O.K. this turned out long but read it anyway and open the linked graphs in separate open windows so you can jump between them when you alligned the pages vertically you can esily spot the differences.
Indeed graphs are difficult to read.
squarewaves alter their shapes depending on frequency response (there should be a flat horizontal line and a steep vertical line without overshoot (that is the membrane not 'stopping' directly after it has been motivate to move and stop again) and a longer Z in the horizontal and the way it swins out (get's smaller in amplitude) says something about resonances and how well they are damped.
No dying sinewave in the 300Hz squarewave says it has excellent damping a little overshoot is allowed.
No steep rising and falling and 'slow' rising says it doesn't do highs very well.
No horizontal line but sloping downwards says it's high centric.
In the 30Hz squarwave you can see bass behaviour and if the headphone is bright or not.
a small peak indicates it has fast treble compared to lows and mids.
steep downwards slopes (straight) says lows roll off.
when it is curved in the downwards slope this says something about the frequency response between 30 Hz and 300Hz.
When you look at the frequency crves and squarewave of the CAL and D5000 you will see similarities and the CAL even seems 'better' than D5000
www.innerfidelity.com/images/CreativeAurvana.pdfwww.innerfidelity.com/images/DenonAHD5000.pdfThis led me to believe you could just as well buy a CAL instead of the very expensive D5000 and only would have a difference in comfort.
BUT there is another thing and that's the distortion at 100dB SPL (so when playing above average levels but not very loud yet)
Notice how the lows (100Hz) in the CAL reach 6% to 7% (that's not quite the 0.01% amps have !) where as the D5000 is around 0.3%.
This is ruthlessly showing the weakness of the CAL (bass department) and the squarewave overshoot of the D5000 says it's brighter/more detailed.
Also the little humps in the impedance graphs show resonances at certain frequencies.
The higher this hump the worse it is.
The V6... (7506)
www.innerfidelity.com/images/SonyMDRV6.pdfNotice the fierce overshoot (sharp edge) and the not linear behavior in the 30Hz SQW.
Also look as the 300Hz SQW and notice how the 3kz resonance is rather big in amplitude and long in time.
Also the impedance graph shows this weakness.
See how the distortion rises in the lows (similar to CAL).
The price of this HP is also a give away in this area eventhough it has some strong points as well.
But which headphone doesn't ?
Here is another one that acc. to graphs would sound pretty bad:
www.innerfidelity.com/images/GradoPS1000.pdfa bucketload of resonances (look at 300Hz SQW), sloped and rolled of lows (30Hz SQW) distortion going thorugh the roof (10% to 60% in the lows), bad damping (see how long it swings out on the impulse) yet it sounds lovely to most who hear it.
This can partly be explained by the distribution of resonances and harmonics (a bit like tube-like distortion where the upper harmonics have a certain decay opposite the fundamental waves) which makes the added distortion add instead of distract something to the MUSIC signals.
Notice distortions are several orders higher than even the worst measuring amps.
Knowing this you can also see the reason why certain amps that measure really bad in distortion figures or have limited FR range this still can come out as 'nicer sounding' than perfect amps.
Simply because the added harmonics 'add' to the sound in a positive way and do not degrade. Where as a different harmonics behavior with considerable lower numbers but no decay in amplitude for the higher harmonics may give better NUMBERS but sound worse.
Needless to say below certain levels this becomes irrelevantly low in amplitude and is 'masked' by recorded harmonics( which are higher in amplitude).
An amp that measures perfectly and adds absolutely nothing or inmeasurable low will sound good to technical people that appreciate it yet may sound to clinical for music lovers who simply prefer 'additions'.
This is what's difficult and enhances the idea that graphs mean little to nothing, yet they do when you also use the ears.
Now for example take the LCD-2 and look at those FR graphs and squarewaves and see how they are approaching the 'ideal' squarwaves
www.innerfidelity.com/images/AudezeLCD2.pdfSee the (grey lines) FR and squarewave, notice the impedance graph and above all the distortion graphs.
Here is another intersting one:
www.innerfidelity.com/images/BeyerdynamicDT1350.pdfNotice the (grey lines) FR and squarewave, notice the impedance graph and above all the distortion graphs and take into account this is a portable and several times cheaper.
a headphone that represents the SQW and flat (grey lines) FR one would expect to sound the best as it does reproduce what's been applied to it (line electrostatic, magnetostatic and some horn speakers do) but that will only be the case to those that listen to the technical aspects and appreciate those.
There will be many people that like their Bose or beats much better !
Even for me it isn't easy to accurately predict how a headphone will sound and that will only apply to those with somewhat similar hearing and preferences as well.
I have looked at and listened to (at the same time) a lot of squarwaves and pink/white noise signals to understand that matter and link measurements to real life occurances and this gives me an idea about what is important and what not.
Really complicated stuff but in the end it really doesn't matter that much.
It's you, your hearing and preferences/taste that should decide what sounds good or not.
Those using it as a tool in audio engineering need accuracy instead of niceness/pleasant sound.
These simply do not combine very well, maybe with the exception of the better HP's such as orthos .