We are lucky for measurements because we have the ones that I have made and also the ones that come with the earphones made by Harsch. The Harsch ones should and most certainly are more accurate than mine using a 711 simulator. We will start with their measurements.
Now to accompany them I want to include what Samuel Harsch himself had to say regarding creating a flat earphone.
“To get a flat sound, you must have a bump at about 4kHz on the measurement system, to compensate the open ear gain of the ears The Etymotic ER4 is an example of a flat sound, with a fully compensated open ear gain). If you have a flat curve on the measurement, you will get a V curve by listening, like the PFE232 or PFE 1xx with black or green filter).
It is also common to have a dip around 7-10kHz, that is simulating an anti-resonance produced by the concha with an angle of 30% on open ear (which is what you have with speakers).
With earphones, we generally increase the level of bass to mimic what happens with speakers into a room (recordings are made to be listened with speakers):
Several things are to be taken into account:
1. Masking effect of the head:
At low frequencies, each ear received sound from both speakers with a very small phase shift, so the pressure of both speakers is summed in both ears, resulting a boost of 6dB in low frequencies.
At mid and high frequencies, the masking effect of the head become more and more important as the frequency increase, and so above about 2-3 kHz, each ear receive more sound from one speaker than the other and with a significant phase shift, so there is no sum anymore.
2. Room gains.
Speakers are generally designed to be flat in anechoic conditions, but their directivity increase with the frequency. So more energy in low frequency are radiated into the room and reflected to the user.
3. Room modes
Room mode are standing waves between the walls that also increase the level of low frequencies.
4. Room Absorbtion
Rooms are generally more absorbent in high than low frequencies.
All that stuffs must be taken into account in designing earphones.”
Very interesting indeed. By all means as well the measurements back up exactly what he has to say. Slightly boosted bass, 4kHz boost and 7kHz drop out. Under this strain of research they should indeed sound flat, which is quite against what Josh is perceiving, at least in his subjective opinion and from comparing other earphones. With a specific and likely smartly worked out compensation curve these will likely measure very flat so while Josh may have been angered by the sound, to the best of Samuel’s research he has likely done a VERY good job. It just was likely nothing like what Josh would have achieved having used the U-Tune machine himself, maybe that would have been what Samuel calls bass light?
This is the issue, speakers, earphones and headphones are very different ways of listening and the numbers that work for one won’t necessarily work for either of the others. There are so many variables to sound and the way it is perceived, these measurements and values that are quoted don’t really tell much of the story. Of course they give you an idea but the UERM and this earphone sound more different than the graphs suggest. Also, myself and Josh hear and perceive things near as anything identically, and I know I wouldn’t consider this flat either, at all.
Now to move on lets look at my own measurements. While not as accurate as the 711 Samuel uses and with some obvious differences visible my measurements will show differences to other products accurately. The scale differences in our measurements though will cause a lot of the visible differences but the feel of the curves are very similar. The biggest difference is the bass position relative to 3kHz and above on with it being more prominent on the 711, this is something we have already established with the Vibro Veritas I use.
First and foremost I have to give credit to the top notch channel matching with these earphones, extension being very impressive to roughly the quoted 18kHz. But moving on from that I want to look more at what this means in terms of comparisons to other graphs. Lets take the Ultimate Ears Reference Monitor for example. This is something known widely for being one of THE FLATTEST CIEMs but even by many including myself is said to have a slightly boosted mid bass.
The bass is certainly more dominant on the Harsch. While at 2kHz they sit at the same point in dB, the Harsch have moved ahead 3dB by 100Hz while the UERM have stayed flat. I would not say that makes them heavily bassy but more bass than the UERM that I believe to be boosted in the mid bass as is. The midrange of the two is very similar but when it comes to that 7kHz drop that Samuel talks about the SH-3 does it much more dramatically, likely playing a hand in the issues Josh has with the treble, although I could be wrong of course. From 4kHz to 8kHz the Harsch drop 12dB, the UERM only 4-5dB. A significant difference. They both also show similar all off after 100Hz one of the bigger complaints with the UERM. One thing I have to compliment the Harsch on though is how much better the channel matching is than on the UERM, WOAH it is considerable.
I want to finish this with I am very new to measuring and the the whole concept it entitles. I am not perfect at perceiving graphs and know there is much more to sound that simply these and even a perfectly flat measurement could have loads of bass decay that makes it come across bassy, maybe the case with the Harsch, I don’t know I haven’t heard it. All things like this are helping me learn and that is certainly a good thing!