Off Axis Mic Rejection

I did an experiment with a bunch of mics recently to measure the off axis rejection.
Basically this is a practical experiment for the situation of recording a singer playing acoustic guitar where you want to isolate the guitar and vocals as much as possible by directing the null at the opposite instrument / voice.
Off course this is dependent on the room and looking at the results, you could  think that the -17 dB rejection may be the minimum possible due to the room sound.
I used a tuner with an audible tone in front and to the side of each mic (A 440  at 6"). The dB rejection is the difference in the recorded level between the front and side.
I love using the sm7b but clearly it doesn't isolate well!


Ribbon          Fig8                 -17 dB
SM7b         Cardiod         -4  dB
Beyer m260 HypCard         -13 dB
Re55         Omni         -2  dB
Shure 330 SuperCard      -11 dB
Beyer m201 HypCard         -9  dB

DaleM7         Cardiod         -7    dB
                Fig8                 -17  dB
GrpBuy C12 Cardiod         -4    dB
                Fig8                 -8.5 dB
TimC C12       Cardiod         -7    dB
              Fig8                 -17  dB
BzNz U67      Cardiod         -7    dB
              Fig8                 -13  dB

That's interesting about the SM7.  I find off axis drum leakage sounds bad compared to an omni, yet it's so close to omni in your test. 

I only measured at the one frequency, which would might explain it.
A good omni mic has pretty consistent pattern at different frequencies while a cardiod might not.

For example, the tuner can change frequencies in the test tone. In playing around, I saw that a md421 had -11 dB rejection at A (440) and -6 dB rejection at Bb. Only semitone higher.  It seemed weird at the time so I just set that mic aside, but now thinking about it maybe it was a true result.

I've always went with Figure 8 in these situations with the null pointing at the performers head. Your test seems to bear that out.

You might consider the Beyer M88 and Sennheiser 441 as hyper-cardioid candidates.

Interesting that the group buy CK12 doesn't have a deep null like Tim's capsule.

Regards,
Mark

I've always went with Figure 8 in these situations with the null pointing at the performers head. Your test seems to bear that out.

Click to expand...

A big reason I've been looking into this is for a situation where picking up from the back side is undesirable. Recently recorded an Americana band, playing live, with a quiet singer/guitarist, and drums 30ft away in a big barn. The ribbon on guitar was not working due to drum bleed.
Now that I have built some nice sounding multipattern condensors, though, dialing in the 'optimum' to reject some back and some off-axis is possible. The coolest would be to adjust the pattern from the control room while listening!

You might consider the Beyer M88 and Sennheiser 441 as hyper-cardioid candidates.

Click to expand...

I had both of those mics in the past and ended up selling them. I didn't find either to be very useful for some reason. i Tried the 441 quite a bit in this kind of application before selling it. It just didn't sound great to me.

Interesting that the group buy CK12 doesn't have a deep null like Tim's capsule.

Click to expand...

Before the test, I was wondering if a pricey capsule would have better front / back matching, and better rejection. Sure enough.

dmp said:

I've always went with Figure 8 in these situations with the null pointing at the performers head. Your test seems to bear that out.

Click to expand...

A big reason I've been looking into this is for a situation where picking up from the back side is undesirable. Recently recorded an Americana band, playing live, with a quiet singer/guitarist, and drums 30ft away in a big barn. The ribbon on guitar was not working due to drum bleed.
Now that I have built some nice sounding multipattern condensors, though, dialing in the 'optimum' to reject some back and some off-axis is possible. The coolest would be to adjust the pattern from the control room while listening!

Click to expand...

In those situations you may not have any choice but to isolate the singer from the rest of the band.

I agree about the remote pattern adjustment, that would be sweet.

Regards,
Mark

The RE20 has very good rejection (my choice for live scratch vocals) and also sounds not half-bad on acoustic guitar. Also no proximity, so you can get it fairly close. A pair of those might work.

Dylan

dmp said:

I only measured at the one frequency, which would might explain it.
A good omni mic has pretty consistent pattern at different frequencies while a cardiod might not.

Click to expand...

Have you measured at 90° from the capsule or 90° from front of the mic?

