• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Is it possible to flatten out a non measurement microphone?

Volutrik

Member
Joined
Apr 2, 2023
Messages
49
Likes
20
I was wondering if one could flatten out a microphone by inverting the frequency response found in a graph. I would like to get a taste of the result without having to buy an expensive microphone for this... There's a Superlux microphone, which is pretty cheap, but requires an audio interface because of the XLR input. There' a website called Audio Test Kitchen, which contains a database of FR's for a lot of microphones (please let me know if there's a bigger database out there). I know that the answear will most probably be that this method isn't ideal as using a proper measurement microphone, but I'd like to get some brief explanation of why that is. Also, even if I do this for testing purposes only, would the result differ so much from the one I would get with proper measurement microphone that it would be 100% far off?
 
Yes, if you trust the curve. That's how measurement mics work... They aren't necessarily flat but they come with a calibration file. Usually they are individually calibrated with the calibration file tied to the serial number. But hopefully there isn't that much variation with the same model.

The microphone should usually be omnidirectional.
 
I have heard of some people reporting they have inverted a mic's published graph and gotten pretty flat. But these were for $400+ mics, where the graphs are ilkely pretty good. Also, you need to know if your software wants an inverted or non-inverted FR curve (i.e. some software is subtractive and some is additive).

Some manufacturers will test many mics and report an averaged FR response for the mic. If they have good QC, then the sample-to-sample variation is not too bad and inverting will likely be close. A real measurement mic is flat by design. Or a real corrected mic, like the UMIC or MA-1, has its own unique correction file.

I am not sure about the audio test kitchen's graphs. Are they single sample or averaged over many mics?

There are a couple normal mics, like the Shure SM81 that's pretty flat. But I think for room correction, you want an omni with even polar pattern.. There's another thread on ASR about that

The dBx RTA-M is only $125. https://dbxpro.com/en/products/rta-m

I think some people use the inexpensive Behringer ECM8000 for room measurements too.

The UMIC-1 measurement mic is $79 new. I'm sure you could get those used for cheap. This is a USB mic, which might or might not work for your application.
 
... that this method isn't ideal as using a proper measurement microphone, but I'd like to get some brief explanation of why that is.
Dynamic, condensor, ribbon are the main types of mikes. The directivity can show different patterns, which also alter the frequency response, that along the pattern even might differ with distance.

On Linkwitz' site you can find some hints on how to use a 1,50$ Panasonic electret/condensor capsule as a quite satisfying measurement mike. Even the standard connections work perfectly fine. Believe it or not, I recorded supersonic (sound, not speed) bats in my backyard with it--40kHz! The Panasonic is said to be out of production, but the dozen I ordered were at least fully applicable replica.

As it comes to compensation, I didn't care so far. The overall balance, of the Panasonic, due to its tiny size, is perfect, despite of my hefty feedback'ed application. Further more I reproduced the speaker chassis manufacturer's frequency responses so often with many different specimen, that I trust my ears for the last dB or so anyway.

Hope this helps ;-) Very good move to start measuring!

As to answer your question directly. You surely (pun intended) cannot use a Shure vocal mike for measurement. Please specify.
 
1. If the mic puts out nothing at some freq, boosting that freq gives you only noise.
2. If you intend to use it for measuring room+speaker response, the important freqs are 20-700 Hz. If the mic has output at these freqs, the idea might be valid.
3. Omni small diaphrgm mics are the best candidates for this idea.
 
These guys list almost all microphones with basic specs and usually maker supplied response graphs.

I've had very good luck doing what you have in mind. Depsite what is often claimed, the published graphs must be pretty good as I get excellent agreement with my Umik 1. Web Plot Digitizer is an online bit of software that can take a graph and convert it to a CSV file which REW can use. Watch the video instructions and it will make sense.

Now if it is a ribbon or cardioid you'll never get the low end correct. From say 300 hz and up they can do okay. Best is an omni. Better still if it is a small diaphragm omni.

OTOH, unless that Superlux is dirt cheap and an omni, is it worth all this instead of the Umik 1 or Dayton Audio UMM 6 usb microphones. These are convenient, REW can use them easily and you don't need an audio interface just a PC or Mac with a USB port.

Here is a comparison of the Umik 1 vs an Avantone CK1 with omni capsule. The CK1 (in green) is a pencil condenser costing $150 so no savings vs the Umik, but I had one to compare. This is with a cal file from the published CK1 response. Without it there is a hump in the 4- 8 khz region and a dip in the 800-1200 hz region. As you can see it rolls off above 15 khz even with a calibration file. I obtained this by measuring a speaker with each microphone in the same location 1.5 meters from the speaker. The speaker is the source of the up and down response. I would consider that quite usable a result with either.


1680515579658.png


Here is the same comparison with a CAD M179 microphone (in green). An LDC multi-pattern microphone in this case set to omni.

1680515533368.png
 

Attachments

  • 1680462988847.png
    1680462988847.png
    72.1 KB · Views: 85
Last edited:
Back
Top Bottom