[xnor]

1) Well, what about using the Olive/Welti target curve and calculating the "distance" from it for each headphone?
 
Something like the HD800 would never score an A for me personally due to the treble peaks, but with some EQing would probably be an A-rated dynamic headphone due to the low distortion, extension at the frequency extremes ...
 
2) I would do the same with distortion, like 3% at 20 Hz wouldn't matter much, but 3% at 1 kHz would result in a severe score penalty.
 
3) Think about the distortion numbers in general.
 
4) Smoothness of frequency response should be calculated like 1) from raw data.
 
5) A dip at ~100 Hz may only be cosmetic if distortion is still low. Easy to EQ. So the score penalty should be a lot smaller.
 
6) Who cares about the impedance curves?
 
7) Channel balance: again, scored by calculating it over the whole frequency range.
 
8) (Semi)open: shouldn't matter at all. Different strokes for different folks or applications.
 
9) Treat every headphone equally. Ignore price and marketing. If a 100$ headphone beats a 1400$ one so be it.
 
 
Score for each category could be from 0.0 (complete and utter failure) to 1.0. Different categories may need different weighting for the final summed score.

 

[anetode]

   
Frequency response analysis is hard. So much research is still going into it that I really don't feel safe making a strong comparison against any model yet. We have the Lorho modified DF, we have the Olive-Welti curve, there's an analysis for 1995 that proposed another curve, Sennheiser seems to have their own curve that almost all of their headphones match, as does Beyerdynamic, etc, etc. I tried to avoid that problem by trying to analyze the FR as minimally as I could, even though it probably is the most important metric of them all here, for this reason.

How about judging response below 1khz equally in all phones and then considering the deviation from their ideal curve as stated by manufacturer.
 

 
Judge how far the headphone deviates from that initial hump and how smooth the transition to the dip at low treble appears. If there are any out of place peaks or sharp slope changes then deduct points, or something like that
 
 
Quote:

   
8. I agree that we need to have concrete measurements on this one. Anetode's suggestion of 1dB in bass and mid-range and 3dB in the treble seem like reasonable ballpark figures. I think we need to be a bit more exacting,though. I think we encounter measurement difficulties past 10khz, so we should probably only look up to 10khz. I feel like some of the IF measurements suffer from a poor seal, which also makes those ones hard to compare against. I also feel like this one SHOULD have unit-to-unit variation taken into account, but the data doesn't exist yet for most headphones.

 
To clarify I meant 3db past 10khz, though I agree that it might be wise to limit this to below this threshold since sensitivity to level differences is lower at high treble frequencies. Another approach might be to determine a metric for a stable center image. Also the low bass response will depend on fit, so a difference there might be accounted for by earpad wear, positioning on the user's head and clamping force.

 

[anetode]

   
9. Someone raised a good point about how a closed headphone shouldn't be penalized for being closed. There are closed headphones which can sound better than open headphones, after all. I can't think an elegant way to credit a closed headphone for that, though, and I don't personally think it exists at the flagship level yet. Does anyone have any ideas on this one?

 
Sony R10, Fostex TH900, Audio Technica W3000 (& others, like the "leatherheads"), LCD XC, Ultrasone Signature Pro (the least sucky of the Ultrasone line). Of course not all of these headphones would score well on your metrics, but I'm fairly sure that the R10 would satisfy all criteria.

 

[erikfreedom]

I have never had any treble peaks problems with my hd 800 and they are not modded. I have added a moon audio deluxe black dragon cable and I match them with a yummy class a solid state amp. the hifiman ef-6.
 
huge bass, huge soundstage and great clarity without any treble peaks.
 
hd 800 has no problems. hd 800 has problems when you stick with stock cable and match it with amp that is bright as hell.
 
go for premium copper cables and warm class a tube or solid state amps and you are good to go.
 
hd 800 need love and warmth to sound their best. and when they do, they rule.

 

[anetode]

 
hd 800 need love and warmth to sound their best. and when they do, they rule.

I've been beating up and tossing around my HD800s in drunken spousal rages. I keep getting warned that if I continue such abuse the state will take custody of my HD650s. Still sometimes I lock the HD800 out in the cold St. Paul winter, just to teach her a lesson. No wonder she's shrill.
 
And don't even get me started on yummy amps

 

[xnor]

  I have never had any treble peaks problems with my hd 800 and they are not modded. I have added a moon audio deluxe black dragon cable and I match them with a yummy class a solid state amp. the hifiman ef-6.
 
huge bass, huge soundstage and great clarity without any treble peaks.
 
hd 800 has no problems. hd 800 has problems when you stick with stock cable and match it with amp that is bright as hell.
 
go for premium copper cables and warm class a tube or solid state amps and you are good to go.
 
hd 800 need love and warmth to sound their best. and when they do, they rule.

