I don’t think I was attacking anyone in particular. Here and here are some more useful sources, the latter being written purely in layman’s language and introducing an element I didn’t know yet, the effect that sound sources often behave as point sources for low frequencies, while behaving as radiating surfaces for higher frequencies. That adds a very logical explanation, at least to me. But I won’t repeat all that writing here 😉 The first one requires payment (I didn’t) or membership (I am not).
Received a ISEMcon EMX-7150 mic few days ago,
yesterday I downloaded calibration data and checked it for the first time on a Moto M2:
DUT is the "Scandinavian Connection":
The right one, on a similar position.
Active speakers with FA123, 2x Scan Disco 26W in CB, SB15NBAC, TAC/GB (different to the DXT on the pic);
bucking magnets on mid and woofers, passive notches in series to mid and tweeter tuned to cone/dome breakup resonances.
Did only some shots from the hip by hand without mike stand, but got pretty happy results.
Distance was approx. 0,5m, mic aimed between mid and tweeter, so SPL @ 1m was roughly above >90dB (91~92 dB or so):
Well, that measured noisefloor looks very nice for both the mic and the speaker...
(FR dip at 120Hz and peak at 50Hz are caused by room influence)
Used a Monacor ECM-40 mic on MPA-102 preamp and internal Asus Xonar PCI Soundcard with ARTA SW for the last ~12 years. Got a calibration file with it and have send it again for calibration about 6 years ago. So FR measurements were fine, but the ECM-40 had too much own noise and k2 to make reliable distortion measurements.
Have changed my setup completely now to improve performance and flexibility. From PC with Win 10 to Laptop with Linux Mint. Will keep ARTA for FR measurements, but changed to REW for distortion measurements now (also excited to test the FSAF implementation in REW beta soon). And have changed the DIY ARTA Switchbox to a DATS V3 for impedance / electrical measurements. Got it all running under Linux (ARTA and DATS with WINE) and signals level calibrated.
So far I'm very happy with my purchase and feel prepared to do a lot of distortion measurements reliably now. I should have done that much earlier and can only recommend to pay a few bucks more for a good mike...
yesterday I downloaded calibration data and checked it for the first time on a Moto M2:
DUT is the "Scandinavian Connection":
The right one, on a similar position.
Active speakers with FA123, 2x Scan Disco 26W in CB, SB15NBAC, TAC/GB (different to the DXT on the pic);
bucking magnets on mid and woofers, passive notches in series to mid and tweeter tuned to cone/dome breakup resonances.
Did only some shots from the hip by hand without mike stand, but got pretty happy results.
Distance was approx. 0,5m, mic aimed between mid and tweeter, so SPL @ 1m was roughly above >90dB (91~92 dB or so):
Well, that measured noisefloor looks very nice for both the mic and the speaker...
(FR dip at 120Hz and peak at 50Hz are caused by room influence)
Used a Monacor ECM-40 mic on MPA-102 preamp and internal Asus Xonar PCI Soundcard with ARTA SW for the last ~12 years. Got a calibration file with it and have send it again for calibration about 6 years ago. So FR measurements were fine, but the ECM-40 had too much own noise and k2 to make reliable distortion measurements.
Have changed my setup completely now to improve performance and flexibility. From PC with Win 10 to Laptop with Linux Mint. Will keep ARTA for FR measurements, but changed to REW for distortion measurements now (also excited to test the FSAF implementation in REW beta soon). And have changed the DIY ARTA Switchbox to a DATS V3 for impedance / electrical measurements. Got it all running under Linux (ARTA and DATS with WINE) and signals level calibrated.
So far I'm very happy with my purchase and feel prepared to do a lot of distortion measurements reliably now. I should have done that much earlier and can only recommend to pay a few bucks more for a good mike...
Could you present us some sort of a comparison/overlay of both mic 's distortion and SPL measurements?
I'm trying to choose between the EMX-7150 and the MicW M215 (and, perhaps another or two). What was the deciding factor(s) for you?
@Boden : I might do that, I'm also curious - probably not within the next 4 weeks, but around christmas I'll have some free time for this.
@Paul Ebert :
I wanted to step up from my 80.- € mic; there are some nice offers of pre-polarized 48V type with improved FR and capable distortion limits; MicW M215 as mentioned, popably Earthworks M23, NTI Mini SPL, Haun MBC 550. But all these cost much more that 300,- € I have paid for the Isemcon EMX, at least with the purchase options at my location.
