Polarity anomaly - UMIK-1 vs. UMIK-2

[AustinJerry]

Was the 2/24/2023 measurement done with the Java driver and EXCL device?

No, the ASIO4All driver was used. The 7/5 measurement used the Excel drivers. Do you think the driver selection would make a difference?

Did you look at the MDAT file? Something peculiar is going on with the Dirac Off/On measurements.

 

[thothsong]

Definitely worth trying the ASIO driver again, to see if the polarity changes.

What about the Dirac off vs on measurements is peculiar?

 

[AustinJerry]

Definitely worth trying the ASIO driver again, to see if the polarity changes.

What about the Dirac off vs on measurements is peculiar?

OK, I ran a fresh set of measurements using both the UMIK-1 and UMIK-2 and switched between using the Java Excel driver and the ASIO driver (MDAT attached). All of the UMIK-2 measurements using both drivers are showing negative polarity, while all of the UMIK-1 measurements are showing positive polarity. Very strange, but consistent.

Another peculiarity that I am trying to understand is the impulse alignment. While the UMIK-1 and UMIK-2 impulse responses are showing inverted polarity, I still would expect them to line up on the horizontal time scale. This is not the case, as shown here (a difference of .0995 ms.):

I am at a loss trying to understand what I am looking at here. Which mic is correct? Does it really matter for the casual measurements I am likely to make? What about a Dirac calibration--will there be a difference depending on which mic I use?

And never mind my comment about a difference between Dirac off and on--after looking at it again, I don't see an issue.

 

[thothsong]

Sorry, I don't have an explanation for the polarity or the timing difference.

 

[AustinJerry]

If you had a choice of which mic to use, which would you pick? Or does it really make a difference?

 

[thothsong]

In that old post, you showed UMIK-1 and UMIK-2 SPL measurements that were basically the same. In your most recent measurements, they're very much different. If you were careful in placing both mics in the exact same position for the most recent measurements, then it looks like an SPL change has also occurred at some point. In your most recent measurements, there's a clear and consistent difference in the 20-200 Hz range between the two mics, and significant variations at higher frequencies. If your environment and audio equipment haven't changed since the beginning of last year, then I'd try comparing current measurements with measurements from last year, UMIK-1 vs UMIK-1 and UMIK-2 vs UMIK-2, especially in the bass range which should be more consistent, to see if either mic has remained stable.

 

[AustinJerry]

Thanks for your suggestion, but for now I'm just trying to understand the polarity difference. While my equipment and environment have remained relatively stable since the "old post" and today's measurements, there could always be some variables, like slight differences in mic placements, or some slight differences in room furniture. I am hoping at some time that JohnM will weigh in on the polarity issue, but John has some personal issues he is dealing with right now. I also have the open ticket with MiniDSP, so they might have some insight.

 

[sam_adams]

The differences in the impulse timing could be linked to something as simple as processing delays in the mics themselves—different ADCs—or a slight change in the position of the mic between measurements. The difference in the timing—99,5µS—is equivalent to two wavelengths at 20KHz or 720˚ of phase. The overarching issue is one of polarity. I am most interested in what the response from MiniDSP will be regarding this issue since it seems to have gone unaddressed for a considerable length of time. @AustinJerry, you have gone the extra mile—as always— to help us understand this issue. Thank you. Now, let's see what MiniDSP has to say.

 

[thothsong]

The differences in the impulse timing could be linked to something as simple as processing delays in the mics themselves

I don't understand how that would change the timing relative to the timing reference. FWIW, I don't see a difference between my UMIK-1 and UMIK-2.

 

[sm52]

AustinJerry,​

Arrange with someone near you that you will come to him with your microphones and take measurements on his computer with both microphones. Java excl drivers, all windows enhancements disabled.

 

[FargateOne]

Thanks for your suggestion, but for now I'm just trying to understand the polarity difference. While my equipment and environment have remained relatively stable since the "old post" and today's measurements, there could always be some variables, like slight differences in mic placements, or some slight differences in room furniture. I am hoping at some time that JohnM will weigh in on the polarity issue, but John has some personal issues he is dealing with right now. I also have the open ticket with MiniDSP, so they might have some insight.

Dear Austin, if you are still there, I just bought a new UMIK-! and it has reversee polarity issues. How did you solved yours?

 

[dcibel]

There's an "invert" check box on the sound card preferences window.

 

[FargateOne]

There's an "invert" check box on the sound card preferences window.

In REW you mean not in Win11. ? I know but I don't have this option for DIRAC. I know Dirac can deal with polarities but I don't think that it is the right thing to do to measure with a bad mic.

 

[dcibel]

If all the measurements are inverted, their relative polarity is still captured. I'm no Dirac expert but I don't see how inverting all measurements would provide a different result for a room EQ scenario.

FWIW it's quite common for amplifiers to invert the output, no one notices as long as they connect red and black terminals correctly.

 

[FargateOne]

Here the answer of Devteam miniDSP
There are some instances where the mic capsule of the UMIK-1 is inverted on the output.
Does it make sense? If yes should I ask for a replacement?

 

[sm52]

If yes should I ask for a replacement?

Yes. If they agree. But instead of Umik-1, it is better to take Umik-2. The Umik-2 has its own clock generator.

