TOPIC:

I would like to correct recordings with UMIK-1 with the calibration file. Please read my approach below and correct me where are wrong or provide additional information.

Since I am only working in the frequency domain, I can do an FFT and use the calibration file as a multicplecation factor on the output of the FFT.

Polynomial interpolation on the calibration file is good enough to find the factors needed for teh frequencies found in the output of the FFT.

How do I extrapolation the calibration for 0 to 10.054 Hz and above 20016.816 Hz? Which additional user-name additional calibration files are out there (here) to map via their and mine minidsp calibration files an estimated extrapolation usable for my mic? I am just reading documentation for REW (no experience with that so far) and my speakers and workplace are not that good enoug, I think, to get an idea of what is going on beyond the provided calibration file.

I have extended the frequency response by creating a smooth model. For this I used two different polynomial fits for extrapolation before and after the specified calibration curve and and linear interpolation for values in the specified curve.

The result can be seen here i.imgur.com/QkiUVbG.png and is specific for my microphone but can be reused for other microphones. I have no idea how the curve is suppose to look from 5 Hz - 10 Hz, perhaps somebody that has seen it can confirm it. I based my approach on what I found when looking at a large number of calibration files, see i.imgur.com/Fxr72v4.png

@ Pander,

Just so I get it right, you're using say a recording software (e.g. Audacity) that basically doesn't allow you to get include the calibration file into the equation. So what you're doing is correct the recording afterward? Sounds neat..

I think that your linear interpolation sounds good. Seems like you already found the location of our calibration files but unfortunately we don't know how the mics will perform from 1~10Hz... I wonder if we can find that info later. Using some feedback from other UMIK-1 users could indeed be one good way to do that testing.

DevTeam

I am using sox and have already made one general patch. The next patch concerns using calibration file directly aften any FFT has been calculated. Could you tell me a bit more about the calibration file? I want to make sure I have everything correct. I have a very technical background but working with audio has been almost 15 years ago.

In the clibration file is the microphone response in terms of power attenuation in dB measured for certain frequencies? Correct?

What do I do with the sensibility factor?

Yes, pander, it's indeed power attenuation in dB.
The sensitivity factor is really here for SPL measurement/i.e. calibration of the capsule sensitivty. Not useful in your case I believe..

I am using the mircophone also for SPL. The system I have build is to analyse low frequency noise pollution. This a very undiscovered terrain. Environmental inpectors of my city government have measuring devices which are very limited, but calibatred. My system with th UMIK-1 is very accurate but uncalibrated, but with reference recordings with their devices present I am going to calibrate my system. However, I am looking for someone in the community here who would like to verify my approach.

If you would like to help me (find someone), please leave a message to do this outside of the forum. Once this has been solved, with very simple tooling on Linux you will get SPL measurements and automated plots.

@ Pander,

It indeed sounds interesting. Let's see how we can help. We'll contact you by email.

DevTeam