I already had been using a UMIK-1 for about 3 years to measure my system, and actually have been pretty happy with it. But because I'm an engineer, and because there's really no way of knowing whether the calibration of a mic is correct without some kind of reference, I decided to buy a Dayton UMM-6 to compare against it and see whether I could learn anything.
The first UMM-6 was DOA, so I sent it back. The second one worked fine, so I set up a comparison. I actually measured the UMM-6, then the UMIK-1, and then the UMM-6 again. I used my system in the basement as a reference, using PR pink noise; I know the actual spectrum isn't totally flat, but the object here was to compare the mic curves, and not to do an absolute calibration on the mics themselves. I quit REW between each measurement, in order to minimize the possibility that it might hang on to the wrong cal file and bias the measurement. I used the moving mic technique with averaging (something I've used a lot over the last couple of years) to minimize position errors and got very consistent results. I actually measured the UMM-6 at both 0° and 90° to see if there was a significant difference caused by orientation; I used one of each orientation in my graph since otherwise the two UMM-6 traces would be indistinguishable - they were that close to each other from run to run.
The result: there's a rather large discrepancy between the mics, which is greatest around 10 kHz. Blue trace = UMIK-1; other traces = UMM-6. Both of the cal files look reasonable (and yes, the cal files from the two UMM-6 mics were actually different, so I don't think they're cheating by sending the same cal file with each mic) and show a max difference of ~2 dB - nothing like the 10 dB I see (which is a sanity check on whether the cal files actually were applied correctly). Maybe somehow there is an error due to the wrong cal file being selected, but that doesn't seem consistent with the results as I view them.
Tonight I'm going to insert a graphic EQ into my system and attempt to verify that the UMIK-1 trace is the more accurate - and if it isn't, I'll be sure to note that. I have no bias toward one or the other of these mics, but I'd sure like to know whether my ears agree. The system sounds OK now, but I recognize that I may just have gotten used to the error and am accepting it as "right". Like the man who has two clocks, I am now worried that I don't have the correct time.
If anyone sees a glaring error here, please chime in.
The first UMM-6 was DOA, so I sent it back. The second one worked fine, so I set up a comparison. I actually measured the UMM-6, then the UMIK-1, and then the UMM-6 again. I used my system in the basement as a reference, using PR pink noise; I know the actual spectrum isn't totally flat, but the object here was to compare the mic curves, and not to do an absolute calibration on the mics themselves. I quit REW between each measurement, in order to minimize the possibility that it might hang on to the wrong cal file and bias the measurement. I used the moving mic technique with averaging (something I've used a lot over the last couple of years) to minimize position errors and got very consistent results. I actually measured the UMM-6 at both 0° and 90° to see if there was a significant difference caused by orientation; I used one of each orientation in my graph since otherwise the two UMM-6 traces would be indistinguishable - they were that close to each other from run to run.
The result: there's a rather large discrepancy between the mics, which is greatest around 10 kHz. Blue trace = UMIK-1; other traces = UMM-6. Both of the cal files look reasonable (and yes, the cal files from the two UMM-6 mics were actually different, so I don't think they're cheating by sending the same cal file with each mic) and show a max difference of ~2 dB - nothing like the 10 dB I see (which is a sanity check on whether the cal files actually were applied correctly). Maybe somehow there is an error due to the wrong cal file being selected, but that doesn't seem consistent with the results as I view them.
Tonight I'm going to insert a graphic EQ into my system and attempt to verify that the UMIK-1 trace is the more accurate - and if it isn't, I'll be sure to note that. I have no bias toward one or the other of these mics, but I'd sure like to know whether my ears agree. The system sounds OK now, but I recognize that I may just have gotten used to the error and am accepting it as "right". Like the man who has two clocks, I am now worried that I don't have the correct time.
If anyone sees a glaring error here, please chime in.
It has done a worthy job of making my system sound right with most music.
