"Where does the tone come from in a microphone?"

[GeorgeToledo]

Bernhard Weingartner designed CK12, C12, C24 to have flat response especially in Fig8.

Konrad Wolf.

 

[Marik]

In da 70s & 80s, doing Hilbert Transforms was much more difficult than today. In 20+ yrs designing and measuring speakers, I've found only 2 drive units which were non-MP.

So how do you measure and what do you look for to determine if the drive unit is MP, or non MP? Is the sheer fact "it doesn't have a crossover" is enough to say--this is MP?

Best, M

 

[kingkorg]

Does it even make sense to talk about minimum phase if we consider room acoustics, all the reflections, and angle of incidence?

 

[kingkorg]

Bernhard Weingartner designed CK12, C12, C24 to have flat response especially in Fig8.

Thank you for bringing this up, I keep forgetting. The way true CK12 behaves in multipattern is way different to k47, k87, edge terminated k87 based capsules.

 

[ricardo]

So how do you measure and what do you look for to determine if the drive unit is MP, or non MP? Is the sheer fact "it doesn't have a crossover" is enough to say--this is MP?

• measure the Amplitude Response
• do a Hilbert Transform on this
• compare it with the measured phase You have to remove the phase shift due to any delays.
• If the same, the 'system' is MP

There's quite a number of Hilbert Transform pairs (like FFTs & IFFTs) depending on whether its Log response, Re & Im etc.

There's a number of visual clues we used in da old days when doing Hilbert Transforms was difficult. See various obfuscating papers by Richard Heyser

Today, you can do it with one line of MATLAB. See my two AES papers for details. Dis beach bum is sorry he can't provide copies cos several HD crashes.

One of the 'necessary but not sufficient' requirements for non-MP is that there must be at least 2 different paths. The two drive units I mentioned were

• Celestion HF1300 treble unit cos its 'phase correcting plate' in front of the diaphragm and
• Wharfedale Super 8RSDD, a famous double cone 8" unit. This would have 'no crossover'. It is likely that other double cone units would also be non-MP.

 

[GeorgeToledo]

I think that the focus on minimum phase is ignoring important non linear parameters.

For one, an amplification circuit of a microphone is not necessarily amplitude linear. It’s overwhelmingly likely not to be.

 

[abbey road d enfer]

I think that the focus on minimum phase is ignoring important non linear parameters.

Not "focus". The point has been raised by some members that phase response is a one-to-one derivative of the frequency response, which is true only in MP systems. Nobody suggested it was a cause for "bad sound".

For one, an amplification circuit of a microphone is not necessarily amplitude linear. It’s overwhelmingly likely not to be.

AFAIK, non linearities don't produce phase aberrations; I may be wrong...

 

[GeorgeToledo]

Not "focus". The point has been raised by some members that phase response is a one-to-one derivative of the frequency response, which is true only in MP systems. Nobody suggested it was a cause for "bad sound".

AFAIK, non linearities don't produce phase aberrations; I may be wrong...

I’m not sure where your quote of “bad sound” comes from.

What I am saying is that I don’t think a mic can be fully characterized by impulse response because it isn’t a minimum phase device. I think there was an earlier assertion in the thread that all mics were minimum phase. I question that because part of the definition of minimum phase is that a device has to be linear. Many amplification circuits in mics are not linear, setting aside linearity of the transducer. Which probably isn’t either, since capsules, ribbons, dynamic cartridges exhibit harmonic distortion.

This isn’t really directed at your comments though, I think mainly Ricardo(?). I think it is hard to debate that the amplification stages in many (all?) microphones are not linear.

I don’t really want to keep feeding the fire of the youtuber who got this all rolling though. Maybe some other day and in some other context.

 

[ricardo]

I think there was an earlier assertion in the thread that all mics were minimum phase. I question that because part of the definition of minimum phase is that a device has to be linear.

The obvious way to test this assertion is to measure loadsa mikes for Minimum Phase. I've probably done this more than anyone on this forum though I've done far more speakers.

I don't think the 'linearity' of mikes is enough to affect this. Mikes are certainly a LOT more linear than speakers.

I've designed mikes which might be non-MP cos multiple paths (the Soundfields and a couple of other Tetrahedral mikes) and I was pleased to find that they were MP ie their phase was as God intended

I might yet persuade Marik to start checking for MP so he might come up with evidence to the contrary I've provided some hints on where to look

As to whether it is necessary for good sound, the "Is Linear Phase Worthwhile?" AES paper shows large amounts of non-MP are mostly inaudible though I can certainly dream up test signals that are.

IIRC, there's an AES paper by Vanderkooy & Lipshitz, using their ABX box that shows there ARE real life signals where 'Phase Distortion' can be detected but weren't objectionable. It's just after my AES paper, the result of some discussions I had with them at Hamburg

 

[metalockpick]

Because the wavelength gets shorter as frequency climbs (12kHz is about 25mm), a larger diaphragm than the wavelength will yield an extremely directional speaker at high frequencies.
In a microphone, sound can go through the holes in the capsule and back to cause phase cancellation. It can bounce off the windscreen and cause phase cancellation. It can work its way around to the back diaphragm and cause phase cancellation. And I suppose if the diaphragm were too big and the frequency high enough, you could get cancellation from the same wave hitting two different places on the same diaphragm.
For these reasons, I don't think you can equate the transient response with the frequency response in a microphone.
Additionally I think you can only call it "minimum phase" when measuring at a particular frequency, at a single point in space, and in anechoic conditions. To claim otherwise is a gross oversimplification. Good test equipment exists, and is not supposed to be used to make particular units to have "favorable" readings. Additionally, The idea that a "dual cone" loudspeaker is "minimum phase" is complete rubbish.

