Frequency Response, Sensitivity, etc., Measuring Microphone

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

Zottel

Well-known member
Joined
Aug 15, 2015
Messages
73
Location
Leipzig
Hey folks!

My mic is working now  Thanks again for the people that helped me out!

I’m back with some questions about measuring a microphone. Sadly, Gefell (wo promised me to measure my mic and do a kind of shootout against an original U47) told me today that this measuring test won’t take place. They don’t have time for that.  :( So now I’m on my own. Horrifyingly I got to finish my bachelor thesis till the 30.9. and I don’t have a clue how to measure these parameters without the cool equipment Gefell got in their HQ.  :-\ So I got some ideas and I would love to get some input from you guys. Keep in mind that I don’t want to promote anything with these results. I just want to check how close my DIY mic is to the original.:

Frequency response
My first idea was to use a measuring microphoneand use it as some kind of reference, put the U47’s (original and mine) in a quiet room and use a loudspeaker and a signal generator. Increasing the frequency from low too high and writing down the values shown on a FFT-Plugin.
The curve might look like crap, because of the room and the frequency curve of the loudspeaker, but at least it’s possible to see how big the difference between these mic’s is.

Sensitivity
Same kind of setup but using a signal generator and the loudspeaker to generate a 1kHz sine wave with 94 dBa.(Using the measuring microphone to check the 94 dBa)
When 94 dBa is reached I will be measuring the Voltage at the XLR Pins of the U47’s power supply, that would get connected to the Mic preamp? Does this make any sense? It’s not allowed to open the original U47/power supply. So there is no way in getting in there somewhere.

Self-Noise
Rest of these things will be really hard to measure I guess. For self-noise I would have to build some kind of Rauschbombe right? But again… I can’t open up the original u47… so this won’t make sense anyway?

Klirr factor
Same setup like when I’m measuring the Frequency response. Using a FFT-Plugin to analyse. But not sure if this makes any sense at all. The room is a “kind of dead” recording room and I guess the loudspeaker will start to klirr on its own.

Intermodulation-factor & Differenzton-factor
Same problem as Klirr-factor.

What do you think… will these tests make any sense at all?
Do you guys have any idea how to do it? Without buying stuff worth 1000€.

Cheers
Zottel
 
Zottel said:
Frequency response
My first idea was to use a measuring microphoneand use it as some kind of reference, put the U47’s (original and mine) in a quiet room and use a loudspeaker and a signal generator. Increasing the frequency from low too high and writing down the values shown on a FFT-Plugin.
The curve might look like crap, because of the room and the frequency curve of the loudspeaker, but at least it’s possible to see how big the difference between these mic’s is.
Assuming the actual response of the measurement microphone is known, you would get a good impression of the frequency response of your microphone, although tainted with issues; the resposne of a directional microphone is different of that of a pressure mic, because the wave front is not perfectly planar, and due to the room, the response below 200Hz cannot be considered accurate.
Sensitivity
Same kind of setup but using a signal generator and the loudspeaker to generate a 1kHz sine wave with 94 dBa.(Using the measuring microphone to check the 94 dBa)
When 94 dBa is reached I will be measuring the Voltage at the XLR Pins of the U47’s power supply, that would get connected to the Mic preamp? Does this make any sense? It’s not allowed to open the original U47/power supply. So there is no way in getting in there somewhere.
It's typically done with a pistonphone, but that would not fit on a cardioid mic (it works only with pressure mics), so again you must use the comparison method, which implies you know the sensitivity of your reference mic. You need to measure the level at the output of the mic pre and leave the settings unchanged when swapping mics. The difference in measured voltage is equal to the difference in sensitivity. You must use a sweep around 1kHz in order to average the frequency response irregularities.
Self-Noise
Rest of these things will be really hard to measure I guess. For self-noise I would have to build some kind of Rauschbombe right? But again… I can’t open up the original u47… so this won’t make sense anyway?
This is very difficult because you need a room that has less noise than the mic's noise. I don't believe you can make a proper evaluation without a silent room/enclosure.
Klirr factor
Same setup like when I’m measuring the Frequency response. Using a FFT-Plugin to analyse. But not sure if this makes any sense at all. The room is a “kind of dead” recording room and I guess the loudspeaker will start to klirr on its own.

Intermodulation-factor & Differenzton-factor
Same problem as Klirr-factor.
You're right; your speaker is likely to distort much more than your mic.
What do you think… will these tests make any sense at all?
Do you guys have any idea how to do it? Without buying stuff worth 1000€.
Ask Microtech Gefell if they want to help you with your thesis; they may be interested in helping a young passionate  that they may also want to hire later...

 
  I've seen some work from a Univ in Chile IIRC, long time ago, couldn't possible point to any help right now, I don't remember if I read the paper or saw a presentation from the author.