 
I'm sorry but this doesn't make any sense to me.
 
Cable: A high resistance cable will increase output impedance and therefore increase bass a bit with dynamics like the HD800. The parameters of cables are well known, especially at audio frequencies.
Amp: Usually doesn't change the frequency response of the headphone, except for high output impedance. Topology doesn't matter.
  Non-broken amps have a ruler flat frequency response in the audible range. EF-6 has ruler flat FR.
 
I don't think the EF-6 or other amps with low-distortion, flat FR amps are "warm" etc. I never got subjective reviews. In one review they say the EF-6 is "natural, transparent, quite detailed", in the next one they say "warm, laid-back, less detailed", in the next one they say "shimmering and velvety like nacre" (I'm not kidding, that's from audio/stereoplay.de).
 
People seem to hear whatever they like to hear, which is why some more objectivity for headphones (especially flagships) is badly needed.

 

[Sonido]

How would you guys rate the Bowers & Wilkins P7, the B&W current flagship? I attempted to rate it on the the same scale, but I was not too sure on a couple, especially #4 and #5. Here's my assessment: http://www.head-fi.org/t/681976/bowers-wilkins-p7-over-the-ear-headphones/720#post_10099920
 
I gave it a 6/9. For a closed headphone that's not bad.
 
I think the OP of the reddit post is better qualified for rating for consistency purposes.

 

[Gowry]

I wonder how the LCD-X would fit in this lineup. One thing that gave me pause when ordering them instead of the LCD-3 was that the distortion looked to be about double that of the LCD-3. Note that double very little is still very little, though. Where does distortion become audible? I'm used to the HE-500 and wanted to keep more treble. While the LCD-X are still darker, from the graphs I can see that they have more treble than the LCD-3.

 

[Clemmaster]

You won't get the HE-500 treble with the LCD-X, unfortunately.
There's some elevation in the treble region of the FR but they are not in the good areas IMO. They make the hiss on songs noticeable but don't convey enough openness.
Hifimen do both much better (IMO, again)

 

[Gowry]

You won't get the HE-500 treble with the LCD-X, unfortunately.
There's some elevation in the treble region of the FR but they are not in the good areas IMO. They make the hiss on songs noticeable but don't convey enough openness.
Hifimen do both much better (IMO, again)

I have both and wholeheartedly agree. The HE-500 seems more open and actually seem more realistic because of that. I'm still getting used to the audeze sound, though. 

 

[gatucho]

I see a lot of talk regarding FR design goals. Nonetheless it seems that at this point defining a single FR profile as THE ONE is just silly, considering that there are a LOT suggested by different sources.


Precise measurements are only half of the equation, the author of the original articles points himself that it is possible to predict whether a speaker will be PERCIBED to sound good or not using  a set of measurements similar to those he proposed. So PERCEPTION is still at the center of the issue. We would have to agree that different manufacturers have different ideas of which FR is ideal.


At this point measurements are only useful to predict the character of a given HP. Under this consideration and assuming that each FLAGSHIP closely represents the corresponding manufacturer FR ideal, well performed measurements only allow an insight of the characteristic Sennheisser, et Al sound.



An example of how, ironically, subjective this FR measurement comparison can be is the FR of the hd800 vs the t1. I will say it again, since it seems that the t1 is not very liked around here, comparing the RAW uncompensated FR responds of these HP (see headroom measurements) THE HD800 HAS MORE RESONANCES. In this regard the author proposes that resonances are indicative of a poorly designed diaphragm. So, is the hd800 more poorly designed than the t1?


This is an example on how objective measurements can be used to support ANY argument if a subjective interpretation of such measurements is used. The sciency folks around here dont seem to understand this simple fact (a thing which I know, suffer and exploit for a living)

 

[castleofargh]

the point is to say that a "flagship" should at least be technically as good as an ok cheap headphone. I can see no rational way of justifying bad technical specs on what is supposed to be elite stuff.

 

[Sorrodje]

   I can see no rational way of justifying bad technical specs on what is supposed to be elite stuff.

 
because Maybe Bad specs sound sometimes better/more enjoyable ?
 
I read the complete article and it's interesting but I would really like to know what's the relation between " good specs" and "good sound".. No Troll in my mind here but a true curiosity.  

 

[xnor]

  I see a lot of talk regarding FR design goals. Nonetheless it seems that at this point defining a single FR profile as THE ONE is just silly, considering that there are a LOT suggested by different sources.

