ultra-thin diaphram, fast, small diaphrgam condensers?

[sunflute]

Originally Posted by jnorman

wow marco - look what you started...

about all i can determine from this extended discussion is, no matter what you think about transient reponse, slew rate, and frequency response, you are WRONG!!

it seems to me that any decent condenser mic should easily be able to handle the transient peaks from a flute. percussive instruments, like a snare drum, is where you really start to hear the differences in mics in the way they respond to sharp transient signal changes, and in those cases, it seems evident that mics with smaller, lighter diaphragm materials can more quickly react to transient changes, though that capsule must be backed up by electronics that can also handle quick recovery and level/output stabilization.

With all due respect jnorman, not all condenser microphones are equipped to handle the complex transient character of the flute in the high register (which like yourself, we both agree it is one of the most difficult instruments to record).

For example, just about everyone here likes Schoeps but IMHO this is one microphone that is not flattering for the flute.

Actually, I have found this thread quite enlightening in that it has pointed me in the direction of microphones I have not tried yet, like the Josephson 617 or the Neumann TLM50.

Another thing about this thread that is very enlightening (and confirms my previous intuition about mics for the flute), is the discussion about the difference between impulse response vs. frequency response as mentioned earlier. i.e. Ribbon microphones having excellent impulse response vs. poor frequency response.

I don't think it is fair to call us WRONG for wanting to know more.

Peace
Marco

[sunflute]

Originally Posted by audio ergo sum

Please read this document:
http://www.neumann.com/download.php?...d=lect0036.PDF
and explain why the ribbon has the best transient response graph, but the worst FR graph...

Again, Thank you for another wonderful article.

[sunflute]

Originally Posted by Audiop

It's been explained in this thread several times. Flat extended response with no sharp roll offs at either end.
/Peter

Thanks Peter. I will look into Earthworks mics as well. Peace.

[Audiop]

Originally Posted by audio ergo sum

Again, I think theory does not translate 100% to the real world systems and particularly transients. The membrane deviations that are relevant for low level signals in condenser mics are close to molecular dimensions. There is a certain amount of nonlinearity and chaos when the membrane picks up a single impuls/wavefront, not a periodic stationary signal.
The deviations from the theoretical models are proven to be significant enough to mean sonic differences to our very sensitive ears.

Such behaviour is mapped with distortion measurements since any non-linear compliance of the capsule translates to harmonic distortion and intermodulation distortion.

Another factor is psychoacoustics. The ear does not detect phase shifts in periodic signals above 3-5 kHz.

Let's say it's not very sensitive above 1-2kHz for any signal but if the amount of phase shift is big enough it will likely be audible.

But it does detect phase shifts/smearing in aperiodic transients !!!

To some degree yes, depending on the signal and amount of phase shift.

Doesn't the outer world dictate the physics actually:-)

I told you so, didn't I? ;-)


/Peter

[Audiop]

Originally Posted by csbarnet

Peter, you have not produced a single document of any kind to back up your assertion that microphones can be modeled accurately as a completely minimum phase system. All you have done is make appeals to 'known physics'.

By your rationale, I could just as easily call your postings a 'joke'.

If you want to really talk science, please provide sources.

Okay. Can we all agree that a loudpseaker driver is a min-phase device?

Can we all agree that speakers and microphones are the very same thing?

A dynamic microphone has a membrane of some sort with a voice coil attached. Voice coil is in the gap of a magnetic motor. A dynamic speaker driver is exactly the same thing.

The only difference is dimensions and that when used as a speaker we put in electrical energy at one side and get acoustic energy at the other while the opposite is true for a mic.

Now tell me how one can be min-phase but the other not??


Condensers you say? They are pretty much the equivalence of ESL's.

Also, if microphones would react different to static signals and transient signals and somehow be something else than minimum-phase, how is it that you can measure various kinds of stimul input to a speaker and get the exact same reading after calculation?

If the mic would be some mysterious gadget that some people like to believe, speaker measurements would be very different.

Now, give me anything that points to speakers or mic's not being min-phase. And please not another AKG mumbling without any substance what so ever. ;-)


/Peter

[didier.brest]

Originally Posted by Audiop

Such behaviour is mapped with distortion measurements since any non-linear compliance of the capsule translates to harmonic distortion and intermodulation distortion.

Generally such measurement is not available for a microphone. The manufacturers specify only the maximum SPL for a given harmonic distortion amount. And there is no evidence (at least to me) that such measurement for non linear behaviour of a microphone against stationary sine signals would be useful for evaluating its behaviour against transients.