  I do remember they were measuring electroacustical properties with limited resources, could be the speaker or the mic, the thing is they where ruling out the room as a variable using some windowing technique, which will be limited in LF by the size of the room but is very helpful from there up. I don't remember the details, just saying there's something out there that may help you with this. Probably it was some repeated short measurements to do some clever processing in matlab and getting away with it having a smaller margin error independently from the environment.

  The alternative is to use a natural anechoic chamber, go to an open field, away from any building, and if possible road. and mesure there. I don't know how hard is to find something like this near to you. The only reflection you need to worry about is the floor, maybe a barrier blocking the direct reflection will do for the HF, placing the source close to it would do it for LF.

JS
 
Indeed, windowing is the key to better, almost anechoic acoustic measurements.
The CLIO (Audiomatica) manual is the source for all relevant information.
Ther are three free software that implement windowing: REW (Room EQ Wizard - although primarily intended for home-cinema tuning is quite apt at mic measurement), ARTA (http://www.artalabs.hr/) and AURORA (http://www.aurora-plugins.com/)
The latter is the most powerful, but not the easiest, since it's a set of plug-ins for Adobe Audition, not an integrated package.
Actually, none of them is primarily designed for mic testing, but they're all suitable, and all have the same limitations due to the size of the working space (unreliable measurements under 200 Hz). Note that windowing is "not an exact science"; one has to fully realize the compromises it implies.
 
First of all - please do not expect measurement data to reveal anything significant about how the microphone sounds. We are still far from being able to quantify most important parts of what makes subjective quality.

Second - why not ask at your local technical university (HTWK Leipzig?) or a similar institution - they often have setups that they love to use for something real..

Last - when doing the measurement, make sure to get plots of sensitivity vs. directional patterns at different frequencies - this is probably the parameter that tells you most about the capsule mechanics

Jakob E.
 
abbey road d enfer said:
The CLIO (Audiomatica) manual is the source for all relevant information.

In another forum  I write ...
Today, frequency responses are best obtained by a quasi-anechoic method which lets you do ‘anechoic’ measurements in a domestic room.  I would want to do this even in a big anechoic chamber.

Chapter 10.4.3 ACOUSTIC FREQUENCY RESPONSE

of the Clio 8 manual is the best explanation of this. http://www.audiomatica.com/wp/?page_id=34

I highly recommend Clio as an inexpensive R&D tool.  The calibration of their inexpensive measurement mikes are one of the few I believe.  ClioQC is THE system for production testing.

Eric Benjamin’s Extending Quasi-Anechoic Measurements to Low Frequencies
http://www.aes.org/e-lib/browse.cfm?elib=12875
explains some advanced methods to extend useful measurements to lower frequencies.

These methods measure an impulse response and use FFTs to get the (complex) frequency response.  The comparison is done by (complex) dividing the response of the DUT by the response of the reference in the Frequency Domain.

There’s only a few free programs that allow this to be done conveniently.

 Prof. Angelo Farina developed the Log Sweep method that is now the favoured method from Clio to the latest Audio Precision to yours truly.  If you have Audition, try his Aurora plugins.  There is a version for a special version of Audacity.
 Our own Les Watts uses & recommends the full version of ARTA.  There is a free version.

Read the Clio manual regardless of what you use.  I haven’t used any of these methods but have my own version of Angelo’s method in DOS.
The Clio manual is the ONLY good accurate explanation of quasi-anechoic.  If you come across any more, please let me know.

Alas, I'm told that Audiomatica no longer have that section in their Clio 11 manual .. but it should be in the older manuals.
 
The Earthworks paper:

"How Earthworks Measures Microphones"
'How to measure microphones and the implications relating to measuring loudspeakers.'

http://www.earthworksaudio.com/support/technical-articles/
http://www.earthworksaudio.com/wp-content/uploads/2012/07/how-earthworks-measures-mics.pdf
 
Thanks for the reply!  :)
abbey road d enfer: Gefell told me in June that they would be helping me with all this stuff. But sadly they have a lot of work to do right now. That’s why they had to cancel.  :(

I asked all of the technical university’s around. Non could/would help me. But I came across a company named Mikrofontechnik Leipzig. They told me that they would be willing to help me out on this although they are not sure if they have the tools to meet all my “needs.” Going to call them on Monday. Let’s see what they have to offer.  :)

Hopefully they can help me with the main things. Because I don’t get all that other stuff you guys talked about. Blows my mind completely!  :eek:

Do you guys know if Klirrfactor etc. is measured with or without the capsule?
 
Zottel said:
Do you guys know if Klirrfactor etc. is measured with or without the capsule?
Distortion (english for klirrfaktor) is measured with the capsule. Actually, distortion happens in the capsule AND in the electronics. The distorsion test recommended in the standard technical manual for Neumann is carried only on the electronics, because, although it is relatively easy to find a good audio generator with low distortion, it is much more difficult to find a loudspeaker capable of producung less distortion than the capsule. Most microphones are measured for 3% distortion, and that happens with very high spl levels (typically >130dB spl). At this level, most loudspeakers produce about 20-30% distortion!
 
