OMNITRONIC MM-2USB Mic?

[Mikko Pihlajamäki]

Looking for budget solution mic to use instead on umik-1.
Found this:
Omnitronic MM-2USB (https://www.steinigke.de/en/mpn13030923-omnitronic-mic-mm-2usb-usb-condenser-measurement-mic.html)

Frequency responce looks obviously too good, but could this be used as an entry level mic?

 

[luegotelodigo]

I've measured several 1/4" electret capsules and I don't believe that specs. They claim it's pressure gradient and omnidirectional... it's a shame.

 

[Matthew J Poes]

I suspect it's about as good as any of the other 1/4" capsule based USB mics on the market, maybe even made by the same manufacturers. At LF's it's fine, at mid/HF's its suspect.

I bought a 1/2" capsule measurement mic as an upgrade to my USB and other 1/4" capsule mics. Mine came with a NIST certified response, so I assume mine is accurate. The response is generally similar to the USB mic, but once you get past 3khz I see differences. By 15khz I see BIG differences. If I was designing a speaker or full range EQ, I would want a lab grade mic. If I was just doing basic DIY home acoustic testing, I think these are ok.

I think there biggest fault remains the noise floor. It's high enough to mess up decay based measurements as it decays into mic noise before room noise typically. It's mostly a prblem at HF's.

 

[Matthew J Poes]

I've measured several 1/4" electret capsules and I don't believe that specs. They claim it's pressure gradient and omnidirectional... it's a shame.

If I understand correctly, the response is corrected in the preamp built into these. Earthworks uses a panasonic 1/4" capsule and has a dead flat response. A fellow reviewer friend uses one for his measurements and its both similar to my 1/2" measurement mic and has been independently tested and verified. The USB mic I own was tested by Cross-spectrum and while it does have and need a correction file, the correction isn't so great, so I have to assume they apply some correction internally (as earthworks does). I think there biggest problems is that the response remains inaccurate above 15khz, quality control is so bad you never know what your going to get, and they have VERY high self noise. My Dayton USB mic has a noise floor that rises as high as 30dB by 10khz.

My 1/2" mic is a free field mic and the NIST file shows it is flat to nearly 30khz, its only 3dB down at 31.5khz, the upper limit of it's response table.

If you aren't designing or characterizing speakers and you aren't using it for EQ above say 500hz, does the response issues really matter? I've always figured that these are good enough for basic DIY acoustic measurements and Bass EQ. As long as you don't get one of the ones whose QC was so bad as to screw up the LF response, they all see pretty flat down to 20hz or so, good enough for EQ purposes, right?

 

[luegotelodigo]

Are you sure Earthworks use a Panasonic capsule? I don't think so, they have far less distortion than the Panasonics I know.

My ECM 8000 have a sensitivity about -38 dB@ 250 Hz but it quickly drops over 130 dB due to the enormous distortion. Second harmonic distortion has a maximum about 130 dB (-20 dB) while 3rd harmonic keeps rising. You can't measure harmonic distortion of high-end loudspeakers using this kind of mic. I think noise was 32-33 dBA and it's not bad given the small diaphragm.

I love measuring with cheapest gear but it has limitations.

 

[Matthew J Poes]

Are you sure Earthworks use a Panasonic capsule? I don't think so, they have far less distortion than the Panasonics I know.

My ECM 8000 have a sensitivity about -38 dB@ 250 Hz but it quickly drops over 130 dB due to the enormous distortion. Second harmonic distortion has a maximum about 130 dB (-20 dB) while 3rd harmonic keeps rising. You can't measure harmonic distortion of high-end loudspeakers using this kind of mic. I think noise was 32-33 dBA and it's not bad given the small diaphragm.

I love measuring with cheapest gear but it has limitations.

It’s true the mic has a relatively low peak limit but I guess I don’t see the problem for average users. Most people measure the linear responses of their system and use that to apply eq at low frequencies, or to assess room decay measurements. It seems to work ok for that.

For anything serious, you are right, it shouldn’t be used.

As for Earthworks, I was told that by an Earthworks engineer a while back, but to be honest, now that I think back on it, he may have been only referring to the M23. Ill email him and see if they will state what capsules are used.

 

[luegotelodigo]

SPL and phase of my ECM 8000 compared to a PCB 378B02 (20 dB attenuated), no smoothing, not bad

 

[AudiocRaver]

Matt is our authority on measurement mics, and has earned the title. Agreed, for speaker measurement or any pro application, a solid NIST-traceable mic is a must. For me, even for room EQ, above 15 kHz or so is the gravy range, and most speakers/rooms seem to have rolloff at or above that and are pretty uncorrectable, so my approach above the main HF rolloff point is live with what you got.

