I was reading through threads on reddit about Noble and came across one that linked to this
measurement test blog post about the Kaiser 10U with a lot of surprised responses that didn't offer up much further explanation as to why the results were surprising. Any insight as to why the test results were surprising (I am new to higher end audio).
The answer my friend, as already mentioned before, is "scale." Now, for the novice, "scale" may not mean much, so let me create an example. The surface of the skin, when viewed under a magnifying glass, may look a bit more clear, you may see the skin pores a bit more clearer. But when the surface of the skin is viewed by a microscope under a lot of power, the skin surface now looks like the pocked marked moon surface, and viewed under more power, the skin surface may look like the Grand Canyon.
The magnification value/power level, represents "scale" on the frequency graph. If you bother to look at the graph closely, you will find that the difference between 50 Hz and 750 Hz is a meager 5 dB. All things being relative, that isn't much, but under a "high powered" microscope, that 5 dB drop can be appear as a huge trough.
Let me preface this post by saying that I have not heard a K10 myself, but my post is based on some knowledge I have gathered about the K10 on the internet. And also, my post is geared more towards the FR measurements..
1. I just compared the measurements of some of the IEMs and Headphones on this site (http://clarityfidelity.blogspot.com/) to Tyll's measurements (http://www.innerfidelity.com/headphone-measurements) and they seem to line up for the most part. To me it suggests that the person has taken effort to calibrate the measuring equipment/software and hence the measurements can be taken with certain level of credibility. YMMV
2. If I have to take an educated guess why people are surprised with the measurements, I think it has to do with the peak at 8kHz, which would suggest a piercing treble, and that seems to be in contrast to most K10 owners' experience, that the IEM is smooth in the high frequencies
3. All that said, measurements above 3kHz should only be used to get an idea how an IEM might sound in the high frequencies, and should not be taken as is, because the measurements above 3kHz vary based on shape and size of the ear canals. So it is not a good representation of how a person might hear it.
Point 2.
Regarding the piercing treble peak, all things being relative; that peak is about 2-3 dB above the response at 50-60 Hz. Keeping that in mind, the Er4P considered by many to be the king of neutral, has a peak that is around 10 dB greater than the band with of around 50 Hz - 1k Hz, few people, if any at all, say the Er4P is piercing. I also don't consider the K10 to be piercing either, and I would think, that a product design that has been on the market for 3 years, has been reviewed over 20 times, has received over 20+ consecutive 5 star reviews (no product has that record) and has held the number one position, longer than any design on the history of HeadFi......
piercing highs, would have been mentioned by now, and I think that would have impacted its ranking. (just my opinion)
If anything, most reports state that the highs are not piercing, but a bit relaxed (relatively speaking).
Now this isn't gospel, there may be those that say differently, as this is sound we are discussing, so the interpretation of sound will be subjective and certainly has the potential to be different from person to person.
The K10's have the single
best high frequency response of any IEM I have experienced. This was what first drew me in when I listened to the demo models. So far, I have not heard anything better.
As always...YMMV...
Your opinion seems to shared by many
The curve seen in that blog seems slightly exaggerated because of the scale, as mentioned by a few users in the Reddit thread, but having seen some measurements of my k10 custom, I'm not the most surprised, apart from a clear bass presence, there's also a slight presence in the 1-2k and more prominent bump in the 5-8k region, those are all seen in multiple separate measurements I've seen. They extent of the 5-8k peak would arguably be larger with the aluminium model given what I hear (and perhaps more geared towards the slightly higher frequencies), but it's really the scale that's exaggerating things here, scale it down closer to how most sites publish measurements and it will look fine.
As you said, scale is the key