For 300,- alternatively, there is Mipro MM-100 that looks not bad, but not much data available. Serious competitor to the EMX is Audix TM-1 Plus with comparable properties, calibration data and price. Isemcon convinced me somehow by advertising with transparent technical data and specs that I will get a capable and reliable product. The Audix might be as good so the decision was tight.
I think with such an Audix or Isemcon mic pre-polarized capsule technology is pretty much maxed out, so why pay 150-300 more for Earthworks or MicW for "only" comparable performance (at least I assume this)...
@Paul Ebert :
I wanted to step up from my 80.- € mic; there are some nice offers of pre-polarized 48V type with improved FR and capable distortion limits; MicW M215 as mentioned, popably Earthworks M23, NTI Mini SPL, Haun MBC 550. But all these cost much more that 300,- € I have paid for the Isemcon EMX, at least with the purchase options at my location.
For 300,- alternatively, there is Mipro MM-100 that looks not bad, but not much data available. Serious competitor to the EMX is Audix TM-1 Plus with comparable properties, calibration data and price. Isemcon convinced me somehow by advertising with transparent technical data and specs that I will get a capable and reliable product. The Audix might be as good so the decision was tight.
I think with such an Audix or Isemcon mic pre-polarized capsule technology is pretty much maxed out, so why pay 150-300 more for Earthworks or MicW for "only" comparable performance (at least I assume this)...
Last edited:
This counts for you as well.
It's a two way street!!
One that you keep on ignoring.
Since there is no constructive input at this point, it's not even a discussion.
Btw, can't appreciate al the emotion, harsh comments and personal attacks.
Don't think it's very respectful, mature or professional.
Let alone helpful.
To give a general answer; Yes I have done this in the past
I already wrote that a couple of times.
To get back to this, two reasons: difference in distance (even in the far field) between the mics lead to difference in sound level at the diaphragms. The closer the mikes to the source, the bigger the difference. The other: lots of actual sound sources, including the human voice, as a sound source, behave differently at different frequency regions. Which in turn again leads to varying, frequency related level differences between the two mikes.
b_force, no one is making harsh comments or personal attacks. Just expressing surprise that you have conducted an experiment some of us have done in the past with very different results. 😉
I'm sorry I'm too lazy to do the math ... which would be above the heads of most of this forum anyway 😊
But you CAN help us understand your discovery by telling us a bit more about your experiment.
- What sort of directional mike were you 'simulating' with your 2 omnis? A Fig-8 or a cardioid?
- What omnis did you use?
- How did you test for proximity? Speaking into the 'simulated directional' mike at several distances is valid for us.
- When you make a Fig-8 from 2 omnis you have to apply 6dB/8ve LF boost from your lowest frequency to the frequency where the distance between the 2 omnis becomes significant and the Fig-8 pattern breaks down. IIRC, this is shown in the B&K notes for their Sound Intensity wand. Otherwise, your 'directional' mike will have VERY poor LFs. Between what frequencies did you have your EQ from?
BTW, markbarkk's explanation for your two omnis is a very good one. My $0.02 is that what he describes happens with a 'point' source where the spl (what an omni measures) varies inversely with the radial distance squared. And you need to put in the LF EQ I describe above for a 'flat' directional mike.
If you ask acousticians what is the best mic for measurments, they will very often come with very expensive ones like the Gefell M 373 or the HBK (Brüel & Kjær). But those are not needed for speaker design i think. Even cheap USB-mics can be used if you use them right and if they are calibrated (best don't trust factory calibration if you need preceision). The right calibration is more important than all the rest.
But to be honest, a cheap Dayton or MiniDSP (UMIK) does the job quiet good today, especially if it's right calibrated. If you use better, you first need the right (anchionic or semi-anchionic) space to measure speakers right. If an Earthworks or DPA makes you feel more assure, off course you should get them (they are more precise when calibrated), but i don't think it matters that much, especially for diy-amateurs. You could off course also buy a Kippel NFS system if you got a 100K for it, but if you can't build a decent speaker/crossover with an Umik, it's more your skills that lack, not the mic..,
But to be honest, a cheap Dayton or MiniDSP (UMIK) does the job quiet good today, especially if it's right calibrated. If you use better, you first need the right (anchionic or semi-anchionic) space to measure speakers right. If an Earthworks or DPA makes you feel more assure, off course you should get them (they are more precise when calibrated), but i don't think it matters that much, especially for diy-amateurs. You could off course also buy a Kippel NFS system if you got a 100K for it, but if you can't build a decent speaker/crossover with an Umik, it's more your skills that lack, not the mic..,
@waxx
In general I agree. I developed lots of speakers in the last decade with a 80.- mic.