What kind of technical manual are you talking about?
Hopefully the MTL guys are capable of measuring these things  :eek: Otherwise I'm kind of fucked :D
 
Zottel said:
I’m back with some questions about measuring a microphone.

The reference tool for measuring microphones is called a microphone calibrator.

http://www.ebay.com/itm/Reed-SC-05-Sound-Level-Calibrator-for-1-2-Diameter-Microphones-0-5dB-/331966649136?hash=item4d4ac11330:g:YM4AAOSwgmJX1Owi

They are basically a miniature single-frequency signal generator and loudspeaker. Their performance is highly predictable and stable.  This gives you a way to measure sensitivity at a convenient frequency in the middle of the audio range where most microphones are relatively flat. 

The usual way to extrapolate from a mic calibrator to general frequency response measurements is to use a measurement-style microphone. They are omnis and have fairly predictable frequency response that is very flat over a useful, predictable  range,partially because they are small diameter and omnis. There sensitivity is not so well controlled, and therefore we have the mic calibrator as our reference..

http://www.ebay.com/itm/Behringer-ECM8000-Professional-Measurement-Microphone-Mic-ECM-8000-689076111713-/182281444156?hash=item2a70d21b3c:g:p2EAAOSwsB9WDWiM

You can then use this mic to understand the performance of whatever sound source you are using to test your experimental mic.

 
arnyk said:
The reference tool for measuring microphones is called a microphone calibrator.

They are basically a miniature single-frequency signal generator and loudspeaker. Their performance is highly predictable and stable.  This gives you a way to measure sensitivity at a convenient frequency in the middle of the audio range where most microphones are relatively flat. 
Mic calibrators are almost useless when it comes to directional microphones (cardio or fig-8), not only because of the form factor, that makes proper acoustic coupling very awkward, but also because directional microphones are not pressure-sensitive, they are velocity-sensitive. A mic calibrator produces a specific sound pressure that does nor necessarily translates linearly into velocity. The most accurate sensitivity method for directional microphones is a comparative method.
 
abbey road d enfer said:
arnyk said:
The reference tool for measuring microphones is called a microphone calibrator.

They are basically a miniature single-frequency signal generator and loudspeaker. Their performance is highly predictable and stable.  This gives you a way to measure sensitivity at a convenient frequency in the middle of the audio range where most microphones are relatively flat. 
Mic calibrators are almost useless when it comes to directional microphones (cardio or fig-8), not only because of the form factor, that makes proper acoustic coupling very awkward, but also because directional microphones are not pressure-sensitive, they are velocity-sensitive. A mic calibrator produces a specific sound pressure that does nor necessarily translates linearly into velocity. The most accurate sensitivity method for directional microphones is a comparative method.
Isn't that what he was saying? Calibrate your omni pressure measurement mic with the mic calibrator, use that to measure your loudspeaker / chamber performance and then take that into account when measuring your mic under test.
 
Matt Nolan said:
abbey road d enfer said:
arnyk said:
The reference tool for measuring microphones is called a microphone calibrator.

They are basically a miniature single-frequency signal generator and loudspeaker. Their performance is highly predictable and stable.  This gives you a way to measure sensitivity at a convenient frequency in the middle of the audio range where most microphones are relatively flat. 
Mic calibrators are almost useless when it comes to directional microphones (cardio or fig-8), not only because of the form factor, that makes proper acoustic coupling very awkward, but also because directional microphones are not pressure-sensitive, they are velocity-sensitive. A mic calibrator produces a specific sound pressure that does nor necessarily translates linearly into velocity. The most accurate sensitivity method for directional microphones is a comparative method.
Isn't that what he was saying? Calibrate your omni pressure measurement mic with the mic calibrator, use that to measure your loudspeaker / chamber performance and then take that into account when measuring your mic under test.
My reading of the earlier post was that he suggested using the calibrator on the MUT. Language issue maybe...
 
abbey road d enfer said:
My reading of the earlier post was that he suggested using the calibrator on the MUT. Language issue maybe...

Perhaps: "The English and the Americans are two great peoples separated by a common language" ;-)
 
arnyk said:
abbey road d enfer said:
My reading of the earlier post was that he suggested using the calibrator on the MUT. Language issue maybe...

Perhaps: "The English and the Americans are two great peoples separated by a common language" ;-)
Made only worse by the fact I'm neither English nor American...
 
abbey road d enfer said:
arnyk said:
abbey road d enfer said:
My reading of the earlier post was that he suggested using the calibrator on the MUT. Language issue maybe...

Perhaps: "The English and the Americans are two great peoples separated by a common language" ;-)
Made only worse by the fact I'm neither English nor American...
Feel lucky that it probably makes a lot of other things better ;-)
 

Latest posts

Back
Top