[jnorman]

marco - of course you are not wrong to want more information - that is always a good thing. i was simply making a joke about peter and didier (and others) going back and forth so much with the comments like:
"That article is a joke."
"By your rationale, I could just as easily call your postings a 'joke'."
"But sure, if you feel the science community is wrong about this, feel free to rewrite history and meet me for drinks at the Nobel dinner. ;-)"
"That is wrong. It is much more complex."
"Wow. When inaccurate, incomplete, and over generalized information is put together in a seemingly rational way, statements like these are often the result."
"That is not exactly true. "
"No it doesn't:"
"Wrong."
"How wrong you are."
"To say something positive, you sure don't let facts confuse you... "
etc, etc

anyway, i doubt that lack of transient response is the reason you were not pleased with the schoeps. i also was less than happy with the flute sound i got out of my CMC64s. i rather liked the DPA 4011s in ORTF about 7-feet out, though.

carry on...

[sunflute]

Thanks jnorman,

But in the midst of the crazy back-and-forth some wonderful information has been distilled.

I am particularly grateful for the two articles posted by audio ergo sum from Georg Peuss (Neumann). Very enlightening.

Like you said.....carry on.

Anyways, like I said in my original post....I Really like the KSM141, this mic professes to have a 2.5 micron diaphragm as well as a fast, transformerless preamp body. It almost feels ribbon like but with a better frequency response. I think it will be a nice match-up for my crowley & tripp mics. Peace.

Originally Posted by jnorman

marco - of course you are not wrong to want more information - that is always a good thing. i was simply making a joke about peter and didier (and others) going back and forth so much with the comments like:
"That article is a joke."
"By your rationale, I could just as easily call your postings a 'joke'."
"But sure, if you feel the science community is wrong about this, feel free to rewrite history and meet me for drinks at the Nobel dinner. ;-)"
"That is wrong. It is much more complex."
"Wow. When inaccurate, incomplete, and over generalized information is put together in a seemingly rational way, statements like these are often the result."
"That is not exactly true. "
"No it doesn't:"
"Wrong."
"How wrong you are."
"To say something positive, you sure don't let facts confuse you... "
etc, etc

anyway, i doubt that lack of transient response is the reason you were not pleased with the schoeps. i also was less than happy with the flute sound i got out of my CMC64s. i rather liked the DPA 4011s in ORTF about 7-feet out, though.

carry on...

[Audiop]

Originally Posted by didier.brest

Generally such measurement is not available for a microphone. The manufacturers specify only the maximum SPL for a given harmonic distortion amount.

Not my intention to be argumentative just because but what manufacturers publish or not is of little interest to me. I care more about physics, science and methods and often do my own measurements which is educational and also make more sense since you then can compare various gear and be sure that the test set up is identical hence not interfering with the results.

And there is no evidence (at least to me) that such measurement for non linear behaviour of a microphone against stationary sine signals would be useful for evaluating its behaviour against transients.

Then we use transient signals and not stationary signals as stimul. :-)



/Peter

[Audiop]

Originally Posted by audio ergo sum

The membrane deviations that are relevant for low level signals in condenser mics are close to molecular dimensions.

Yup, very small indeed.

There is a certain amount of nonlinearity and chaos when the membrane picks up a single impuls/wavefront, not a periodic stationary signal.

What do you base this on?

The membrane has some tension from the mounting. An acoustical wave of any sort pushing on the membrane increases the tension and distortion from nonlinear stretching of the membrane should increase with amplitude.

The deviations from the theoretical models are proven to be significant enough to mean sonic differences to our very sensitive ears.

Which models and where can I see proof of it?


/Peter

[d_fu]

Originally Posted by audio ergo sum

? to what is it related then? I did not see that part.

I was referring to page 3 - which I believe is mostly about the original M50 (and its particular construction), but some of that will also apply to the TLM.

[didier.brest]

Originally Posted by Audiop

Not my intention to be argumentative just because but what manufacturers publish or not is of little interest to me.I care more about physics, science and methods and often do my own measurements which is educational and also make more sense since you then can compare various gear and be sure that the test set up is identical hence not interfering with the results.

Your measurements might be of interest to the members of this forum if you would present them. Even if not stated so clearly, the point at stake at the beginning of our discussion was whether one could have a good idea about the transient response of a microphone from its specs, especially its frequency response, as available generally from the manufacturers so that an user could choice a microphone rather another one if this criterion is determining. I think that the answer to this question, which may be of higher interest to most readers of this topic than the maths and the physics, is rather negative.

