Trace arithmetic working correctly?

[David Eisan]

I have both an Omnimic and an Earthworks M23.


Here are two readings of the same speaker, One with Omnimic (dark green) and the other with the Earthworks M23 (purple).

I was hoping to subtract one from the other than add (or subtract it, not sure yet) it to the Omnimic Calibration file so the Omnimic was Calibrated to the Earthworks M23.

However, A-B does not work as I expected.

Instead of the difference between the two graphs producing a graph in the +/-several dB showing the actual difference between the two graphs, I have something else. I actually don't know what I have (light green).

I time aligned both graphs before doing the math

(This was just a test, I know I need nearfield data properly merged with both to do this correctly)

 

[John Mulcahy]

You should divide the traces, Omnimic/Earthworks.

 

[David Eisan]

I tried dividing the traces, and I didn't get the difference between them.

I lowered the DB on both mics so they are both on the scale.

 

[John Mulcahy]

Post the mdat file. If you need 1/6 smoothing to see clean traces you probably don't have measurements that are suitable, with too much room contribution. Better making nearfield measurements of a tweeter and a woofer with each to see the mic responses with much less room contribution.

 

[David Eisan]

It looks like it worked this time. I windowed the measurements to 4ms. I am only looking to change the data >300Hz.

Thank you,

David.