 

[AustinJerry]

Dear Austin, if you are still there, I just bought a new UMIK-! and it has reversee polarity issues. How did you solved yours?

I have not revisited this issue since the discussion two years ago. I don't recall ever getting a response from MiniDSP Support, but I could be wrong. Out of curiosity, I ran a few test measurements this morning. I now have three MiniDSP mics, the original UMIK-1 with its CSL custom calibration file (labelled "Old UMIK-1" in the screenshots), a newer UMIK-1 with the calibration file downloaded from the MiniDSP web site (labelled "New UMIK-1"), and the original UMIK-2, also with a custom CSL calibration file.

First observation from my measurements today is that all three mics measure very close to the same, which is the desired result:

Second observation is that the impulse anomaly still exists between the Old UMIK-1 and the UMIK-2. Strangely, the impulse for the New UMIK-1 matches the UMIK-2:

I have not noticed any impact on my overall Dirac calibrations from the impulse anomaly, so I have not pursued any further explanation for why it exists. And contrary to what @sm52 seems to imply, I have never heard a definitive improvement in the UMIK-2 over the UMIK-1.

 

[sm52]

And contrary to what @sm52 seems to imply, I have never heard a definitive improvement in the UMIK-2 over the UMIK-1.

See specifications. Umik-2 has a processor. Any processor needs a clock generator to work. The Umik-1 does not have a processor. It doesn"t need a clock generator. It receives the sample rate from the USB bus. Therefore, many people write that under the same conditions, the impulses of umik-1 move slightly in time.

UMIK-2

USB Reference Measurement Microphone, 32bit/44.1~192kHz, ASIO drivers, 1/2" capsule, Low noise floor, Unique Calibration file, REW and Dirac Live integration

www.minidsp.com

UMIK-1

The UMIK-1 is an omni-directional USB measurement calibrated microphone providing Plug & Play acoustic measurement. From speaker & room acoustic measurement to recording, this microphone provides low noise and accurate results you can rely on.

www.minidsp.com

 

[AustinJerry]

Here is a lengthy discussion concerning differences between the two mics.

https://www.avnirvana.com/threads/umik-1-vs-umik-2.8397/

 

[FargateOne]

Second observation is that the impulse anomaly still exists between the Old UMIK-1 and the UMIK-2. Strangely, the impulse for the New UMIK-1 matches the UMIK-2:

So Great to hear from you Sir!
It is interesting to see that our 2 OLD mics and our news one have polarity reversed. Maybe one day the production line of Umik changed the outputs without any notice. Or we are somehow linked by the same curse...I received this peculiar response from DevTeam:

There are some instances where the mic capsule of the UMIK-1 is inverted on the output.​

Indeed Room EQ Wizard (REW) can invert results in a couple of different ways:​

Of course I knew that REW can reverse polarity but not Dirac. I was afraid that it could affects the DLBC filters results. You seem of the opinion that it is not the case.
I still have to decide if I'll push for a replacement.

Any processor needs a clock generator to work

Internal clock is somehow important I suppose. But here in Canada since near 1 year, ...something happened in the canadian market...don't know what...Umik-2 are more than twice the price of umik-1....so I am not sure that it worth it for my amateur needs.

 

[AustinJerry]

So Great to hear from you Sir!
It is interesting to see that our 2 OLD mics and our news one have polarity reversed. Maybe one day the production line of Umik changed the outputs without any notice. Or we are somehow linked by the same curse...I received this peculiar response from DevTeam:

There are some instances where the mic capsule of the UMIK-1 is inverted on the output.​

Indeed Room EQ Wizard (REW) can invert results in a couple of different ways:​

Of course I knew that REW can reverse polarity but not Dirac. I was afraid that it could affects the DLBC filters results. You seem of the opinion that it is not the case.
I still have to decide if I'll push for a replacement.

Opposite polarity is not an issue as long as polarity is consistent across all speakers. It would be an unusual case if you were to alternate between mics with opposite polarity during a single calibration session.

 

[FargateOne]

It would be an unusual case if you were to alternate between mics with opposite polarity during a single calibration session.

Not at all. The old mic became to give weird results 2 weeks ago(bizarre 60Hz spikes without any logical source other than the mic) and I decided to buy a new one.

 

[Keith_W]

How to test if your mic has correct polarity or not:

Step 1: Take a speaker driver and a 1.5V battery. Connect the red input to + and black input to -. If the speaker pops out, it has correct polarity.
Step 2: Sweep the driver with your microphone and REW and examine the impulse response (or step response). If the deflection is positive, then your measurement system has correct polarity.

If it goes the wrong way, simply choose "Invert" in REW's preferences.

 

[sam_adams]

Test the entire signal chain for correct polarity using the attached file from my post over on AVS.

 

[FargateOne]

Test the entire signal chain for correct polarity using the attached file from my post over on AVS.

Given my relatively limited knowledge, I would need more details, please. Does the WAV file need to be read in REW to go to the AVR? and I want to check the microphone, what should I look for in REW?
I was not concerned about the polarity of the speakers until now, which have all been “positive” for 10 years or are you and @Keith_W saying that it might be possible that all the 5 satellites pop in instead of popping out with + battery connect to red input of each 5 satellites...