making cardioid from omni + fig-8

Recording Techniques, People Skills, Gear, Recording Spaces, Computers, and DIY

Moderators: drumsound, tomb

Post Reply
User avatar
chconnor
takin' a dinner break
Posts: 170
Joined: Thu Jan 04, 2007 4:23 pm
Location: Bellingham, WA
Contact:

making cardioid from omni + fig-8

Post by chconnor » Tue Sep 16, 2014 11:36 am

(Nerd-out warning; abandon all sense of utility ye who enter.)

Question for anyone interested:

Back in the day, they made a cardioid pattern by combining an omni with a coincident figure-8. Since the omni hears in-phase all around and the fig-8 hears out-of-phase from the back, the result is that sound from the back nulls, and voil?, cardioid.

Nowadays cardioid capsules are built to do this by themselves, of course, and we get fig-8 most often by combining a front capsule with a polarity-flipped rear capsule.

But the usual dual-capsule route generates an imperfect cardioid, with a fair amount of rear pickup in the higher frequencies.

Say you wanted to record two harmony singers at once, and they want to face each other. You could put up two cardioids. But what about this kooky plan: put up an omni and a fig-8, coincident, lobes pointing at the singers. For the first singer, you do omni + fig-8, for the other signer, you do omni - fig-8.

(Looked at alternatively, it's just an omni mid-side with the singers on the sides.)

I'm wondering if there is any advantage to this... better rear rejection? Easier to stand close together since the mics are vertically over each other? The polar plot has a perfect null at the back, but that's just the math; obviously as the frequency varies it won't be total. Maybe the practical reality is that it is worse than regular ol' cardioids for whatever reason? Anyone know?

One way to find out, of course, and i plan to do the test some day. Just wanted to post because I thought it was interesting. Maybe in some contrived situation it would be handy (you have a single ribbon mic you love and want to use on both singers but want separation of tracks, or you don't have two cardioids but do have an omni and fig-8, etc.)

RoyMatthews
re-cappin' neve
Posts: 778
Joined: Fri Dec 30, 2005 12:07 pm
Location: Sunnyside Queens, NY

Post by RoyMatthews » Tue Sep 16, 2014 11:51 am

Sort of reminds me of my thoughts once?

http://messageboard.tapeop.com/viewtopi ... highlight=

Thoughts?.

Are you looking to record them in stereo? If so just go with MS or XY or Blum or whatever. Are you looking to record them in mono? Then, I dunno, Just go with an omni and move them to balance. You could consider 2 fig 8's in a Blumlein setup but in mono. In the end it's like omni but with nulls at the floor/ceiling. That's more for room control and not vocal balance.

Your way might work but wily some polarity issues you might not gain much. I'd just give a pass on an omni mic and see if the balance works. Move whoever is louder back and then double. In the end you'll get the same result with less thinking and phase weirdness.

Your way, in mono, will cancel out the figure 8s anyway. In stereo, move the singers to the front of the stereo image and you still need to worry about their acoustic balance. Front to back is moot. Omni or Figure 8. Pick one and stick to it.

Sorry. Kinda day drunk. In the end, Audio doesn't have a front/back perspective. It has depth. So you're in either stereo, which is L/R and depth from the listener's POV or mono. 5.1 notwithstanding.

Does this make sense? Feel free to prod me on this.
"If there's one ironclad rule of pop history, it's this: The monkey types Hamlet only once."

RoyMatthews
re-cappin' neve
Posts: 778
Joined: Fri Dec 30, 2005 12:07 pm
Location: Sunnyside Queens, NY

Post by RoyMatthews » Tue Sep 16, 2014 12:11 pm

Just want to add?

Yes, you can create a cardioid pattern from a mix of Bi/Omni. See the Josephson 700 series. http://www.josephson.com. Any pattern is essential a mix of Bi/Omni. But you have to know why you may want to.

In the end if you want a hard left/right image, with level control, nothing will beat recording 2 mono sources and hard panning. If you want to try stereo mixing then anything will work really. Remember, mathematically, M/S is the same as any coincident pair when all is said and done. Really. Cardioid and Bi is no different than XY. MS with 2 Bi mics is the same as blumlein.

There are pluses and minuses but you have to figure them out.

