0cosX 2.0 0.0cosX
subcard 0.7 + 0.3cosX 1.4 0.6cosX
card 0.5 + 0.5cosX 1.0 1.0cosX
hypercard 0.3 + 0.7cosX 0.6 1.4cosX
Fig-8 0.0 + 1.0cosX 0.0 2.0cosX

(N.B. trig is needed for vector summing since angle X as stated above is w.r.t. each mic's main axis. And of course the trivial cases: Fig-8's summing to give a zero strength omni, and omnis subtracting to give a zero stength Fig-8).

Tom
Live at The London Palindrome - ABBA

[marwutz]

This thread is the nerdiest thing I've ever read through on the whole internet. :D

[Mark A. Jay]

Originally Posted by Tom McC

Mark, some more comment for completeness:

It actually pertains for EVERY "V = a + bcosX" directional mic pattern - e.g. hypercardioid, cardioid, subcardioid, Fig-8 etc.- that the summed signal from two coincident, opposite-facing mics will generate a perfect omni pattern, and their difference signal will generate a perfect Fig-8 dipole. The only changes through the series will be the strength of the resultant mic signal:

pattern equation => sum signal difference signal
omni 1.0 + 0.0cosX 2.0 0.0cosX
subcard 0.7 + 0.3cosX 1.4 0.6cosX
card 0.5 + 0.5cosX 1.0 1.0cosX
hypercard 0.3 + 0.7cosX 0.6 1.4cosX
Fig-8 0.0 + 1.0cosX 0.0 2.0cosX

(N.B. trig is needed for vector summing since angle X as stated above is w.r.t. each mic's main axis. And of course the trivial cases: Fig-8's summing to give a zero strength omni, and omnis subtracting to give a zero stength Fig-8).

Tom
Live at The London Palindrome - ABBA

Tom - thanks very much for this. I will have a look; the defining equations are where I needed to start (rather than end up) and I truly didn't think to research them before posting, so this one's my mistake for not having started at the beginning.

Apologies to all.

[Mark A. Jay]

Originally Posted by Mark A. Jay

1) Flip right-facing phase (call this 'b'); -1*(b) = -b
2) Sum both mics (we'll call the other mic 'a'), thus; a+(-b), or equivalently, [a-b].
3) Create a copy of the track from (2).
4) Pan this signal to the left, left channel gets [a-b]
5) Invert the sum [a-b]; -1*(a-b) = -a+b
6) Pan this hard to the right channel.

This yields:

Left channel: [a-b]
Right channel [-a+b]

What we have (above) is a difference signal in each channel, inverted in phase with respect to one another. In the classic M-S approach, we have one signal that is inverted in phase and placed in the other channel; there is no difference information (but were you to subtract the left channel from the phantom (phase-altered) right channel, the difference would just be a 6 dB gain bump, that is [a] - [-a] = 2[a].

So, now it's crystal clear to me.

What I completely glossed over was the fundamental principle of the figure-8 type of pickup - it's a differential (difference) transducer. Indeed, if the instantaneous pressure on either side is the same, then the output of the mic = 0, and therefore, we have a delta-P type transducer.

Thus, my initial assessment that in the math above, we get two difference signals, one inverted with respect to the other was in fact correct. However, what I failed to consider was that signal 'a' (in my example) is by its very nature a difference signal. The mistake's mine for not having thought through the fundamental differences between the cardiod and the figure-8.

I stand corrected.

[ysmgj]

i tried the omni in vertical xy last weekend with the only omni's i have which are little blondies, which also are rather directional when it comes to HF stuff.

i was tracking a 13 string koto in mono to be blended with a 17 string bass koto later on, hence the mono,... and while the blondies are nice i wasnt getting the sound i was after, so on a whim i tried the vertical xy,....

while its certainly not anything more than mono, what i got sounded better than one single blondie. a single blondie might start sounding harsh on louder passages but the vertical xy seemed to smooth it out a bit.

probably to do with what spicemix mentioned on the first page of this thread, doubled surface area, better S/N but with the benefits of these lil omnis.

but it was the last take of the day and we were just messing about,... I'll be tracking again on friday so if we get something post worthy i'll put it up.

[ysmgj]

noob question:

i have 4 blondies, so if i wanted to do this:

verticalXY-A and verticalXY-B

and then put that into either AB, or ORTF, etc,etc,...

but i only have a 1 nice 2 channel BG1, and i want to send all mics through it,...

how much power do these little things take, would i be able to power the pair on one channel of phantom if i managed to split the cable?