Why I HATE USB Measurement Microphones 

Awesome, I'd love to see a practical example of the transfer function method. cheers.

Closing off the port changes the tuning of the box and compliance of the driver. Splicing or normalizing based off this will give incorrect results. Isolating the port with a baffle may be a better option, if I’m understanding your method correctly.

@@jtmcfarland3512 The port needs to be sealed off for a good close-mike measurement. The only purpose of this part of the process is to derive the transfer function, so it doesn't matter if the tuning of the box is changed. Sometimes I use an entirely different speaker of approximately the same size to derive the transfer function if it's not possible to seal the vent of the speaker that I want to measure (e.g. if the speaker is a passive radiator, rather than a vented, alignment.). The transfer function basically describes the change between the measurement with the mic at point A and then at point B, so if you do any subsequent measurement at point B, you can calculate what the measurement would be at point A by subtracting the transfer function.

@@Grommet2007 I understand the second paragraph, but I’m not sure I agree with the first. My amateur experience in WinISD suggests the “tuning” of the box (size, order, compliance, damping material, etc) directly effects the transfer function. The ideal way to measure would be to have the port on the other side of an IB (simulated obviously) from the driver in order to get near field measurement of the transfer function without the port & vice versa. With a large enough baffle, you should be able to get an accurate gated response of each. Then you can do whatever comparison you like. However, each time you change the port dynamics and certainly the driver itself, you’ve changed the transfer function of the system and will need to take new measurements at BOTH locations.

@@jtmcfarland3512 Nope - the transfer function is simply the difference between the measurements at point A and point B, due to the delta difference from the device being measured and whatever impact the environment has on the measurement. You're talking about the transfer function of the system - I'm talking about the transfer function of the measurement environment :-).

😢

If you tick "Adjust clock with acoustic reference" option in preferences - a recent addition to REW - it will compensate for clock differences between the input and output by adding a reference timing signal at the end of the sweep. This will take care of "most" of the clocking deviations inherent to USB mics.

yes, a valid point.. in 'uncontrolled conditions' at such high frequencies the wavelengths are so small that everything counts, from exact positioning of the equipment to even the position of the person doing the measurements.. that being said, for all practical reasons super accurate measurements of the high fqs are not needed since healthy ppl over the age of adolescence can't hear much of the second half of last octave plus there's v little acoustic information there anyways :)

This is what I'm planning on doing.

so you would say the umk1 would be good beginner one for measure my hifi speakers and headphones or is the cheap Behringer ECM8000 enough? hm but then I need a sound interface and a notebook in my living room

You need a new calibration file, not a new mic. Although, I don’t know any company that calibrates mics as a service.

@@AudioJudgement Awesome video besides the matter of Interface selection. Isn't that most important specifically when going for XLR Microphone and the one that does 5hz-30khz. Because if the frequency response of Interface not being flat would negate the benefit of XLR Mic being absolutely flat in response. Please look into Audio Interfaces and make video on that as well. Which is absolutely flat and is the best.

@@KING_DRANZERYou have to calibrate the interface separately. Just use the loop cable and recalibrate as needed. No need to have a perfectly flat interface. I’d be more concerned with other features.

You don't need to do that. They're just trying to make you a "subscription" customer.

If the port is on the rear, it certainly seems he should be adding an offset. I’m a little confused about this method.

Room EQ Wizard

Hello! It will be one large course to replace 4 of my old courses. It will be finished in 6-12 months

Interesting. It seems that's a generic calibration file. One that fixes some major response flaws that every behringer mic has. However, a true calibration file is made specifically for your mic. And you need to download a calibration file based on your mic's serial number.

Yes, the udemy course needs an update, as there is no mention of the loop-back. All the measurements in the course are valid with the single exception of nearfield measurements of ported enclosures (where a loopback is necessary). I'm currently working on a new updated course (complete overhaul), but it will take several months until it's finished.

This is great info. It seems there’s not only a delay in the port resonance (which seems to be solved by the loopback method), but there’s also an acoustical offset as well as unique baffle diffraction which will change depending on the location of the port relative to the driver. 😅

Couldn't find information regarding umik-2 frequency response. Maybe send minidsp an e-mail. Let us know as well if you find out.

