4 mic array phase concerns

[jnorman]

like several others around here lately, i find myself using a four microphone main arrray in almost every situation these days. i am using a center ORTF pair of KM140s, flanked by a pair of DPA 4090 omnis (spaced about 40cm) - all on one stand mounted on a single bar. (i started using this setup shortly after doing some testing of ORTF vs spaced omnis as a main pair, posting the tests here at GS, and getting a response from ben maas indicating that the two pairs complemented each other well when blended, and that neither pair by themselves was giving a solid picture of the sound stage. and btw, does this type of array have a name?)

i generally use the ORTF pair as the main part of the mix, and blend in the omnis as needed to add depth and width, and to control the amount of room ambience included. it works like a charm - all except for one session i did where i had an odd phase issue. each pair alone was fine, but when blending them, i found i needed to flip the phase on one of the omnis to make it sound right - i could not figure out why, so i thought i would ask some of you more tech-savvy GS'ers to help me understand.

so, with a 4-mic array as described above, what are the possible concerns over maintaining proper phase relationships between the two pairs? where might i be introducing possible phase problmes with this setup, and how can i make sure i avoid phase problems? thanks.

[pkautzsch]

What I suspect to be a possible issue is the relatively close spacing between the ORTF pair and the omnis. Left omni - Left ORTF distance is about 11 cm. As you probably have both panned hard left, that can get you the first notch at 22 cm wavelength (or 1,545 Hz) for sound coming from 90° - for sound from other angles it's going to be different values (shorter wavelenght/higher frequency). The only sound that will not be subject to some sort of (theoretical) comb-filtering is sound from exactly 0°.
How much this matters in practice I don't know. I've only once set up both ORTF and spaced on the same bar, and this was an either/or approach. When Tony Faulkner does similar setups with parallel fig-8 and wider spaced omnis, the fig-8 will not pick up the sound from 90° that's most susceptible.

[boojum]

Originally Posted by jnorman

<SNIP>

i generally use the ORTF pair as the main part of the mix, and blend in the omnis as needed to add depth and width, and to control the amount of room ambience included. it works like a charm - all except for one session i did where i had an odd phase issue. each pair alone was fine, but when blending them, i found i needed to flip the phase on one of the omnis to make it sound right - i could not figure out why, so i thought i would ask some of you more tech-savvy GS'ers to help me understand.

<SNIP>

Jim, as you may know I run this setup a lot. I have a CMC64 ORTF in the middle of a 40cm AB DPA 4061. I like it for the same reasons as do you and it creates little visual intrusion in live work. I was interested in your "phase flip" comment so I tried it on a pull from last night's choir rehearsal. I flipped both tracks of the right side of the DPA 4061 and lost some of the stereo effect. Running it without the phase flip I get a good, open sound.


Here is a short clip one straight, one flipped, same recording. The "flipped" track should be titled "right omni channel" not "right omni channels."

Is this the same effect you are getting??

[jnorman]

yeah booj - normally everything is rock solid with no phase issues. just had the problem one time, and it was only one of the 4 tracks that i had to flip phase on to fix it. yes, usually flipping phase results in loss of bottom end.

glad to hear you are getting some recording done, though if i know you, you are doing it all for free...

thanks for posting the clip - nice job. i did a gig up at a church in woodburn, but i couldnt get a decent sound from it -just too much extraneous noise in the venue. i am starting to think that only really good venues give me an improvement over what i can do in my studio, where i can control everything (though it took me a LONG time to get to where i could create a fairly realistic reverb using digital gear...)

[Gutbucket]

Hi John,
I've commonly run a 4060 omni pair outside of a blumlien center pair (more recently a Tetramic, but usually end up with something between blumlein and wider angled hypercards) and have noticed something similar on a couple ocassions when mixing to stereo. Not quite the same scenario since my center pair in either case is coincident, so only 3 actual mic positions vs 4, and the omni spacing was usually more like a meter.

The solution which worked for me was to EQ the two pairs seperately, omnis primarily contributing bass frequencies with reduced mids and highs and low end rolled off from the center pair. That reduced the phase interaction problems I was hearing by sort of dividing frequency range duties, retained the nice wide room feel and depth of the omnis and cleaned up the mids. If I recall, I added back a bit of top end from the omnis to sweeten the ambient room sound and the top end air somewhat.

