I think it is the Q of what is being fixed that bugs me.
If I try to eq the calibration of my oldest dayton mix, I can see it is around a Q=10 to fix. So in reality, the mic naturally has q=10 hump right where I have noticed problems in previous measurements. I am now wondering if the Q of the correction is to blame.
Good advice. I might use multiple mics no matter what. I am looking into calibration.
o I thought I'd compare the calibration files of a few mics. Huge props to Audix support for sending me some random cal files. Comparing the audix cal (blue shades), my 2 dayton mics(red shades), and an a m23r earthworks I found online (green)
The cal for the earthworks looks less extreme. The dayton look the worst. The audix in the middle.
That's also a great thing of my earthworks M50 - it doesn't need calibration up to 20kHz!
But there is no reason a microphone with omni capsule has a narrow peak of 1,5dB at 330Hz and the next one doesn't (Dayton red) - this STRONGLY looks like measurement tolerance! I'm sure these cheap mics are not calibrated in a proper anechoic room, that's probably a small chamber in a chinese production environment with noise and disturbances around. Don't take anything +-1dB too serious with the Daytons!
Audix already looks better but for e.g. Audix C ... that's not a microphone behaviour, that's room behaviour!
Earthworks calibration seems more on track what to expect.
Taking into account the cost of a calibration ... I would consider an Earthworks already as low budget.
Have you checked Dr Jordan’s site? For those who don’t trust Hifi selbstbau, but both are quite reliable. You don’t need full calibration, frequency correction is enough. But you know. Just as you don’t know your unchecked Earthworks…
The Earthworks M23 has a calibration file. The M23R (Reference) is tested to have +/- 0.5 to db without a calibration file. So not sure where you found a "calibration file" for the M23R but I'm skeptical of it's authenticity.
Disclaimer: I own the M23R. I tested it against the Dayton EMM and Behringer ECM8000 HERE
Disclaimer: I own the M23R. I tested it against the Dayton EMM and Behringer ECM8000 HERE
The stuff below 1kHz is almost certainly due to poor mounting in your 'calibration' rig. You need to get the mike diaphragms within 1mm of each other with NOTHING else nearby for calibration.
Did you measure the mikes separately?
Can you show a pic of your measurement rig?
This is more like I expect but is still wonky. I don't expect even a Behringer ECM8000 to show wobbles until it nears 10kHz
Post your cal file.
The cal in my graph came from a user on audioscience. It does only adjust less than .5db?
This is just a display of the calibration files provided to fix the mic. I did no measuring.
Then IamJF's comments
apply. I don't believe ANY of these except the Earthworks.
Hello,
I look at this from the point of view of bin width which is determined by sample size.
Neither the hardware ( microphone or amplifier and other stuff ) nor the software know or care about there being a difference between Harmonics or Noise. Both are summed in the sampled bin width. The wider the bin more noise is included in Harmonic calculation. large sample size and averaging noise over several measurements will lower the noise included in the calculation of Harmonic distortion.
Yes the stuff that shows up on a distortion plot may not be distortion at all, but only be noise.
Thanks DT
Ok, now earthworks sent me 3 cal files. They aren't all set to the same zero but this gives me an idea. (i did adjust the earthworks -1.6db to each). Yellow is the earthworks cal I found on another site.
Reds= my dayton mic calibration files, Blues = Audix calibration files, and Greens = earthworks m23R calibration files.
I also own a 23R - there is no cal file. You sure you have the 23 or the 23R?
You got you calibration file from a user on a forum? That’s not how this works… only a manufacturer should send you a cal file unique to your mics serial number.
Measuring mikes in a non-anechoic space, you must make a Quasi Anechoic Measurement by getting rid of the reflections. eg p116 et al of Clio 7 manual.
Otherwise you are just measuring room effects.
Only the Earthworks measurements above seem to do this properly and I'm suspicious even of those from designing and making Measurement Mikes circa 1980. Maybe electret capsules have become worse in the last 44 yrs 😊
Certainly the original Panasonic capsules used by Earthworks for their early mikes are now Unobtainium and there doesn't seem to be a similarly consistent replacement.
The LF cutoff needs to be estimated (measured? guessed?) by other means but it's usually very simple to do even if you don't use Benjamin's method above.
If anyone is making measurement mikes and can advise us, please share your trade secrets.
BTW, the Clio 7 manual is the ONLY good & accurate explanation of Quasi Anechoic I've found on da WWW or da AES library. If anyone knows different, please tell. Alas IIRC, later Clio manuals omit this treasure.
Otherwise you are just measuring room effects.
Only the Earthworks measurements above seem to do this properly and I'm suspicious even of those from designing and making Measurement Mikes circa 1980. Maybe electret capsules have become worse in the last 44 yrs 😊
Certainly the original Panasonic capsules used by Earthworks for their early mikes are now Unobtainium and there doesn't seem to be a similarly consistent replacement.
The LF cutoff needs to be estimated (measured? guessed?) by other means but it's usually very simple to do even if you don't use Benjamin's method above.
If anyone is making measurement mikes and can advise us, please share your trade secrets.
BTW, the Clio 7 manual is the ONLY good & accurate explanation of Quasi Anechoic I've found on da WWW or da AES library. If anyone knows different, please tell. Alas IIRC, later Clio manuals omit this treasure.
I got some new files direct from earthworks. I asked for 3 random files. We can ignore the one I found. (see the post 310)
There is a cal for the 23R. https://earthworksaudio.com/support/ecf/
I don't own any mics other than the 2 daytons.
In da 21st century, the method of choice to measure response & distortion of speakers & mikes is Prof Angelo Farina's method used by AP, ARTA, Clio, bla bla and REW's ESS method.
It produces the most accurate Response & Distortion in the shortest possible time in the presence of noise (ALL measuring environments have noise).
You get a handle on how much noise by just measuring again with the speaker muted.
Last edited:
Ahh I see. Interesting! all mics I've purchased previously mention in their documentation getting a cal file, the 23R I got made no such mention so I assumed they didn't have/need them.
Yes bla bla, Farina's method method is used by the APx analyzer and APx500 software on my bench.
The software is unable to separate noise from harmonic peaks.
If the test frequency F is 1kHz then the measured SPL at at 2F is used to calculate H2 Distortion.
Noise from a laminated steel core inductor will measured and calculated as if it were real distortion from the driver. The software does not know the difference.
Audio Precision has a pretty good application note, slightly less Jurassic
https://ampslab.com/RECOMMMENDED READ/LOUDSPEAKER MEASUREMENTS/AUDIO PRECISION.pdf
Thanks for this fluid. It is indeed a useful document for speaker measurement.
But I think the Clio 7 manual is a better explanation of Quasi Anechoic .. especially for DIYers wanting to measure stuff in their living room.
Actually it does. The equivalent 'swept filter' for each Harmonic is so narrow band that very little noise is picked up. That's why Angelo's method is the fastest method to get a given accuracy in the presence of noise.
If the 'noise' is actually a 'distortion', like your 'steel core noise', it will still be picked up at the correct level by Angelo's method. It might be swamped by each Harmonic but then you'll know which is more audible / important 😊