@@AudioJudgement The email response was this: “UMIK-2 calibrated response is 10Hz to 20kHz.” (Link to owners manual that I am omitting in this quote) “please read for more details. Maybe the next step would be to read a bit details from the manuals/product datasheet so you can learn a bit more about our products? We'd be happy to help answer any questions you may have.” “Hoping this info helps and feel free to contact us if you have further questions.” So, they are saying 10 HZ to 20 KHZ, but I see no documentation of that in the user manual or any of the links or charts anywhere in their documentation. By pointing me to the user manual one might expect to see those numbers listed somewhere or perhaps a chart in REW but I had already read that and the data sheet before emailing. This time I checked the link on Home Theater Calibration, found in the user manual, and read through the docs at that link. Still, I would like to see it in print but they claim in their email it is calibrated to measure down to 10 HZ.

It's most likely true. When I googled for this spec I saw a graph comparing the response of UMIK-1 and UMIK-2. The range of 10-20 Hz of the UMIK-1 rolls off quite significantly compared to UMIK-2. My guess is that UMIK-2 does have an extended frequency response and you should buy it if you want an affordable mic that can measure below 20 Hz.

@@williamhicks2763Open the calibration file as text. Please let us know what the min and max frequencies are. I assume there are some zeros and not only values for non-zero offsets. Idk.

@@jtmcfarland3512 I will when I get it. I saw I could save a few dollars by ordering directly from miniDSP and then realized it was being shipped from Hong Kong. I guess that I can let you know if it ever gets here. Three days after the order the status only says “Confirmed.”

That seems like a VERY high noise floor. It should be around -95 dB. Are you sure it wasn’t the setup or cables or something? I do agree that if you’re using multiple microphones at once they all need minimal variation if you’re only using one calibration file. For most people and many companies, a single calibrated mic even with

@@jtmcfarland3512 Hi. I tried every microphone with the same cable and same measurement program. REW. I’ve also used 3 different calibration file for each microphone. While playing constant noise, SPL values I measured was around 65 dB for 1 microphone, in the other It was 66.5 and other was 63.5. I corrected the other microphone by changing the calibration gile gains and eqs myself. There was a little bit bigger difference in the Noise floor.

@@AliBekirKlckaya That’s interesting. I don’t remember REW having an option to calibrate SPL separately for each input. Separately adjusting gains to match for each input device would be necessary and is recommended by multiple manufacturers. For the types of uses in this RU-vid channel, we’re generally not concerned with absolute dB and really only need relative dB data. Likewise, noise floor really isn’t an issue for our purposes either as long as it’s not ridiculously high. Our measurements are ideally taken at 80 dB, so even a 30 dB noise floor is easily filtered out since dB is not a linear scale. YMMV

@@jtmcfarland3512 I calibrated each calibration file by using one at a time, from the same position. It should be noted that as the noise floor changes, the distortion graph in the lowest levels changes as well. Its like you cant measure below 0.2 percent distortion because of the higher noise floor in one microphone compared to other. Yes, dB relative data would be unfeffected. However, I’ve noticed frequency response difference up to 3 dB between microphones. Especially around higher frequencies, such as 9,-15kHz. I can provide the calibration file number of each microphone. .UMM 6 microphones are good for average consumers. But I’m a loudspeaker tuning guy, and tuning 3dB off will sometimes makes or breaks the speaker in the eyes of reviewers nowadays..

@@AliBekirKlckaya You’re clearly working on higher accuracy stuff than me, but my noise floor is currently around -35 dB (not great, but usable). With measurements at 80 dB that’s a 115 dB difference. That means even a 3 dB variation in my noise floor only shows up as a teeny tiny fraction of a dB at my measurement level. (110 is 10^11 times louder than the floor) 3 dB from 0-3 is not the same as 3 dB from 100-103. Very different.

I do have in plan to make a complete overhaul of my courses, But it will take some time.

👍👍👍

Like you made zig for measuring T/S parameters using sound card and REW.

Makes sense. He seems to imply that doesn’t work by suggesting an audio interface is better. I’d be curious to know more too.