See: http://www.ribbonmics.com/pdf/technique.pdf

In the end, you won't "break" the L/R barrier. And stereo recording you will do will add "middle", center, information.

Ugh. I feel like I'm not making sense. Keep in mind I haven't recorded in a while but feel like I have a good grip on stereo micing.
"If there's one ironclad rule of pop history, it's this: The monkey types Hamlet only once."

User avatar
chconnor
takin' a dinner break
Posts: 170
Joined: Thu Jan 04, 2007 4:23 pm
Location: Bellingham, WA
Contact:

Post by chconnor » Tue Sep 16, 2014 12:27 pm

Thanks for the response -- my point is just to wonder if there is a functional difference between two cardioids facing away from each other and my theoretical setup which uses an omni and fig-8 (with O+8 and O-8 as the two channels) to achieve the same thing. I'm not trying to get a stereo image of anything, just two isolated tracks.

In theory, the omni/fig-8 method would produce perfect nulls in the rears, but that graph is just an idealized computation. E.g. the average cardioid capsule has a perfect-looking cardioid plot in the mid-range, but you start seeing the signal creeping in from the rear when you go low or high. I'm wondering how the omni/fig-8 would do in comparison.

I'll post if i ever get around to testing it...

RoyMatthews
re-cappin' neve
Posts: 778
Joined: Fri Dec 30, 2005 12:07 pm
Location: Sunnyside Queens, NY

Post by RoyMatthews » Tue Sep 16, 2014 12:52 pm

I see.

Drunk response?


2 Cardioids hard L/R in phase, 1 Omni

2 Cardiods out of phase, 1 Figure 8/Bi

I left the word "phase" in there but I meant "Polarity".

I just want it known, I applause the thought. Seriously...

But you're just trying to make a carded pattern when the company who deign the pattern figured it out.

I mean, if you have an Earthworks (pretty ideal omni) and an 4038 (beloved Bi mic) then fuck it, see what happens when you make a cardioid mic out of the combo. But if you have a u87 then just just that.

Yeah, in theory, omni+fig 8= cardioid. But that's what all mic companies strive for.

In the end it all depends on the mic combinations. I can't speak how a 414/eb in bi combines with an 87 in omni.

Try it. See how you feel. That all that matters. But you won hit on much earth shattering info.

In the end, omni+fig 8 _wont_ hit on anything idealized unless they were the same mic and if so, that's cardiod. If not, it might be cool bit veeeeeeery specific to the combo. Earthworks 40k + a 4038 won't be ideal.

Again, try it. it's all the combination of what you have and what works. But, as a drunk guy who's been an engineer for 20 years, talking mathematically, meh. it's not worth the effort.

There isn't, on paper a difference, between 2 cards, facing back to back as there is an omni/bi at the right levels. But there can be certain distortions/ differences that happen with the signal path you have that produce certain, pleasing results. But from what i've learned, you haven't happened upon a new paradigm.

I sincerely don't mean to squash any innovation. You might get an inspiring sound. But you won't get a new, mono, signal.

FYI, Personally, I love this type of thinking and I love to think about it. But, unfortunately, this is pretty basic thinking that we as engineers have forgotten about. Ala Alan Blumlein.
"If there's one ironclad rule of pop history, it's this: The monkey types Hamlet only once."

RoyMatthews
re-cappin' neve
Posts: 778
Joined: Fri Dec 30, 2005 12:07 pm
Location: Sunnyside Queens, NY

Post by RoyMatthews » Tue Sep 16, 2014 12:53 pm

Oh man. I need to work more because i drink too much. But fuck it. figure it out.
"If there's one ironclad rule of pop history, it's this: The monkey types Hamlet only once."

User avatar
chconnor
takin' a dinner break
Posts: 170
Joined: Thu Jan 04, 2007 4:23 pm
Location: Bellingham, WA
Contact:

Post by chconnor » Tue Sep 16, 2014 1:19 pm

RoyMatthews wrote:But from what i've learned, you haven't happened upon a new paradigm.
Yeah, I suspect this is the case. :-) Besides the contrived situations i described above where it might be handy, it's hard to see it as much better (or even different) then using a couple cardioids. But when/if i do the experiment, i'll let the world know.

ashcat_lt
tinnitus
Posts: 1082
Joined: Thu Jan 04, 2007 1:54 pm
Location: Duluth, MN
Contact:

Post by ashcat_lt » Thu Sep 18, 2014 5:44 pm

This sort of polarity based cancellation requires pretty close correlation between the two sources. That is, it will cancel best when all things (short of polarity) are closest to exactly the same. Any variation in frequency or phase response between the two mics will cause things to "leak through", which is I think a big part of the reason why we have the frequency-dependent polar response in directional mics in general.