 

[LevinGuitar]

Which is also another brand name. the actual owner is Samsung.
I'm not privvy to Samsung's work organization, but I would think all mics are now subcontracted to different jobbers. The 414 capsules are probably an OEM derivative of an existing line.

From the ashes, the Phoenix is reborn under the Austrian Audio name.

Seems that the new AKG mics (and capsules) are made in Hungary

https://www.facebook.com/AKG

 

[kingkorg]

Seems that the new AKG mics (and capsules) are made in Hungary

https://www.facebook.com/AKG

Awesome! Quite a lot to be learned from that video!

 

[ricardo]

For these reasons, I don't think you can equate the transient response with the frequency response in a microphone.

Er.rrh! It's called the Fourier Transform. Works with non-MP stuff too Look it up.

Additionally I think you can only call it "minimum phase" when measuring at a particular frequency, at a single point in space, and in anechoic conditions.

No. 'Minimum Phase' is a wideband property. It has precise mathematical meaning. Why don't you look it up. IIRC, one of my AES papers has a simple explanation.

Additionally, The idea that a "dual cone" loudspeaker is "minimum phase" is complete rubbish.

You might want to check what I actually said.

 

[Co.]

I get the feeling, that very early in this thread, someone mixed up the terms "transient response" as in "how (well) does a mic respond to transients from musical instruments" and the term "impulse response", a strictly technical term.
A lot technical terms seem to be getting mixed up and taken out of context.

 

[abbey road d enfer]

I get the feeling, that very early in this thread, someone mixed up the terms "transient response" as in "how (well) does a mic respond to transients from musical instruments" and the term "impulse response",

Do you think these are uncorrelated?

 

[Co.]

Do you think these are uncorrelated?

They are certainly related. But not in a straight forward 1:1 way.
It is totally possible, to have microphones with the same frequency response, but a completely different transient response. This of course is going to lead to a different representation as an IR.
So transient response and frequency response might be represented as an IR, but that doesn't mean, that transient response and frequency response is completely related.
How is this all going to be represented (in an IR) considering proximity effect and off axis response? What is actually the practical value of talk about IR and minimum phase, related to the topic of the thread, the mic test video and its questionable testing methods?

 

[metalockpick]

Er.rrh! It's called the Fourier Transform. Works with non-MP stuff too Look it up.


No. 'Minimum Phase' is a wideband property. It has precise mathematical meaning. Why don't you look it up. IIRC, one of my AES papers has a simple explanation.


You might want to check what I actually said.

The arrival of good PC based FFT in the late 1980s showed that some devices didn't actually have good frequency response, but instead generated noise over wide bands. For example, 1970s JBL supertweters tweeters that had the diamond embossed metal surrounds would create a lot of 5K band noise on their own, more related to the amplitude of input then the frequency. Paper "whizzer cone" or "dual cone" devices do this, which is why even though they are cheap to manufacture, they are not popular today.
When you put a pulse in a device, FFT will measure the behavior over time. That is a reasonable representation that correlates to the performance of a transducer. But what FFT really showed was that the previous generations of test equipment were mostly crap.
Impulse response...yes. I goofed on that.

 

[kingkorg]

It is totally possible, to have microphones with the same frequency response, but a completely different transient response.

This is simply not true, you are breaking laws of physics here. Can you provide any documentation to substantiate this? Or measurement? If two microphones have same exact frequency response, and their upper and lower limit of FR extends to the same point they will have exactly same TR. It seems you are mixing up TR and IR.

However if you have two mics that have same exact FR say 20-20.000hz but FR starts to differ above 20.000hz you do indeed get different TR.

I don't get why make things so complicated. Transient response is determined by the frequency response extension, or bandwidth. You can not limit FR and have good TR, same as you can't have wide FR and claim TR is bad just because the mic sounds dark.

I am of course talking about operation in linear region, and presuming microphones are not clipping.

 

[abbey road d enfer]

They are certainly related. But not in a straight forward 1:1 way.
It is totally possible, to have microphones with the same frequency response, but a completely different transient response. This of course is going to lead to a different representation as an IR.
So transient response and frequency response might be represented as an IR, but that doesn't mean, that transient response and frequency response is completely related.

So you disagree head-to-head with ricardo. You'd better have good arguments, then.

What is actually the practical value of talk about IR and minimum phase, related to the topic of the thread, the mic test video and its questionable testing methods?

Had you read the whole thread, you would know.

 

[abbey road d enfer]

If two microphones have same exact frequency response, and their upper and lower limit of FR extends to the same point they will have exactly same TR. It seems you are mixing up TR and IR.

This assumes that the system is minimum-phase. Actually are they MP? ricardo affirms they are. However, in the mathematical sense, MP implies an infinite frequency domain. Many writers present microphones as being "minimum-phase in the audio bandwidth", which leaves the possibility that they are not outside the specified useful range.
It means that micrs can be considered close-enough to MP, in a restricted BW.
That's the reason why room deficiencies, which are intrinsically non-MP, because they are due to multiple reflections, can however be partially EQ'd with MP filters.

However if you have two mics that have same exact FR say 20-20.000hz but FR starts to differ above 20.000hz you do indeed get different TR.

That's a very important point, actually the difference due to LF response is even more notable on IR.

I don't get why make things so complicated. Transient response is determined by the frequency response extension, or bandwidth. You can not limit FR and have good TR, same as you can't have wide FR and claim TR is bad just because the mic sounds dark.

Define "good Transient Response".