When it comes to evaluation of a headphone, the last tool I use are graphs.
I measured for the following purpose (as written}: "I did some measurements to see if I could equalize it properly". The graphs are for analysis purposes, not to determine how a headphone sounds. You can explain some characteristics though, and explain why equalizing the phone with an ordinary parametric equalizer has its limitations.
I initially tried to tune the headphone based on what I heard, I didn't manage, so I measured a couple of weeks later to see if I could fall back to measurable properties (sometimes - or even many times - your ears are fooling you). The measurements explain some items that I hear:
- Treble somewhat unnatural sounding
- Hollow sound characteristic, voices disappear somewhat into the music, missing energy in the fundament region.
Many people say the X2 is not a bass monster. This you can also explain from the measurements, it has a dip in the fundament region. So there is relative more bass compared to where the fundament of most instruments is situated, but not in an absolute sense! I design loudspeakers myself, and there it is very obvious that the relative magnitude of frequency regions of about on octave matters a lot to how a speaker sounds. Lowering the frequency spectrum of a very narrow frequency range with 0.3dB is hardly audible, lowering a whole octave is easy to spot. Here we talk about 4 dB in the fundament region. I tried to lift it, but then the sound starts to get shut-in. This phenomenon can be explained by the resonance that is present at 220Hz. It might even be that Philips deliberately lowered the energy in that region to compensate for that resonance of the enclosure.
What is not explained with the on-ear graphs immediately is why the Philips sounds more hot in the treble region. There is only a small peak at 9kHz. So I also did a few more measurements with different distances to see which peaks-dips are due to geometric properties of the phone (also explaining why phone characteristics are so different when measuring on-ear), and which ones are intrinsic. You see that the peak at 2kHz shifts, and completely disappears when put on-ear, so it is a reflection. The ones at 5 and 9kHz are certainly there intrinsically.
Looking at the different between the on-ear and near-field measurements, you can see that the Philips has some issues in its treble, where the Sennheiser is pretty clean.
X2: (on-ear is yellow, near-field purple, observe the peaks that remain at 5kHz and 9kHz)
HD650 (on-ear green, near-field blue; a small peak at 2.5kHz, which is assume is on purpose and diffuse field adjustment):
Whether at the end you like the tuning (overall timbre) is a completely different story, and something of personal preference. Records are also mixed differently, so trying to go for an "absolute reference" phone is not a guarantee for ultimate music performance; a reference may work better with some records, and worse with others. So a better measurement does not mean "better sound to you". To give an analogy; suppose you have a time machine, and can go to the 17th century where Rembrandt is painting his famous painting "de Nachtwacht" (Nightwatch). You are there with the best camera equipment, and make a photo of the same setting, go back, and let it print at the same scale as the painting. The photo will probably be more accurate in color, contrast, dynamic range etc. Nevertheless, most people will probably prefer the painting, as the deliberate effects of the light and contrast is what creates a more intense emotion with most observers. I.e., I understand the limitations of a measurement graph, it is for analyzing purposes only, not to claim it creates a better reference headphone.
Overall I think the X2 is a very nice headphone, and I could certainly recommend it to many people. If you like acoustical music, and appreciate it for the wide field of colors and timbre, then go for the HD650. If you prefer the ticking of a plectrum against guitar strings above the resonance of the guitar enclosure, go for the X2. If you look for a headphone that you can tune with a graphical equalizer to tast (and have a decent amplifier to steer it), the HD650 is a better choice, as is supported by measurements.
So, you would never use graphs (which is fine of course), and as an answer I give you even more graphs in return
This is probably the fun of a forum.