[sunflute]

Originally Posted by didier.brest

Your measurements might be of interest to the members of this forum if you would present them. Even if not stated so clearly, the point at stake at the beginning of our discussion was whether one could have a good idea about the transient response of a microphone from its specs, especially its frequency response, as available generally from the manufacturers so that an user could choice a microphone rather another one if this criterion is determining. I think that the answer to this question, which may be of higher interest to most readers of this topic than the maths and the physics, is rather negative.

Agreed

[Audiop]

Originally Posted by didier.brest

Even if not stated so clearly, the point at stake at the beginning of our discussion was whether one could have a good idea about the transient response of a microphone from its specs, especially its frequency response, as available generally from the manufacturers so that an user could choice a microphone rather another one if this criterion is determining. I think that the answer to this question, which may be of higher interest to most readers of this topic than the maths and the physics, is rather negative.

And I'm confident the answer is yes.


/Peter

[BrianHanke]

Originally Posted by jnorman

i was simply making a joke about peter and didier (and others) going back and forth so much with the comments like:
"That article is a joke."
"By your rationale, I could just as easily call your postings a 'joke'."
"But sure, if you feel the science community is wrong about this, feel free to rewrite history and meet me for drinks at the Nobel dinner. ;-)"
"That is wrong. It is much more complex."
"Wow. When inaccurate, incomplete, and over generalized information is put together in a seemingly rational way, statements like these are often the result."
"That is not exactly true. "
"No it doesn't:"
"Wrong."
"How wrong you are."
"To say something positive, you sure don't let facts confuse you... "
etc, etc

James, I was going to post almost exactly the same thing! Great minds think alike...

[BrianHanke]

Just a thought regarding the main topic of this thread: I think if there was a standardized way to publish microphone specifications it would clear up a lot of our confusion. As it stands now, every manufacturer has a slightly different way of graphing the frequency response of their mics and so we are left wondering exactly how different models compare.

A while ago I posted a comparison between a pair of Neumann KM 131s and a pair of Lauten Torches (with the omni capsules). We could all agree that the Lautens sounded brighter. After using them a few days I found they worked well in some kinds of music and not so well in others. Bach sounded better with the Neumanns but Liszt sounded better with the Lautens. Go figure...

On paper both mics look like they have an extremely flat frequency response over their entire range. However, if we dig a little deeper we notice that the Neumann graph states there's a tolerance of +/- 2 dB. That's a lot of potential variance. If there was a gentle rise of 2 dB in the mid range followed by a gentle dip of 2 dB in the treble that could well account for the darker sound of the mics. The graph on the Lauten website is less smoothed, but it doesn't say there's no smoothing at all.

If all microphone manufacturers made extremely detailed graphs (perhaps averaging the results from a random sampling of a large number of mics), with no smoothing and accuracy to within a tenth of a dB, we'd be able to make more informed decisions based on the published specs.

[Audiop]

Good point Brian.

But please don't hold your breath. :-)

What we can do is collect knowledge and outline measurement procedures as a community.

We cold specify a test set up which could be used all over the globe and those results could be truly comparable.

As you say, different measurement methods, smoothing, scaling of graphs and so on makes it difficult to compare manufacturers graphs.


/Peter

[BrianHanke]

Originally Posted by Audiop

But please don't hold your breath. :-)

No doubt this information will never be volunteered by the manufacturers. Independent test results would be a highly valuable resource though. Heck, I'd even be willing to pay to view such data.

[John Willett]

Originally Posted by BrianHanke

Just a thought regarding the main topic of this thread: I think if there was a standardized way to publish microphone specifications it would clear up a lot of our confusion. As it stands now, every manufacturer has a slightly different way of graphing the frequency response of their mics and so we are left wondering exactly how different models compare.

I certainly think this is a good idea.

I know there was talk a few years ago about this and various manufacturers discussing this - I do hope it happens as it would be best for all.

[David Spearritt]

Originally Posted by sunflute

With all due respect jnorman, not all condenser microphones are equipped to handle the complex transient character of the flute in the high register (which like yourself, we both agree it is one of the most difficult instruments to record).

Why is it complex? What frequencies are we talking here? Is this just your feeling? From actual measurements, there is not a lot of energy above 4 kHz.
modern B flute acoustics - multiphonic D#5 B5 D6 G6 A#6

I find the flute difficult to record with any spaced mic techniques because of the awful and very noticeable comb filtering. But when one records flute with coincident techniques, it is no more complex or difficult to get really hifi than any other instrument, IMHO.

