Off axis response SDC microphones

[MicUlli]

Hello all,
it would be fine to catch some opinions regarding the effects of off axis response on the sound (stage) of different stereo mic setups. As a good starting point I have attached some off axis responses of 4 different mic types:
Audio Technica AT2031 (107€), Neumann KM184 (750€), Telefunken TC600 (50 years old, former price 158DM), Primo EM21 (40 years old, former price 98DM).
We all know that the response on axis can (and should) be easily corrected by proper EQ. That is NOT true for off axis response.
For my taste the 50 years old Telefunken TC600 performs best. This mic was developed by MB Electronics, now better known as MBHO

Comments very welcome

 

[kingkorg]

It doesn't make much sense unless you publish the off axis responses together with 0°. 0° normalized of course, and off axis adjusted for level in respect to 0°.
That Telefunken looks like almost omni or wide cardioid. Off axis will also vary widely with distance from the source.

It would make sense to provide 180° as well, which would indicate type of capsule.

If a capsule is a hypercardioid it would logically show better rejection and FR at 90°, but a cardioid would beat it at 180°.

 

[MicUlli]

@kingkorg: ???

The responses ARE already normalised to 0° frequency response. So you can imagine 0° FR for all 4 mics as a straight line at 0dB in the diagrams.

The capsule type is self explaining thru 90° response: AT2031, KM184 and EM21 are cardioids (approx. -6dB) , TC600 is hypercardioid (approx. -8,5dB).

The dB-values in the legend are for 1 kHz. Mesasurement distance was 1m, data were taken in a church (2500m3 volume), quasi anaechonic, by windowing impulse responses before first echo arrived.

As already stated 0° FR is not of importance for this discussion because it can be corrected by proper equalisation.

The diagrams show the differences of the 4 mics very clearly. What do you think, is it (barely) audible?

 

[kingkorg]

If i understand you correctly, you used each mic 0° as a reference for it's off axis response? You divided off axis responses by the 0°, or used 0° as a cal file.

I am a big proponent of importance of off axis response! It's not just important it's crucial even when recording at 0°. It makes much sense that TF will have better rejection at those angles since it's HC. Thanks for heads up on that mic, i wasn't aware of it. But as a HC it will have less rejection from 180°. The energy has to go somewhere, it would be interesting to see how it performs at 180° compared to others.

 

[MicUlli]

Yes, all graphs are scaled to 0° FR. Find attached the responses at 135° and 180°.
For low frequencies the FR is influenced by the proximity effect, should be always kept in mind..
At 1kHz and 135° TC600 has the maximum rejection because of hypercardioid construction.
AT2031 has poor rejection at 180° and high frequencies.

 

[MicUlli]

Here are the graphs for 0° frequency response to complete the measurement set..

 

[kingkorg]

Awesome!

 

[ubxf]

Would the MB Electronic C527 have a similar capsule ?

 

[Paul W]

But as a HC it will have less rejection from 180°. The energy has to go somewhere, it would be interesting to see how it performs at 180° compared to others.

I don't know what you mean by "the energy has to go somewhere."

Is there a conservation law at work that I don't understand?

I would think that when you're combining front and rear signals to create a polar pattern, you're combining information, not energy, and you can combine them in ways where things cancel out, or reinforce, without the missing energy necessarily poking out somewhere else.

For example, at low frequencies where front and back both act more like omnis, flipping the polarity of the back signal will switch between mostly reinforcing low frequencies and mostly cancelling them out.

It also seems like your very neat trick of EQing the back signal differently to make rejection more consistent across different frequencies is another violation of the idea that "the energy has to go somewhere."

 

[kingkorg]

I don't know what you mean by "the energy has to go somewhere."

Is there a conservation law at work that I don't understand?

I would think that when you're combining front and rear signals to create a polar pattern, you're combining information, not energy, and you can combine them in ways where things cancel out, or reinforce, without the missing energy necessarily poking out somewhere else.

For example, at low frequencies where front and back both act more like omnis, flipping the polarity of the back signal will switch between mostly reinforcing low frequencies and mostly cancelling them out.

It also seems like your very neat trick of EQing the back signal differently to make rejection more consistent across different frequencies is another violation of the idea that "the energy has to go somewhere."

Two different things. At physical level, the capsule is canceling what comes from the rear by delaying the sound that comes from the rear, at the diaphragm itself the waves cancel each other out and the energy turns to heat.

At the level of plugin yes we are indeed talking about information instead of energy. Different approach, just like when you flip the phase on one of identical channels in the daw and you get null.

Cardioid and hyper are both formed by combining omni and F8 at varying levels. When you look at any cardioid vs HC graph you see that at the cost of having tighter pattern in the front, you get less rejection at the rear, by the same amount. You have increased the level of F8 component compared to omni. The amount of energy is preserved, the distribution is shifted.

In order to get more rear rejection i introduced the second stage in the daw, which is why i called it "second order cardioid".

You can also achieve the same thing by placing large mass of rock wool behind the mic (without blocking the vents) to block all the sound coming from the rear brute force style. You haven't broken any laws, it's just the sound coming from the rear got converted to heat at another stage.

But if you compare two capsules with comparable construction which are both 1st order directional mics, there is no second stage to achieve superior rejection, you have just shifted the distribution of the rejection.

I hope this makes sense.

 

[k brown]

Off-axis response is particularly important for ORTF; of the common, formal near-coincident stereo pairs, it uses the most of the mics' off-axis sound, since they are at 110 degrees, whereas with DIN and NOS they mics are at 90 degrees.