So, basically, if one of the mics is brighter than the other, you'll have treble bleed from the back side, if one has an extended low end compared to the other, you'll have less directional response in the bottom. Interestingly, it doesn't really matter which mic has more or less of what, as long as they're different, they can't fully cancel.

You might then try to "fix it" by EQing one to better match the other, but since most filters cause phase changes, it will only improve things accidentally, and is probably more likely to make things worse.

It's probably obvious, but timing (distance) differences will also complicate matters, and have a greater impact at higher frequencies.

I tend to think you'd be better off using 2 figure-8s at right angles to each other. The singers won't be facing directly at each other, but will be in each other's null spot, plus you'd have less issues with reflections off each other's faces. And yes, I know that was a hypothetical, I'm just saying...

User avatar
chconnor
takin' a dinner break
Posts: 170
Joined: Thu Jan 04, 2007 4:23 pm
Location: Bellingham, WA
Contact:

Post by chconnor » Thu Sep 18, 2014 5:51 pm

Interesting. I had assumed that the same kind of imperfections would be present in a cardiod capsule, which if i understand things correctly, is achieving the same kind of thing by opening ports in the back plate and allowing phase cancellation and that would result in similar things (e.g. proximity effect). But you're estimating that the cardiod capsule is likely to be better tuned than the wonky two-mic method. Makes sense.

Also, I was assuming the omni and fig-8 would have to be precisely level matched for the same reasons you state; if one is a little too loud or quiet things wouldn't null properly...

ashcat_lt
tinnitus
Posts: 1082
Joined: Thu Jan 04, 2007 1:54 pm
Location: Duluth, MN
Contact:

Post by ashcat_lt » Thu Sep 18, 2014 6:07 pm

chconnor wrote:Also, I was assuming the omni and fig-8 would have to be precisely level matched for the same reasons you state; if one is a little too loud or quiet things wouldn't null properly...
Yep, that too, but it seems that might be easier to deal with.

User avatar
Scodiddly
ghost haunting audio students
Posts: 3499
Joined: Wed Dec 10, 2003 6:38 am
Location: Mundelein, IL, USA
Contact:

Post by Scodiddly » Fri Sep 19, 2014 5:09 pm

Most any cardioid that doesn't do the large-capsule front-and-back trick will just use small ports around the back of the diaphragm. The time delay for sound getting to the back, along with some acoustic resistance (some kind of fabric usually) creates the directional pickup pattern.

Which doesn't mean that it wouldn't be fun to create directional patterns some other way and see what they sound like.

User avatar
Marc Alan Goodman
george martin
Posts: 1400
Joined: Tue Oct 28, 2003 7:57 pm
Location: NYC
Contact:

Post by Marc Alan Goodman » Mon Feb 23, 2015 10:10 am

Scodiddly wrote:Most any cardioid that doesn't do the large-capsule front-and-back trick will just use small ports around the back of the diaphragm. The time delay for sound getting to the back, along with some acoustic resistance (some kind of fabric usually) creates the directional pickup pattern.
That's how essentially all current multipatern mics make cardioid (josephson's excepted).

I don't have a lot of experience comparing cardioid patterns, however I have spent a lot of time listening to omni derived from two opposing diaphrams, versus a single capsule with no ports (pressure omni) like the neumann 53/63/73/83, M50, Avenson/Stapes omnis, Earthworks etc. In my opinion pressure omnis sound WAY better on almost all sources. I actually can't think of a situation where I've preferred additive omni (making up a term there but whatever). I have to imagine you can get if not superior at least significantly different results by deriving cardioid a different way. Now that I think of it I should probably get around to seriously trying out those Josephson mics...
Which doesn't mean that it wouldn't be fun to create directional patterns some other way and see what they sound like.
Bingo.

Post Reply

Who is online

Users browsing this forum: No registered users and 51 guests