I figured you would know this better than I would. 🙂
For an absolute calibration you use a source with extremely harmonic low distortion such as an electrostatic actuator, pistonphone or (as IEC 60268 seems to imply) a very linear speaker at a low level. Another method specified by IEC (which only measures 2nd order distortion) is to measure two signals from two identical sources 80 Hz apart and compute the distortion that way.
I don't have any of those, so I'm going to assume that my ACO 7052 has essentially no distortion at relatively moderate levels (the mic is rated to 3% distortion at 150 dB), make a reference measurement with a source, use that distortion measurement as a baseline, measure the same source with the DUT, and then compute the difference. The problem is that I'm assuming you can't just subtract the difference, so my plan is to actually back out the raw 2nd/3rd/4th harmonic levels, compute the difference, and then recalculate the (hopefully) additional distortion level of the DUT.
Hopefully this will work.
Well that won't work. THD is not frequency response, comparisons like that aren't valid.
Here is a link of a EMM-13D082/H-P48/RM I have bought to do my speaker measurements if anyone is interested.
iSEMcon GmbH
Wolfgang seems very knowledageable.It will be interesting to see how he measures his mics etc
iSEMcon GmbH
Wolfgang seems very knowledageable.It will be interesting to see how he measures his mics etc
And I'm not that interested in THD anyway. Much more useful would be a tracking of harmonics along the frequency scale, such as HOLMImpulse or STEPS does. That would really tell us something.
But at least from the info linked to in this thread, I'm not too worried about distortion as long as the SPL at the capsule is <100dB.
Earl, I don't think you're giving Herb enough credit for knowing what he's doing. Forget THD and look at the individual harmonics. They are indeed a 'frequency response' of dB vs. freq. If you assume that the reference mic's distortion is very low, most of the distortion it measures is due to the speaker generating the signal. It's relatively simple math to get from there to how much extra distortion the mic under test is contributing. Maybe not a perfect NIST test but good enough for our purposes.
I think Earl was responding to panomaniac's comment about directly (implied) comparing the THD of the two mics, and he's right, you can't do that. Hence my plan to try to actually back out the raw harmonic levels and go from there. We'll see.
Earl is also right when he says that it's not a trivial task - B&K provides a lot of data with their individual mics, but the two things they don't provide are individualized absolute phase response and distortion figures, which isn't surprising given those are the two hardest things to actually measure.
Heh. I was actually thinking about this the other night - I just haven't had the time to pursue it. It's still on the list, I just don't know when I can get to it.
Idea for estimating THD at spot frequencies:
Measure a woofer at a specific SPL with a mic distance of 1 meter
keep halving the mic distance and record the spectrum. Eventually the spectrum will change and it won't be due to the speaker since it is playing at the same SPL. This should work OK at low frequencies, anyway...
Measure a woofer at a specific SPL with a mic distance of 1 meter
keep halving the mic distance and record the spectrum. Eventually the spectrum will change and it won't be due to the speaker since it is playing at the same SPL. This should work OK at low frequencies, anyway...
- Status
- Not open for further replies.