Finally, the ultimate frequency curve for a neutral headphone!

New Headphone Response Repository

AS stated in this AES paper from Harman: http://www.aes.org/e-lib/browse.cfm?elib=16768  the current target curves don't line up very well with what we hear from our stereo.  I have been trying with great pain to get sound equal to my stereo through headphones.  It has essentially been fruitless up until now.

I took a set of binaural mics from Sound Professionals, placed them in my ears, sat in the sweet spot of my calibrated home theater and took some measurements--lots of measurements.  My left ear curve looked like this:

  Another round with my right ear (RED):

  Then averaged with Audyssey's Dynamic Volume in use(GREEN):

Things of note about These graphs: below about 1000Hz, the room is in total control, above there it's the head and shoulders position.  You can see that little transition on the graph.  That just happens to be the wavelength of the width of the head(approx 2000Hz).  You shouldn't want to see any of those little wiggles below 2000 Hz and definitely not the 220Hz dip.

So what headphones did I have that measure like this on my head?

NONE.

I was able to get my grubby hands on a set of Beyerdynamic DT235 which sounded a little dark to me and I took and further modified my German Maestro GMP160 to more closely line up with my curve.  Then the Sennheiser HD598 which probably sounds the most correct to me. Here's what I got:

If you smoothed these out a bit:

I'll update this with more headphones and IEMs before long.