[KlarkKentThe3rd]

I am not exactly sure how old this thread is, but I'll bump it anyway.
 
There is something I, as a n00b, do not understand. How reliable are those graphs anyway? I mean, really? How are they made, and who makes them?
 
 
Also, I have figured out two ways to measure neutrality:
 
1) Relative neutrality. Pick a well recorded and well mastered track that has the full spectrum of frequencies. Listen to it through your test (°_°)phones. If some of the instruments (at the lower and higher ends) are difficult to pick out, your pair is not relatively neutral.
 
2) Almost absolute neutrality. Make a high quality binaural recording of your (°_°)phones. Then, listen to the source file and your recording back to back, through a good pair (can be the same one). If the altered!original and altered!altered files sound almost indistinguishable, then your test pair is pretty much as neutral as it gets.

 

[AzraelDarkangel]

On goldenears graph and waterfall, the HD600 is the flatest I've ever seen:
http://en.goldenears.net/8072

 

[Currawong]

I split your thread necro bump off into a new thread for you instead.

 

[ieee754]

On goldenears graph and waterfall, the HD600 is the flatest I've ever seen:
http://en.goldenears.net/8072

I think that HiFiMAN headphones containing '5' and/or '6' in their names are the flattest measuring headphones around (especially in the subbass region).
 
http://www.innerfidelity.com/images/SennheiserHD600.pdf
http://www.innerfidelity.com/images/HiFiMANHE500.pdf
http://www.innerfidelity.com/images/HiFiMANHE6.pdf
 
And probably HE-560, but there are no measurements yet. Also HE-6 have the cleanest waterfall graph I have ever seen, but I can't post it here.

 

[ieee754]

  I am not exactly sure how old this thread is, but I'll bump it anyway.
 
There is something I, as a n00b, do not understand. How reliable are those graphs anyway? I mean, really? How are they made, and who makes them?
 
 
Also, I have figured out two ways to measure neutrality:
 
1) Relative neutrality. Pick a well recorded and well mastered track that has the full spectrum of frequencies. Listen to it through your test (°_°)phones. If some of the instruments (at the lower and higher ends) are difficult to pick out, your pair is not relatively neutral.
 
2) Almost absolute neutrality. Make a high quality binaural recording of your (°_°)phones. Then, listen to the source file and your recording back to back, through a good pair (can be the same one). If the altered!original and altered!altered files sound almost indistinguishable, then your test pair is pretty much as neutral as it gets.

I believe that graphs are extremely reliable and that FR graph can tell you 90% of the headphone sound quality, if you can read them of course. However when you compare two headphone graphs you must remember that they should be measured in identical environment, using the same compensation curve, etc.
 
Your ideas for subjective measurements sound pretty good, I think that's the way most people evaluate headphones using ears, either by how they feel it (most subjective way) or by comparing them to the reference pair.

 

[castleofargh]

 

  I am not exactly sure how old this thread is, but I'll bump it anyway.
 
There is something I, as a n00b, do not understand. How reliable are those graphs anyway? I mean, really? How are they made, and who makes them?
 
 
Also, I have figured out two ways to measure neutrality:
 
1) Relative neutrality. Pick a well recorded and well mastered track that has the full spectrum of frequencies. Listen to it through your test (°_°)phones. If some of the instruments (at the lower and higher ends) are difficult to pick out, your pair is not relatively neutral.
 
2) Almost absolute neutrality. Make a high quality binaural recording of your (°_°)phones. Then, listen to the source file and your recording back to back, through a good pair (can be the same one). If the altered!original and altered!altered files sound almost indistinguishable, then your test pair is pretty much as neutral as it gets.

I believe that graphs are extremely reliable and that FR graph can tell you 90% of the headphone sound quality, if you can read them of course. However when you compare two headphone graphs you must remember that they should be measured in identical environment, using the same compensation curve, etc.

+1
graphs should be used to compare 2headphones, one you own, one you want. and measurements should indeed come from the same source else it really doesn't mean anything.
I got used to Tyll's raw measurements, maybe I would check at the compensated values with some olive&welti eye candy because that compensation is a lot more "in phase" with what I really hear. but yeah anyway I'm now used to work with innerfidelity's raw data so as long as Tyll doesn't change something in his equipment, it's an amazingly reliable source of data for me.
golden ear compensated graphs make me cry a little, they use a detail exterminator, the results are so smooth that they would hide mountains and fault lines on a world map.

 

[bigshot]

The thing is, those smoothed out graphs are probably a lot more representative of what your ears actually hear. I'd actually rather see graphs corrected for human hearing, with Fletcher Munson and a bit of smoothing to reflect the actual sensitivity of the ear.