Another thing about this thread that is very enlightening (and confirms my previous intuition about mics for the flute), is the discussion about the difference between impulse response vs. frequency response as mentioned earlier. i.e. Ribbon microphones having excellent impulse response vs. poor frequency response.

This is unsubstantiated and I suspect it is incorrect. You cannot have a "good" impulse response without an extended freq response.

References to these assertions would help.

[audio ergo sum]

Originally Posted by Audiop

...

What we can do is collect knowledge and outline measurement procedures as a community.

We cold specify a test set up which could be used all over the globe and those results could be truly comparable.

...

I love the idea, but am afraid that even a standardized measurement procedure yields non comparable results, if done in different locations with different equipment etc., considering the sensitive nature of sonic differences.

It would take someone with the expertise and facilities and being independent from the industry to measure all kind of microphones in a single facility.

What would everybody like to see in such a data set beyond manufacturers published data?

Out of my head some of my wishes:

Frequency response:
(all with max. 1/3rd octave smoothing or 1/12th octave smoothing?)

-0 deg 1m distance (ISO standard)

-0 deg 0.3m distance (proximity effect)

-off axis 1m distance in 45 deg steps (45, 90, 135, 180)

Polar graphs full spectrum in octave steps (63Hz-16kHz)

Transient response:

-repeatable standardized test setup (ultrasonic spark discharge)


what else?

[audio ergo sum]

Originally Posted by David Spearritt

...
This is unsubstantiated and I suspect it is incorrect. You cannot have a "good" impulse response without an extended freq response.

References to these assertions would help.

See the linked article from Stephan Peus (Neumann) above.
They measured a ribbon where the transient suggested an f-max of 50kHz while the measured frequency response of the capsule rolled off (more gently though than a condenser but earlier) beginning at 10 kHz !

[David Spearritt]

The Peus PDF seems a very poor translation and paper that cannot be taken seriously. It presents no conclusions. The figures are all out of order, with different scales in both X and Y axis on each scope. The ribbon impulse is a poor representation of the spark recorded with the measurement mic, itself an influence on the result, being as it probably is, a condenser.

Its drawing a long bow to conclude that this paper concludes that a ribbon has a "good" impulse response.

Where is a decent and perhaps "recent" paper to disprove the laws of physics.

[audio ergo sum]

Originally Posted by David Spearritt

The Peus PDF seems a very poor translation and paper that cannot be taken seriously. It presents no conclusions. The figures are all out of order, with different scales in both X and Y axis on each scope. The ribbon impulse is a poor representation of the spark recorded with the measurement mic, itself an influence on the result, being as it probably is, a condenser.

Its drawing a long bow to conclude that this paper concludes that a ribbon has a "good" impulse response.

Where is a decent and perhaps "recent" paper to disprove the laws of physics.

??? I find the paper rather clear. The X-Axis is always 100 microsec, stated clearly. Only the reference impulse has a 20 microsecond scale, stated clearly. Everything is in order and clearly labeled.
The Y-Scale is different, relative to the microphones sensitivity, which makes the most sense. The transient response quality is represented in the Y-plane by the relation between max deflection and post ringing, so absolute scaling doesn't make sense.

The paper does not disprove laws of physics.

The point is not if the ribbon has a good transient response, but that in real world with mechanical components there is no 100% correlation between theoretical models and practical behavior.

The problem here is that you want the real world to comply 100% with simplified theoretical models.

Originally Posted by David Spearritt

...
I find the flute difficult to record with any spaced mic techniques because of the awful and very noticeable comb filtering...

??? Maybe OT, but there is no comb filtering because of spaced mic techniques, not when the mic signals are panned hard L/R.

[d_fu]

Originally Posted by BrianHanke

If all microphone manufacturers made extremely detailed graphs (perhaps averaging the results from a random sampling of a large number of mics), with no smoothing and accuracy to within a tenth of a dB, we'd be able to make more informed decisions based on the published specs.

Do you think such detailed graphs (where the averaging would also act as some sort of smoothing) would really help you choose one mic over another?

I don't mind a bit of smoothing, as long as a tendency can be seen (treble boost or whatever). I would find polar graphs and 90°/180° responses much more interesting.

Originally Posted by David Spearritt

Why is it complex? What frequencies are we talking here? Is this just your feeling? (...)
But when one records flute with coincident techniques, it is no more complex or difficult to get really hifi than any other instrument, IMHO.

David, you keep making good points here IMHO...

Originally Posted by audio ergo sum

??? Maybe OT, but there is no comb filtering because of spaced mic techniques, not when the mic signals are panned hard L/R.

