Approximating frequency response from a recording

[Justaguy]

Background: I bought some MS-TFB-2 in ear microphones so I could play around with measurements and more specifically, the Impulcifier program. One day, I was swapping pads on one of my headphones. I wasn't sure which pad I liked more, and since swapping between pads on that headphone was pretty time consuming, I thought to just record a bit of a song on those headphones with both sets of pads using my in ear microphones so I could get a rough "side-by-side" comparison. In my (lack of) infinite wisdom, I never saved full sweeps of those headphones before I sold them.

Question: Given recordings of a sample of music recorded with in-ear binaural microphones, could an EQ curve be extracted from these files so that a user's current pair of headphones would sound close to the headphones in the recording? Obviously, doing this activity with a sweep would be much easier, but I have since sold those headphones, and I didn't have the foresight to actually save the sweep files I recorded until now.

 

[staticV3]

Question: Given recordings of a sample of music recorded with in-ear binaural microphones, could an EQ curve be extracted from these files so that a user's current pair of headphones would sound close to the headphones in the recording?

Yes. Record the same track with the same in-ear mics, then load both recordings into Deltawave.
It'll analyze the files, match volume, phase, clock drift etc, then spit out a bunch of graphs showing the difference between the two, including the frequency response delta.

You can then take that data and either apply it directly as variable band GEQ in e.g. Equalizer APO, or you can use software like Room EQ Wizard or the https://autoeq.app/ web app to turn it into more conventional EQ presets or Convolution files if you'd like.