 

[Matthew J Poes]

Matt is our authority on measurement mics, and has earned the title. Agreed, for speaker measurement or any pro application, a solid NIST-traceable mic is a must. For me, even for room EQ, above 15 kHz or so is the gravy range, and most speakers/rooms seem to have rolloff at or above that and are pretty uncorrectable, so my approach above the main HF rolloff point is live with what you got.

Hah I appreciate the vote of confidence Wayne. I’m this case I’ll readily admit that @luegotelodigo knows a bit more than I. He formally worked for B&K (right?) and has a collection of high end mics I could only dream of. I have one one good mic and a handful of inexpensive 1/4” models.

But none the less, your point is right on. The response on these cheap mics is good in a restricted range, but both quality control and cheap design compromise their use. The HF response, even corrected, is not a great match to my better mic. I don’t like the idea of using such mics to apply eq up above 15khz and suspect you do more harm than good.

 

[luegotelodigo]

Yes, I worked in B&K for two years. NIST traceability is just traceability and a sensitivity calibration means nothing if you're using the mic to calibrate sound systems. No sensor is perfect and when you use correctly a WM-61 (1,5 $) you get the same results as with the best lab grade mic.

I have a lot of microphones but the ones I use more often are a pair of ECM8000. A few years ago I was lecturing about microphones and I gave some to my students, one of them decided to push the diaphragm of a B&K 4189. From that moment ECM 8000 became a must if somebody wanted to test anything. I have measured their noise, sensitivity, distortion, directivity, and for the typical use in REW they are as good as a B&K.

By the way, in only a few cases I'd eq a loudspeaker above 15 KHz, so few that I'd encourage you not to do it.

 

[Matthew J Poes]

Yes, I worked in B&K for two years. NIST traceability is just traceability and a sensitivity calibration means nothing if you're using the mic to calibrate sound systems. No sensor is perfect and when you use correctly a WM-61 (1,5 $) you get the same results as with the best lab grade mic.

I have a lot of microphones but the ones I use more often are a pair of ECM8000. A few years ago I was lecturing about microphones and I gave some to my students, one of them decided to push the diaphragm of a B&K 4189. From that moment ECM 8000 became a must if somebody wanted to test anything. I have measured their noise, sensitivity, distortion, directivity, and for the typical use in REW they are as good as a B&K.

By the way, in only a few cases I'd eq a loudspeaker above 15 KHz, so few that I'd encourage you not to do it.

The issue is that a number of room corrections automatically correct the full bandwidth and you can’t prevent it. This was my point.

I need a better mic for the lower noise and because I need to provide manufacturers proof of the accuracy of my equipment. I have found enough errors that some manufacturers took issue with my measurements and wouldn’t allow them to be published (outside of my work for AV NIRVANA I also do or support measurements for other places).

 

[Matthew J Poes]

Yes, I worked in B&K for two years. NIST traceability is just traceability and a sensitivity calibration means nothing if you're using the mic to calibrate sound systems. No sensor is perfect and when you use correctly a WM-61 (1,5 $) you get the same results as with the best lab grade mic.

I have a lot of microphones but the ones I use more often are a pair of ECM8000. A few years ago I was lecturing about microphones and I gave some to my students, one of them decided to push the diaphragm of a B&K 4189. From that moment ECM 8000 became a must if somebody wanted to test anything. I have measured their noise, sensitivity, distortion, directivity, and for the typical use in REW they are as good as a B&K.

By the way, in only a few cases I'd eq a loudspeaker above 15 KHz, so few that I'd encourage you not to do it.

One last thing I’ll mention. The ECM 8000 has had some pretty serious QC problems as of late and most people stopped recommending it. cross-spectrum got an entire batch of bad mics. The response was WAY off. Like not even correctable it was so bad. I got ahold of one of these and took it apart. The problem I found that I believe was causing this was a sticker that was on the back of the capsule that I didn’t have on my known good sample and parts values that didn’t match in the preamp. The soldering was also obviously sloppy and with evidence of rework. It’s possible they are all good now, but I’ve been permanently scared off of recommending them. Cross-Spectrum also gets occasional bad batches of the Dayton’s it sells and I’ve known of some bad USB mics from time to time. The big problem here is that for the average person, they have no way of knowing this is true. What are they supposed to do?

I love that these cheap mics exist. Within the bandwidth that matters they are just as accurate as a lab grade mic, as you say, as long as you have a good one. They need to get the QC back under control though.