Off-axis response is of particular importance in Classical recording when using cardioid pairs, as regardless of stereo configuration, most of the sound of the room is being picked up by the back of the mics. If you're in a very nice space, you want to record it as accurately as possible.

Aggravatingly few mfgrs publish 90 and 180 degree responses - even the 'big boys'.

__________

Interesting side note regarding 180 cardioid responses. I once had to record a string orchestra in a church whose 'modernization' had rendered it's reverberation very midrange-heavy. Omnis sounded absolutely dreadful in there. Since most cards have quite scooped response at 180 degrees, I used an AB cardioid pair aimed straight ahead, so the room was picked up almost entirely by the back of the mics, and the scooped sound neatly compensated for the mid-heavy room.

 

[kingkorg]

Off-axis response is particularly important for ORTF; of the common, formal near-coincident stereo pairs, it uses the most of the mics' off-axis sound, since they are at 110 degrees, whereas with DIN and NOS they mics are at 90 degrees.

Off-axis response is of particular importance in Classical recording when using cardioid pairs, as regardless of stereo configuration, most of the sound of the room is being picked up by the back of the mics. If you're in a very nice space, you want to record it as accurately as possible.

Aggravatingly few mfgrs publish 90 and 180 degree responses - even the 'big boys'.

I've long been curious to see an independent measure of the 180 degree response of the MK012 hypercardioid; the published one seems almost impossibly flat, compared to most.

__________

Interesting side note regarding 180 cardioid responses. I once had to record a string orchestra in a church whose 'modernization' had rendered it's reverberation very midrange-heavy. Omnis sounded absolutely dreadful in there. Since most cards have quite scooped response at 180 degrees, I used an AB cardioid pair aimed straight ahead, so the room was picked up almost entirely by the back of the mics, and the scooped sound neatly compensated for the mid-heavy room.

Well designed pencil condensers are superior at 180° rejection and flatness compared to most LDCs. SM7 is also great at rear rejection. Simply due to the fact that the mass of the body is placed behind the capsule. As long the rear chamber is well designed and it doesn't mess with the delay network the body acts somewhat as a baffle.

Regarding my previous post in replyto Dylan, it doesn't change the fact HC will still have more leakage at exactly 180° compared to cardioid under the same conditions.

There are ways to carefully modify existing LDCs to achieve the same effect. Neumann's KK105 and KK104 design gives some ideas.

Even tho i haven't had direct experience with Samar's TF08, I have strong feeling that capsule explores this idea.

 

[k brown]

Here are all the SDC off-axis plots I have in my 'library' (except the RE15, which I guess is an SDD?).

It's interesting that the Crown CM700 and the Primo EM21 off-axis plots are so different, since the EM21 is the capsule in the Crown (effect of the body/vents, I guess?).

Very odd that Primo shows very different off-axis plots for the EM200 and EM204, since all indications are that they the exact same capsule, except for the 204 having it's FET's source grounded internally.

 

[bremusound]

Sony C-35p / C-36p
Same capsule, but the 36 is side-address.

 

[bremusound]

Neumann KM54 (from the 1956 BBC R&D paper)

 

[bremusound]

Hiller M59 with and without „Schallumwegscheibe“ baffle

 

[k brown]

Hiller M59 with and without „Schallumwegscheibe“ baffle

Wow, that baffle ring sure messed up the off-axis response.

Interesting that the Schoeps V4U, which has a similar ring, doesn't show such wacked-out off-axis plots.

 

[ricardo]

The capsule type is self explaining thru 90° response: AT2031, KM184 and EM21 are cardioids (approx. -6dB) , TC600 is hypercardioid (approx. -8,5dB).

The dB-values in the legend are for 1 kHz. Mesasurement distance was 1m, data were taken in a church (2500m3 volume), quasi anaechonic, by windowing impulse responses before first echo arrived.

For directional mikes, sub-cardioids to Fig 8s, you MUST specify the measurement distance. Otherwise the cardioid 180 response is ambiguous because of proximity.

If the distance is 1m, you can tell a lot about the capsule by looking at the 180 response. eg

A perfect cardioid will not show -zillion dB rejection but -27dB @ 1kHz and worsening at 6dB/8ve as frequency goes down. Good cardioids will maintain -15 or -20dB rejection up to 20kHz. This gives the 'scooped' response measured at 1m

Very odd that Primo shows very different off-axis plots for the EM200 and EM204, since all indications are that they the exact same capsule, except for the 204 having it's FET's source grounded internally.

Actually the plots show 2 different samples of the 'same' capsule. It's VERY difficult to match the 180 response of cardioids.

Even a 'matched' pair of Schoeps CM4 are likely to show the differences in your EM200 & 204 graphs. Both are slightly 'sub', the EM204 being slightly more 'sub' than the EM200 sample but Dip. Ing Wuttke would consider a pair of CM4s with these differences 'matched'.

If i understand you correctly, you used each mic 0° as a reference for it's off axis response? You divided off axis responses by the 0°, or used 0° as a cal file.

I prefer to plot on-axis as well as off-axis without de-referencing with on-axis. Of course, all the measurements are de-referenced with my reference mike

There are things you can see very obviously if you keep all the measurements 'raw'. eg often there are peaks or dips in the on-axis response which give rise to corresponding dips & peaks in the off-axis responses. These are design or manufacturing faults which need to be addressed.

I have some mike measurements in MicBuilders Files under Mic Measurements

 

[k brown]

Forgot to include this graph which might be of interest; the Schoeps M221b (w/M934b capsule), which Marc Aubort used for his entire career (without the discs).