Easy enough to play around with.

[rumleymusic]

i generally use the ORTF pair as the main part of the mix, and blend in the omnis as needed to add depth and width, and to control the amount of room ambience included. it works like a charm - all except for one session i did where i had an odd phase issue. each pair alone was fine, but when blending them, i found i needed to flip the phase on one of the omnis to make it sound right - i could not figure out why, so i thought i would ask some of you more tech-savvy GS'ers to help me understand.

Could it be the polarity was reversed on a cable or preamp unknown to you while recording? There shouldn't be any problems combining directional and omni mics in such a way, and the phase relationship tends to be for the most part positive.

[Petrus]

I was also thinking about polarity reversal somewhere in one omni. You could check that by comparing the omni pair low frequency waveforms in DAW.

When I have tried to combine ORTF with AB (bit wider at 60 cm) the result has been better stereo definition but lost clarity in upper midrange. Taping the omni and cardioid mics together and synthesizing a wide cardioid from those has worked better. The spacing and angle can be determined from the excellent "stereophonic zoom" paper which has settings for wide (hypo-) cardioids also.

[Sheikyearbouti]

I recently made my first attempt to do something like that recording a classical guitar (see photo). The AB MK2 pair was enough and I was very happy with what I was getting but I wanted to have just a bit more width (room on the sides) without losing the centre. The flanks were MKH50.

Both me and my customer didn't like the result.

At the end I used the MKH50s at about 4 meters behind the player spaced quite a lot and facing the back of the hall. We mixed just a tiny bit of them and it worked great.

I guess that this kind of setups need quite a lot of experimentation and would probably work better for large sources.

[Unique]

Originally Posted by Sheikyearbouti

I recently made my first attempt to do something like that recording a classical guitar (see photo). The AB MK2 pair was enough and I was very happy with what I was getting but I wanted to have just a bit more width (room on the sides) without losing the centre. The flanks were MKH50.

Both me and my customer didn't like the result.

At the end I used the MKH50s at about 4 meters behind the player spaced quite a lot and facing the back of the hall. We mixed just a tiny bit of them and it worked great.

I guess that this kind of setups need quite a lot of experimentation and would probably work better for large sources.

These 2 pairs have to be on separate stereo bars (therefore different stand or suspension) for at least 4 reasons:

1) ORTF and AB spacing are usually quite different
2) Omni and cardioid microphone distance/height are usually quite different
3) Sennheiser MKH and Schoeps CMC are quite opposite in terms of microphone sound character
4) High frequency diffraction over microphone body is screwing sound capture and pattern of neighbor capsule

[Sheikyearbouti]

Originally Posted by Unique

These 2 pairs have to be on separate stereo bars (therefore different stand or suspension) for at least 4 reasons:

Actually the OP uses one bar. But I guess you mean something like this one (as posted by Sonare and named by him Faulkner array). I have never used anything like it or the one of the OP but I have it for a long time on my to do list

However I don't understand some of your remarks...

Originally Posted by Unique

1) ORTF and AB spacing are usually quite different

Yes, they are, but so what? I am not trying to get the same recording angle with both (nor an ORTF and 40cm AB will).
Actually I was aiming for the larger image of the "ORTF" pair (this by the way is not ORTF as I used super-cardioids and the angle/spacing was more like of a NOS).

Originally Posted by Unique

2) Omni and cardioid microphone distance/height are usually quite different

I don't understand. Do you mean that they have to be placed at different distance to get similar dry/wet image?

Originally Posted by Unique

3) Sennheiser MKH and Schoeps CMC are quite opposite in terms of microphone sound character

I don't find them opposite and I don't see a problem here. I combine microphones of different makes with different character all the time and I like that. However I don't think that (if the mkh50s were mkh20s) there is a huge difference between those. They are both great flat mics and despite their obviously different sonic character sound very similar in a way. You can hear a comparison of the mk2 and mkh20 I made during the same session here.

Originally Posted by Unique

4) High frequency diffraction over microphone body is screwing sound capture and pattern of neighbor capsule

Yes. This probably is true. I was shocked when I first saw some graphs in one AES paper (I think on Neumann's web) showing the effect of a singer's head on the polar pattern of a cardiod.

However, this was just an experiment and it didn't work at all. I had as well phase problems and that is why I posted a pic here.