 

[luegotelodigo]

The ECM 8000 has had some pretty serious QC problems as of late

I have some colleagues developing limitters for pubs and discos and they bought lots of cheap mics this year and I tested them, the pair of ECM 8000 were quite good. I've also checked a lot of friends' 8000 just buy FF comparison in a coincident place and only one was too bad.

The big problem here is that for the average person, they have no way of knowing this is true. What are they supposed to do?

Avoid high frequency EQ, hahaha!

The problem with average people is we have amazing tools like REW and good cheap microphones available but we aren't Manfred Schröder and we want to hit a button to make our stereo perfect. That leads into a "ultra flat frequency onanism" with bad sonic results and feed their antagonist "Hi-End homeopathy AKA modern technology is rubbish" with also bad sonic results.

I will make an exposition with real measurements and will finish with the question "How can we correct this?" and I hope we all agree this is not an easy task, a lot of approaches are respectable but none of them are the absolute truth:

I measure this monitor in the anechoic chamber (1) to get a good response (2), then I move it to the control room at the same distance (3) to get this response (4) I want to correct. Comb-filtering is created by the interference of reflections coming from different angles hence depending on the directivity of the mic (ECM DPF) which is near to omnidirectional but we hear with our ears and they are not omni (Dummy DPF) so below let's say 500 Hz you can correct the curve to avoid the frequency effects of a non minimum-phase artifact that probably you can't hear because your HRTF already made a direction dependent correction. We are in the same point if we start to think our class 1 B&K 4189 microphone is not a good tool anymore and we want a 5128 (high-frequency green man)

Do we agree this is a complex issue or shall I start with psychoacoustics?

(As you surely noticed English is not my first language so please correct me)

 

[Matthew J Poes]

I have some colleagues developing limitters for pubs and discos and they bought lots of cheap mics this year and I tested them, the pair of ECM 8000 were quite good. I've also checked a lot of friends' 8000 just buy FF comparison in a coincident place and only one was too bad.


Avoid high frequency EQ, hahaha!

The problem with average people is we have amazing tools like REW and good cheap microphones available but we aren't Manfred Schröder and we want to hit a button to make our stereo perfect. That leads into a "ultra flat frequency onanism" with bad sonic results and feed their antagonist "Hi-End homeopathy AKA modern technology is rubbish" with also bad sonic results.

I will make an exposition with real measurements and will finish with the question "How can we correct this?" and I hope we all agree this is not an easy task, a lot of approaches are respectable but none of them are the absolute truth:

I measure this monitor in the anechoic chamber (1) to get a good response (2), then I move it to the control room at the same distance (3) to get this response (4) I want to correct. Comb-filtering is created by the interference of reflections coming from different angles hence depending on the directivity of the mic (ECM DPF) which is near to omnidirectional but we hear with our ears and they are not omni (Dummy DPF) so below let's say 500 Hz you can correct the curve to avoid the frequency effects of a non minimum-phase artifact that probably you can't hear because your HRTF already made a direction dependent correction. We are in the same point if we start to think our class 1 B&K 4189 microphone is not a good tool anymore and we want a 5128 (high-frequency green man)

Do we agree this is a complex issue or shall I start with psychoacoustics?

(As you surely noticed English is not my first language so please correct me)

I absolutely agree with you. I’ve never not in this regard.

I did the lay equivalent this in my Dayton Audio MK402 review. I took the speaker outside, fully characterized it, then developed an eq to correct its response (this is not a flat speaker). I then brought it into my office and applied another layer of eq to correct the response below around 300-500hz, most of the focus below 100hz. I think that is the most you can get away with.

I don’t think anyone should be manually applying eq above 300-500hz other than very general shaping such as tone controls (shelf filters). The point I made earlier was targeting Audyssey and DIRAC. These are auto-corrections that apply to the full bandwidth of the speaker. If the mic is off, the correction can be similarly off.

When I get back to my laptop I’ll try to dig up my measurements that highlight this phenomena. What I found is that DIRAC tried to flatten and extend the HF response, but the roll-off was actually the mic response. The true response of the speaker after correction had elevated treble.

 

[Mikko Pihlajamäki]

so yeah, about this OMNITRONIC MM-2USB. Is it OK for room acoustics and hobby level speaker design?

 

[Matthew J Poes]

so yeah, about this OMNITRONIC MM-2USB. Is it OK for room acoustics and hobby level speaker design?

Probably. It’s a complete unknown as we have no test data for it. The other cheap mics we mentioned were accurate enough for such purposes.

 

[luegotelodigo]

so yeah, about this OMNITRONIC MM-2USB. Is it OK for room acoustics and hobby level speaker design?

I know I'm replying after a long time but I tried an omnitronic mic and tested great. It is a good choice for room measurements.