Guess he was referring to phase issues.

BTW, I also find the Neumann pdf a bit inconclusive...


D.

[John Willett]

Originally Posted by audio ergo sum

I love the idea, but am afraid that even a standardized measurement procedure yields non comparable results, if done in different locations with different equipment etc., considering the sensitive nature of sonic differences.

It would take someone with the expertise and facilities and being independent from the industry to measure all kind of microphones in a single facility.

What would everybody like to see in such a data set beyond manufacturers published data?

Out of my head some of my wishes:

Frequency response:
(all with max. 1/3rd octave smoothing or 1/12th octave smoothing?)

-0 deg 1m distance (ISO standard)

-0 deg 0.3m distance (proximity effect)

-off axis 1m distance in 45 deg steps (45, 90, 135, 180)

Polar graphs full spectrum in octave steps (63Hz-16kHz)

Transient response:

-repeatable standardized test setup (ultrasonic spark discharge)


what else?

There are standardised frequencies for polar-patterns:- 125, 250, 500, 1kHz (normally on the left) & 2k, 4k, 8k and 16kHz (normally on the right) - microphones with an extended top end also often show the 32kHz response on the right. These are in the standard and, personally, I get wary when people quote other frequencies or leave some of these not shown.

I agree with your suggestion of 1m and 0.3m.

What else? - a phase response curve would be useful. I have only ever seen this response curve quoted in the Sennheiser MKH brochure for the 20/30/40 series - I don't know if any other manufacturer quotes this.

The only problem I see is the cost of doing all this by an independent - maybe the only way would be for a university to take it on; but it would only be fair if the manufacturers supplied samples direct.

Good idea, though.

[audio ergo sum]

Originally Posted by d_fu

...

BTW, I also find the Neumann pdf a bit inconclusive...

It is, they don't conclude. But they show by experiment a significant unexplained deviation from the simplified theoretical model.

Here a more conclusive graph :-)

[Audiop]

Originally Posted by audio ergo sum

See the linked article from Stephan Peus (Neumann) above.
They measured a ribbon where the transient suggested an f-max of 50kHz while the measured frequency response of the capsule rolled off (more gently though than a condenser but earlier) beginning at 10 kHz !

It's not so. I doubt the author really say that but if he do he lacks understanding of the subject.

The problem is to know what the different parts in a pulse response or square response means.

Things that effect the pulse response:
1-Bandwith
2-In band flatness
3-The order of roll off at the extremes

I made a swept sine and square wave measurement of a AD-DA loop some time ago and sent the square response to a friend (a doctor that teaches acoustic/speech/music/hearing at the royal institute of technology in Stockholm)
and by looking at the square and doing a quick calculation by hand he gave me the -3dB point within tens of a dB from the sine measured response.

Two-three articles have been posted in this thread and none of them supports what the posters think, it's rather a question of misunderstandings and false unsubstantiated claims from the articles (AKG first and foremost).

A very resonant system have a pulse response that oscillates around the null axis, a non resonant system does not. The rise time of the initial flank depends on bandwith and if this bandwith is extended high with a shallow roll off there will be practically no HF ringing. The opposite can be seen in high order filters for example a PCM converter loop. We indeed have flat response up to 20k but since a brick wall filter kick in at the nyquist frequency there will be ringin in the impulse or square response.

Likewise if we look at the low end of the spectrum. If the fundamental resonance of the system (Fs as in speaker drivers) has low Q there will be no overshoot (ringing) in the pulse response. A Q at 0.2-0.5 is close to perfect and has a relatively high -3dB point (shallow roll off as the ribbon). A Q of 0.7 has the flattest response (-3dB point as low as possible) but some overshoot in time domain. Going up to Q of 1 gives even more overshoot and now the frequency resonse is peaking as opposed to being flat for Q's of 0.7 and below.

Again, there is no surprise about the ribbon response and frequency response is tightly connected to time domain behaviour.


/Peter

[Audiop]

Originally Posted by audio ergo sum

It is, they don't conclude. But they show by experiment a deviation form the simplified theoretical model.

No, it's a misunderstanding. I've outlined why several times allready. Please se my last post above. :-)


/Peter

[John Willett]

Originally Posted by Audiop

It's not so. I doubt the author really say that but if he do he lacks understanding of the subject.

Stephan Peus is one of the top microphone designers and certainly *does* know his subject.

He is up there with the best, including Joerg Wuttke of Schoeps, Manfred Hibbing of Sennheiser and Norbert Sobel of AKG (yes, there are a few more, but I don't know them personally).