Individual calibration file for reliable FR measurement is mandatory, but then you're complete to size xovers and validate designs successfully.
Be careful with the USB mics for speaker development; correct acoustical phase measurement requires loopback or timing reference capability - not all USB mics support this.
I've invested now 300.- for a better mic, whereas "better" for me = lower noise and less distortions. The fact is, I see more now in the noisefloor of my speakers and can develop them better further regarding THD.
But as a Hifi-Speaker DIY hobbyist you clearly don't need more than this.
An externally polarized B&K Class 1 capsule is required for 1) Official / Professional certified and proofed measurements 2) measurement of extreme quiet sound and/or extreme loud sound (such capsules have quite low noise and very high SPL capability).
P.S.:
If I would win the lottery tomorrow, it could happen that I might buy one, LOL... 🤣
In general I agree. I developed lots of speakers in the last decade with a 80.- mic.
Individual calibration file for reliable FR measurement is mandatory, but then you're complete to size xovers and validate designs successfully.
Be careful with the USB mics for speaker development; correct acoustical phase measurement requires loopback or timing reference capability - not all USB mics support this.
I've invested now 300.- for a better mic, whereas "better" for me = lower noise and less distortions. The fact is, I see more now in the noisefloor of my speakers and can develop them better further regarding THD.
But as a Hifi-Speaker DIY hobbyist you clearly don't need more than this.
An externally polarized B&K Class 1 capsule is required for 1) Official / Professional certified and proofed measurements 2) measurement of extreme quiet sound and/or extreme loud sound (such capsules have quite low noise and very high SPL capability).
P.S.:
If I would win the lottery tomorrow, it could happen that I might buy one, LOL... 🤣
Last edited:
Yes This ^
The title of the thread is low distortion omni mics. @tktran303 acknowledged early on that a low distortion mic is not necessary for filter development, but it is necessary for someone who wants to measure and characterize distortion products below 0.1%, which is about the typical limit for typical mics.
The title of the thread is low distortion omni mics. @tktran303 acknowledged early on that a low distortion mic is not necessary for filter development, but it is necessary for someone who wants to measure and characterize distortion products below 0.1%, which is about the typical limit for typical mics.
No speaker goes that low in distortion, and for electronic products there are way better ways to measure distortion than trough a speaker because the speaker distortion will mask it. A speaker is a good speaker if the THD is less than 0.5%, very often that is not the case. If you want to test electronics, get an AP or similar measuring device.
The Umik-1 was tested and is accurate to -115dB (0.00018%) several times and extensive discussed on sites like ASR and that of REW. The only issue that it has is the phase because there is no loopback calibration. But there are ways arround that issue.
What is a big difference is the standard calibration, and how much the uncalibrated mic differ from flat. Those expensive brands do that a lot better and are officially certified, a Umik not.
But again, if you feel more assure, buy that Josephon, Earthworks or DPA (those are known good brands) or whatevre you want.
But it's proven more than once that the Umik and similar cheap mics can do almost all that those very expensive can, only with more fidleing to get it right (and more room for error). If i was very rich, i would probally also get a Kippel NFS, that's way more important than a Lambo or other expensive car that would cost more... It's just easier and works faster with little room for error. That's why people/companies buy those.
The Umik-1 was tested and is accurate to -115dB (0.00018%) several times and extensive discussed on sites like ASR and that of REW. The only issue that it has is the phase because there is no loopback calibration. But there are ways arround that issue.
What is a big difference is the standard calibration, and how much the uncalibrated mic differ from flat. Those expensive brands do that a lot better and are officially certified, a Umik not.
But again, if you feel more assure, buy that Josephon, Earthworks or DPA (those are known good brands) or whatevre you want.
But it's proven more than once that the Umik and similar cheap mics can do almost all that those very expensive can, only with more fidleing to get it right (and more room for error). If i was very rich, i would probally also get a Kippel NFS, that's way more important than a Lambo or other expensive car that would cost more... It's just easier and works faster with little room for error. That's why people/companies buy those.
You wouldn't happen to have a links to that?
@IamJF 's thread on microphone distortion showed an Earthworks M50 not being able to do much better than 0.03%.
By "that low" do you mean -60 dB (0.1%) ?
In my experience, it is pretty common to measure a driver or complete system at -60 dB harmonic distortion.
