Originally Posted by Thujone
I keep seeing this over and over again, so I will finally ask the question. What do you piercing highs and female vocals have anything to do with one another? The highest note ever sang by a human being is somewhere around 2000 Hz. The piercing highs on this headphone come from the 8000-10,000 Hz region. If you have a man or a woman make a "sssss" sound, it will be in the highs region (5000+ Hz) regardless - but that isn't considered a note being sung because it doesn't use your vocal chords. That being said, all this talk about female vocals being piercing seems to me like placebo is taking its toll. When I first tried to tackle the question of what mids, highs, and bass were in terms of frequency, I was initially under the impression that sopranos would be considered in the highs category based on several sources on the internet. Then, I found sources with concrete frequency ranges to describe the three, and I found that the "highs" are well above any note that can be produced by any human or normal instrument with the exception of cymbals and artifacts created by the human voice (whether man or woman) such as "ss", "th", "f", etc. Yes, I will admit that timbre of some instruments (snare drum comes to mind) is dependent on the driver's treble extension (ie more pronounced highs brings more crispness to a snare drum), but I still think piercing highs and female vocals have nothing in common. I have plenty of male vocal recordings that are insanely sibilant and other female recordings that aren't... Am I missing something huge that needs to be considered?
Primarily that the human ear is naturally sensitive to frequencies centered around 1 khz (sort of how our eyes are 'tuned' to the color yellow and we tend to be more sensitive to it). The HE-400s 'bump' in frequency response in the 1 khz area, followed by a deep recession through to about 6 Khz may be accentuating a frequency that makes them seem a bit 'hot' compared to more neutral headphones. If you add to that modern recording styles, which can be heavily compressed and include over-pumped up vocals, you get a really bad combination. In my experience, clipped recordings give off a slight 'static' sound with the HE-400s. Once you notice the static, that is more grating than anything else (I thought my phones' were failing at first, then I started looking at waveforms using waveform seekbar in foobar). Also in my experience, audio engineers appear more willing to screw around with female than male vocals, given that women are not only better looking, but usually have much much prettier voices. Notice how many female vocalists are the central focus in many pop recordings that feature a female-lead. They are also 'hot'. Florence and the Machine are my favorite example - her voice and music is lovely but the recording is hot to the point that it is fatiguing even on my lower-fi equipment.
I agree with everything you have said, because I too have complained that those irritated by 'sibilance' a) need to separate sibilant recordings from sibilant speakers and b) need to stop blaming upper treble for sibilance when, in all liklihood, the majority occurs in a much narrower mid-range, not treble band. Hence tweaking response up to about 4 khz might be more effective than 10 khz. On the other hand, cymbals and other percussion make lots of 'ssss' sounds as well, so some folks may be conflating sibilant voices with other sibilant sounds. That is not the same thing to me and should be part of different discussions, but I can see how such confusion would cause people to stress focus on much much higher frequencies with apparent disregard for physical reality.
It seems true that 'audio artifacts' (such as with a female voice) in recordings may bleed into higher frequencies - making vocal sibilance in higher frequencies a 'digitally' if not physically possible phenomenon. You could, for example, just record yourself singing and will likely see audio information in frequencies both higher and lower than you expect. However, this can be misleading. The reason is, most record artists will want to filter out unwanted frequencies (to take out voice 'boom' and vocal sibilance). But also, if the high frequency response were, say, 3 dB lower than the 1 khz response, the frequencies we would expect to hear would be 'twice as loud' as the ones we didn't expect to hear. In other words, sure, they are there, but they are drowned out. The presence of artifacts is not, in other words, proof that all irratating sound is in the upper treble (no one ever seems to cut the bass using the reverse logic, for example, but heavy bass is also fatiguing).
Finally, I do not see the HE-400s as really 'peaking' in the upper treble range, for 2 basic reasons. First reason is that our ears are less sensitive to those frequencies - so boost is expected if you are seeking something approximating 'nuetral'. I think this is true even in the headphone world (which is admittedly very different than the speaker world). Second reason is that, when looking closely at the frequency response curve, if we were to center at 1 khz, the upper treble never actually significantly exceeds the other 'peaks' in response. If anything, the upper midrange frequencies which I would think are more irratating are the most deeply recessed.
As I continue to ponder these headphones, I have recently decided that part of their 'real' issue is that they are quite unforgiving. If you have a bad recording, they will be sure to tell you! If you have a really great recording, they will sound great! A lot of new music is unfortunately pretty badly recorded, IMHO. I read the article about HD-650s wherein the designer of the O2 argues that they are a lot like 'polarized sunglasses' - doing a lot of filtering on their own to reduce fatigue and sort of 'improve' on recordings. I would now love to try them out to see if that seems true in a back to back comparison with the HE-400s, since my other Sennheisers, in doing a little of that, ultimately seemed more veiled than anything else on an initial listen.