TOPIC: Translate 0-degree Calibration to 90-degree

It turns out that the delta between 0 degree and 90 degree measurements on UMIK-1s are extremely consistent, based on 10 measurement samples from Cross Spectrum Labs. So, I wrote a utility to translate the 0 degree calibration files from MiniDSP to 90 degree files.

If you need a 90 degree file to make measurements with the microphone pointed up, but only have the 0 degree individual calibration file from MiniDSP, give it a try (attached file).

If you haven't purchased the mic yet, it's worth it to spend the extra $20 to get an individual 90 degree measurement from Cross Spectrum Labs. If you already have the mic, this is a good alternative.

@ rcohen,

Very neat indeed and a big thanks for sharing with the community!
I'm guessing that you needed that file for Dirac software?

DevTeam

devteam wrote: Pretty much. I am using Dirac for a home theater setup, so I needed to do measurements with the 90 degree orientation. Dirac still worked with a 0 degree file, but it was showing a lot of treble rolloff. As you might expect, correcting the rolloff produced sound with overly boosted treble. So, I had to set the target curve so that Dirac wouldn't correct it to get good sound. It kind of defeats the point of using a calibrated mic, though. With a 90 degree file, I can start with a more accurate curve and see what I'm doing better when I set the target curve.

Thank you so much for posting this. Getting a umik-1 shipped from spectrum would cost me an extra $65 because of price and shipping differences. Nothankyou!

I'm wondering if we can make a 45 degree file also somehow? The reason is that in a home cinema set-up some speakers will be at a different height so placing the microphone at a 45 degree angle for all speakers is possible with just one mic position but placing it at 90 degrees isn't as the angle would hvd to be changed constantly. I'm not sure I'm making sense, I could try and draw a diagram but I'm not very good with drawing.

Anyhow, is it possible?

J19801110 wrote: Yes, if you have reference 0 degree and 45 degree calibration files from another UMIK-1, this utility should work very well to apply the difference to translate your 0 degree file to a 45 degree file.

I included 0 and 90 degree files averaged across 10 samples, but the difference turned out to be so consistent that using a single sample would have given good results.

So, you just need someone's UMIK-1 calibration file set. I don't have any 45 degree files. If you could post them here, perhaps they would be useful to someone else.

Sure, I do have a full set. I'll post it a bit later when I'm home.

Im wondering, how great is the difference between factory calibration and cross-spectrum calibration for the same umik (for 0 degrees that is). Did you have any data on that?

J19801110 wrote:
I also found one, if you need it.

They did appear to be different at the very top and bottom, although it was within the variation, and I didn't have a bunch of samples. No idea which one is more correct. At some point, how much treble or bass you like is a matter of taste, anyway (and the room and the speakers and the recording...)

How much of a differemce are we talking about? Below 1dB or over?

J19801110 wrote: Compare my last graph "my 0" vs "avg ref 0". My 0 is miniDSP (1 sample). Avg ref 0 is 10 Cross Spectrum samples.

If you move the my 0 graph up to overlap them better, it's about a 1 db difference, although it's impossible to say how much is a difference in calibration and how much is sample variation. It is smooth, though, and I tend the treble and bass more than 1 db to my taste, so I'm not worried about which is more correct.

Great. Thank you. Here are the files. Let me know if they're according to what you expected / in line with previous measurements.

Mmm... for some reason I can't attach files.
Third try...

Interesting to see that minidsp now offer a calculated 90 degree offset file themselves - for use with their Direc implementation. The cal file seems pretty similar to your generated one I did wonder if you'd had any luck generating a 45 degree version of your utility - my rear and surround speakers are ceiling mounted with the LCR being in front, so a 45 degree option would be best for me and sending my mic to CS isn't really an option as I live in the UK.

Thanks

pete S

psketch wrote:
That's cool that they started including 90 degree files. Based on the consistency of Cross Spectrum measurements, calculated files seem to be very accurate.

You can use my same utility to generate 45 degree files. You just need to get your hands on a single sample of 0 degree and 45 degree files from the same model mic. It will apply any offset, based on the files you feed it. Based on my tests, averaging 10 samples was unnecessary.

It looks like J19801110 provided some reference 0 and 45 degree files for a UMIK-1 a couple posts up. The narrow band 0 & 45 ones should work nicely, assuming you're using a UMIK-1.

how to implement that file u shared with my UMIK from minidsp?
sorry for stupid question, but i am newbie.. t

thanks so much for helping

@ hendra

Not sure to understand your question... You can get the 90deg directly from the UMIK-1 page. All explained in great details :
minidsp.com/products/acoustic-measurement/umik-1

Devteam