[markanini]

Haven't seen much talk about this, perhaps because some noteworthy implications weren\t explicitly stated.
http://www.innerfidelity.com/content/initial-results-head-acoustics-hats-measurements-harman
 
Especially interesting is the posted curve.
http://www.innerfidelity.com/images/MeasuringHeadAcoustics_FirstPass_Plot_TotalAverage2.jpg

It defines how a pair of headphone would measure specifically on Tylls measurment rig to match Harmans reference listening room. It also defines a target that can be used for EQing any headphone measured on Tylls rig. I tried that today with my Creative Aurvana Live! and I cant say I regret it, sounds like a not bad room for audio. 

 

 

[castleofargh]

the all series of posts about harman target and room measurement was pretty cool. to me there wasn't all that much to discuss because as far as frequency response goes, there weren't any big surprise. we end up pretty close to the harman preferred curve which wasn't all that far off from the old diffuse field curve for headphones(but far enough for me to clearly prefer harman's). a little more bass to compensate the lack of body vibrations(IMO that should be set by the listener based on how he feels about it). more or less of a tilt after 3khz, and the usual predictable variations between headphone and speakers(room, body, head...) where the averaged changes come to look more or less like the kind of curve he got.
 
you have also to consider that using crossfeed(I also have a hard time listening to headphones without some) changes the sound a lot subjectively and objectively, even the FR in most crossfeeds. so one's perception might change depending on how easily we can get fooled by a fake soundstage and how well a given crossfeed works with our own head. in the end the general model while very informative, can only get us so far and individual body and feelings are required to really fine tune it all.
 
also going from harman speakers in harman's room, to Tyll's dummy head, to one pair of headphone averaged at different positions on the dummy head, to our own pair that might not measure exactly the same and our head that might not seal or be as large as the dummy head, to apply an EQ based on that graph without a mean to control the resulting sound. that might not lead to an impressive level of precision down the road and might not be easy to implement for the average audiophile.
 
ok I'm saying it's great and it's not at the same time, but you get what I mean(hopefully ^_^).

 

[markanini]

Thanks for your input @castleofargh
 
Quote:

 
also going from harman speakers in harman's room, to Tyll's dummy head, to one pair of headphone averaged at different positions on the dummy head, to our own pair that might not measure exactly the same and our head that might not seal or be as large as the dummy head, to apply an EQ based on that graph without a mean to control the resulting sound. that might not lead to an impressive level of precision down the road and might not be easy to implement for the average audiophile.

Good points, you can indeed not expect a 1:1 result and just go EQ-ing away. For my EQ experiment I chose to use to my ears to decide: L/R source measurement blend, Y-axis correction threshold(no over-correcting deep nulls), gain scaling. A while back I did a similar experiment when I didnt have the Tyll/Harman HATS data. Then ~20% EQ correction gave the best subjective result. With the new data ~65%.

 
This has told me #1. There's always a place for using you ears to decide. #2. EQ-ing headphons is not easier with good data but more gratifying.