Not exactly on your point, but let me first say that FR is *the* most important measurement. Yes, you could deliberately manufacture an audio component with nice enough FR that is screwed up in other areas but in practice those are rare cases.
 
I don't see why you wouldn't want one target FR. High fidelity means accurate reproduction, so we have one target FR for speakers (flat). The user or manufacturer may add a smooth downwards tilt of the FR curve or even a loudness function, but that doesn't change the flat target curve.
Why should it be different for headphones? Research by Harman and the NRC tries to come up with a better target for headphones that takes human perception into account. The other old curves were based on different sound fields, not perception.
 
 

Precise measurements are only half of the equation, the author of the original articles points himself that it is possible to predict whether a speaker will be PERCIBED to sound good or not using  a set of measurements similar to those he proposed. So PERCEPTION is still at the center of the issue. We would have to agree that different manufacturers have different ideas of which FR is ideal.

Setting a better target curve does not mean manufacturers cannot produce different sounding headphones anymore. Something like +/- 3 dB leaves *a lot* of different sound signatures to be realized. And there's also still some "specialty" frequency responses like extremely boosted bass for bass-heads.
The listeners will still have different head shapes and ears, ear canals.. so headphones will still be perceived slightly different. Different people will prefer different types (on/over/in-ear, closed, open, semi..) of headphones, different frequency responses, etc.
 
BUT, all of these headphones with similarly accurate FR would score similarly.
 
 

At this point measurements are only useful to predict the character of a given HP. Under this consideration and assuming that each FLAGSHIP closely represents the corresponding manufacturer FR ideal, well performed measurements only allow an insight of the characteristic Sennheisser, et Al sound.

Here's an interesting idea: there is no "manufacturer FR ideal". Re-using the same construction (cup, ear pads etc.) and drivers will often lead to similar sound. That's the main reasons you see similar sound from different products by the same "manufacturer".
Examples: Take an HD650 and compare it to an HD800 - completely different sound signatures. Why, because the HD800 is completely different. If Sennheiser had a "house sound" then they would have tuned the HD800 to sound like a HD650 or 600.
Grado: all the same, hence they all sound the same.
Beyerdynamic: OK, they seem to strive for "taunus sound" (boosted bass and treble), but they don't manage to do that with some of their headphones either.
 
 

An example of how, ironically, subjective this FR measurement comparison can be is the FR of the hd800 vs the t1. I will say it again, since it seems that the t1 is not very liked around here, comparing the RAW uncompensated FR responds of these HP (see headroom measurements) THE HD800 HAS MORE RESONANCES. In this regard the author proposes that resonances are indicative of a poorly designed diaphragm. So, is the hd800 more poorly designed than the t1? This is an example on how objective measurements can be used to support ANY argument if a subjective interpretation of such measurements is used. The sciency folks around here dont seem to understand this simple fact (a thing which I know, suffer and exploit for a living)

I disagree strongly.
First of all look at: http://www.innerfidelity.com/images/BeyerdynamicT1SN3964.pdf and http://www.innerfidelity.com/images/SennheiserHD800.pdf
You can compare the raw FR and also the impulse responses. The T1 does have more ringing => more peaks and dips. Also see CSD measurements at the headphone pirates site. The HD800 is extremely clean.
While I like Beyerdynamic for trying to re-invent the dynamic driver, the "fullmetal housing" of the tesla drivers seems to do more harm than good.
 
Headroom graphs give you a rough idea of how the headphones will sound, but beyond that I'm not sure they're that useful. (Still a wonderful effort though.)

 

[donthuang]

   
I disagree strongly.
First of all look at: http://www.innerfidelity.com/images/BeyerdynamicT1SN3964.pdf and http://www.innerfidelity.com/images/SennheiserHD800.pdf
You can compare the raw FR and also the impulse responses. The T1 does have more ringing => more peaks and dips. Also see CSD measurements at the headphone pirates site. The HD800 is extremely clean.
While I like Beyerdynamic for trying to re-invent the dynamic driver, the "fullmetal housing" of the tesla drivers seems to do more harm than good.
 
Headroom graphs give you a rough idea of how the headphones will sound, but beyond that I'm not sure they're that useful. (Still a wonderful effort though.)

 
1.The housing is simple a quasi helmholtz resonator , work to trap back waves, not  just "fullmetal".  and its actually abs plastic frame combine metal grill.
 
2.This kind design somehow sensitive to system's actual damping (not just damping factor number) , clearly some low output impedance amp not suit this.