Given my mic and equipment, I think the limit of resolution in the above plot is -60 dB. I don't trust the accuracy below that. But this shows that this speaker system is at least -60 dB from 500 Hz - up...
I don't know how you measure and how truth the graph is, but i know that Erin's Audio Corner uses a Kippel in standardised measurement protocols and with what should be one of the best woofers arrround he gets this;. In the range between 100hz and 3Khz the distortion is not even fully below -50dB.
most drviers used by diy's don't even come near this. And this is on low power, at high power those distortions rise fast. That is what is wrong with the post about the measurment mic test, they measure at extreme high volume untill breakup. But you never usea a measurment mic like that for speaker calibration, you stay arround 90dB max. How it will break up is already visible then. But if you feel the need to upgrade, feel free (it's your money), the Gefell M373 would be the best say almost all acousticians i know who need such microphones. HBK (B&K) a exellent second.
But don't think because someone uses a Umik he can't make a good speaker, that is straight ********. You can do the far most with it, it's only on high or very low volume that you need those very expesnive ones, because there they make the difference, on normal volume not (for speaker design).
But don't think because someone uses a Umik he can't make a good speaker, that is straight ********. You can do the far most with it, it's only on high or very low volume that you need those very expesnive ones, because there they make the difference, on normal volume not (for speaker design).
Yes, and this Klippel NFS-based plot from Erin is one of the motivations for @tktran303 to start this thread.
I believe his current theory is that the microphone used in the NFS system has self noise that is high enough to mask distortion below approximately -60 dB. Other mics were able to duplicate the distortion plots that Purifi published, which are about 15 dB lower than what the NFS system measures.
I am not going to argue the case for @tktran303, he is more than capable of defending his hypothesis... and I am not familiar with all the data and analysis that has gone into this thread. So I will leave it to him.
I believe his current theory is that the microphone used in the NFS system has self noise that is high enough to mask distortion below approximately -60 dB. Other mics were able to duplicate the distortion plots that Purifi published, which are about 15 dB lower than what the NFS system measures.
I am not going to argue the case for @tktran303, he is more than capable of defending his hypothesis... and I am not familiar with all the data and analysis that has gone into this thread. So I will leave it to him.
Microphones could get way lower but are noise limited at lower levels. THD of the M50 starts to be higher as the noise in the 100dBSpl area, and then K2 dominates BY FAR, higher orders are way lower. But depending on the analysis your THD measurements are more or less sensitive to noise.
So it's possible to measure lower THD but <0,01% it's getting very hard to do.
When measuring at 86dBSpl 0,01% THD would mean an SPL level of 6dBSpl ... normal environments can not do that. The -60dB from the graph are already very likely the measurement boundary - depending on the THD analysis.
I agree with @waxx. You can go a long way with a simple electret condenser mic like one mentioned in the VituixCAD2 manual. In fact, that's what I used:
The exploration into the precision of measurement of harmonics started because I found discrepancies between various sources, which is explained in the original post.
The importance of doing such measurements, is not yet clear to me. One of the main constraints is the background noise, as @IamJF explained.
How important is it to have HD and IM down -50dB or more, when one's room has background noise of ~30dB(A)?
The exploration into the precision of measurement of harmonics started because I found discrepancies between various sources, which is explained in the original post.
The importance of doing such measurements, is not yet clear to me. One of the main constraints is the background noise, as @IamJF explained.
How important is it to have HD and IM down -50dB or more, when one's room has background noise of ~30dB(A)?
Last edited:
A room at night in a silent environment can go very low in background noise! As long as the fridge or projector (or kids) aren't on 😉
But for me THD is more a technical measurement to check if my system is linear and reproduces what I send into it. It shows proper frequency ranges of drivers, problems with motors, design principles (e.g. ring radiators and wide surround tweeters have high H2). And it is for sure better to have lower THD as higher THD (unless you want "sound" -> ring radiators).
What HD can be heard in what circumstances (SPL level, frequency, signal spectrum, signal dynamics) ... is more complicated as making low THD speakers, so I stick with low THD speakers. 😎
But for me THD is more a technical measurement to check if my system is linear and reproduces what I send into it. It shows proper frequency ranges of drivers, problems with motors, design principles (e.g. ring radiators and wide surround tweeters have high H2). And it is for sure better to have lower THD as higher THD (unless you want "sound" -> ring radiators).
What HD can be heard in what circumstances (SPL level, frequency, signal spectrum, signal dynamics) ... is more complicated as making low THD speakers, so I stick with low THD speakers. 😎