I tried to place the sound source 90deg from the capsule. In some cases I had to estimate if the capsule wasn't visible through the grill.

I've checked the specs; your results are consistent.

I think the issue with the 7b rejection might have to do with how far back the capsul sits in respect to the "front" of the mic.  I'd be interested to hear if your off axis was on the side of the pop filter or on the side of the capsule mount.  I recall seeing a plastic barrier of sorts that hooked onto the set screws of the 7b a while ago for this very purpose.  Basically looked like a black plastic "U" that hugged the pop filter (not the bulbous one the flat one).  Dont recall if it was a DIY thing or a commercial thing, but might help squeeze off an extra db or two.  Love that mic also, great sound.

Well, in fact a cardioid is supposed to reject -6dB at 90°, so it's not that far off.
And as dmp has found out, the directivity pattern of a phase-shift based mic in a domestic environment is quite ragged.
I think that type of measurement should be done with filtered noise or warbled tones.

I don't use the foam windscreen on the sm7b. With it off, it is possible to see the capsule and that is what I tried to be 90deg off axis from. 3" back or so.
Interesting idea for a mic mounted baffle - reminds me of a story I heard of a horizontal baffle (plexiglass?) built around the singer with the singers head poking through...
The whole inspiration for this experiment was the poor rejection of vocals when micing guitar with the sm7b.

It will be interesting to try some of these other mics in recording and see if the findings hold up in practice. I'm keen on trying the m260, since it captures a pretty nice acoustic guitar tone.

dmp said:

The whole inspiration for this experiment was the poor rejection of vocals when micing guitar with the sm7b.

Click to expand...

I think the same disappointment is to be expected with any cardioid mic. Hypercard (supercard) have a more rapid angular decrease; however I fear the off-axis sound would be worse.

It will be interesting to try some of these other mics in recording and see if the findings hold up in practice. I'm keen on trying the m260, since it captures a pretty nice acoustic guitar tone.

Click to expand...

Yes, it's a very interesting experiment, since we're often confronted with this situation. Thank you for keeping us informed.

FWIW, I LOVE using ribbons and fig 8 patterns setup in funky ways for this kind of recording.  The royer ribbon mics are especially amazing at this.

Im sure this is a given for some, but others may not have seen or tried it, so...

Taking the fig 8 mic, set it on on its SIDE (not up and down)  the vocal mic gets rotated in a way that the side of the mic is pointing towards the guitar and the guitar mic's side is pointing up towards the vocal.  It wont get 100% separation but its quite enough that you can mix them without one interfering too much with the other.  Only issue you might run into is phase issues (esp if the artist moves around a bit). But thats gotta be taken care of before recording (unless you like nudging wavforms around on the screen by samples... and even that doesn't always take care of a problematic setup). 

That being said, I've never used the 7B for these duties, but if that mic sounded that much better than the alternative, I'd probably have the artist separate the takes (guitar/voice)... they hate doing that, but sometimes it just sounds way better.  Also good to try an MS setup for recording the guitar.  Great natural wide sound from a single guitar.

I have tried the ribbon mic with its null oriented towards the guitar. It's fine when the player is sitting still.
Anyway, if he plays too soft, it never gives good results.
In fact most of the time I end up not fighting against spillage but rather working with it. It becomes a matter of finding mics and mic position that work together positively. I get good results using an SDC (ATM450 or MTG294) positioned closer than usual (the vocal mic will more or less compensate the resultant lack of space) and an SDC for vocals in the standard position (6-10 inch).
But I must say that soundwise, separate takes are always better, although the performance may be lesser.

In fact most of the time I end up not fighting against spillage but rather working with it.

Click to expand...

That's very cool. I would like to work on that as well as improved isolation.
There was a really interesting article in TapeOp last year of Gillian Welch and David Rawlings. recording with four mics. They talk about whenever they have to tear down and set up it takes a long time to get the mic placement just right.

SDC for vocals? I never get good results with SDC compared to LDC

Sorry, typo, I met an LDC.
A long time ago, when my best mic was an AKG C451, I used it on vocals with some success, though.