Originally Posted by slosvt
My take on the cell phone calibration is that the software "calibrates" the near field measurements as "reference" and makes EQ adjustments based on the measurements from the MLP. This is very much like profiling a cheaper/quicker colorimeter (i1 Display Pro) to a more expensive/slower spectrometer (i1 Pro) for display calibration.
The take away is not that the "field" device is accurate, it is that the measurements are repeatable and offset by a "reference" device. Based on the use, the phone's mic should be more than adequate for the EQ calibration generated in the app *if* the measurements are repeatable.
No where have I seen Mark say the phone's mic is equal to that of the UMIK-1. However, I do feel that Mark was comfortable with the results when he did an A/B comparison.
Edit... Now I see the quote above from Andrew Jones explaining what I tried to in more detail...
Let me emphasize again the process I am using for the "calibration" process.
When we think of calibration, we generally assume that we have one device that is perfectly known and is used to then measure another device. In this case one would assume a perfectly known microphone being used to measure the response of a subwoofer.
However, lets stop and think for a moment: How do we measure a microphone? we need a known sound source. But how do we know what the sound source does? we need a known microphone!!!! Suffice to say there are ways using the reciprocity theorem that get us around this conundrum.
So, knowing this, how do we use an unknown microphone to "calibrate" an unknown source?
First, we don't actually calibrate anything! Not the speaker, not the microphone, nothing!
Second, we don't make a measurement that tells us what the speaker response actually is!
What we do is assume that the response that was engineered into the subwoofer by the designer (in this case....Me :-) ), is the response that you want to hear, unencumbered by the effects of the room.
Given this, all we have to do is make two measurements, one close to the subwoofer where the room has almost zero effect, and one at the listening position. Now we simply apply EQ until the second response curve matches the first. Bingo.... no need to know what that curve should actually look like, no need to know what the microphone does.
The mic has to be able to respond with some sufficient signal to noise ratio down to the lowest frequencies.
The nearfield level must be such that the mic does not overload but high enough that we get sufficient s/n ratio.
This does not make the subwoofer "better" than the designer intended. It does not apply EQ to extend the response to be flatter and wider than the designer intended. It just lets you hear what the designer wanted you to hear. If that's not the sound you like, you should pick a different designer!! ;-)
If you still feel that you should be using a "better" microphone, then go and buy the Dayton plug in mic available from parts express. It will plug directly into the 3.5mm socket on your phone. It may make you feel better about the process.
I hope this explanation clarifies our approach to calibration.
I'm sure I'll hear from all of you if it doesn't! :-)