About IEM : in-ear monitors aka intra-aural monitors

Trying to answer my ultimate question : can you equalise well enough a cheap earphone so it is as good as a costly one ? Can you buy an Ety ER6 and make it sound as good or better than an Ety ER4 or Shure E4 ?
1) I wanted to test if for good earphones, linear distortion (aka frequency response) was the main difference, and if non linear distortions (THD, IM,…) are low enough to be neglegted.
2) Another important point : are earphones of a minimal phase nature ? If yes, doing a minimum phase equalisation corrects the amplitude and also related phase non-linearities.

I also saw some measurements especially at headroom and by Knowles, and I wanted to understand those measurements :

  • do IEM show important dips and peaks at high frequencies ?
  • is the response in mid and bass very linear ?
  • is the distortion of balanced armature really high ?
  • is the HF response of IEM a by-product of distortion or a real HF extension ?

So I did some basic tests on IEM earphones I own, Etymotic E6, Shure E3c, Sony EX71.

    My measurement setup consist of a Panasonic electret WM60ay mic at the end of a conical cavity (total 24mm long, about 17mm long between IEM end and mic which has 6mm diameter, a volume of 2cm3) to mimic the ear canal. Warning : those measurements are only valid as a comparison basis, it’s not a true metrological and scientifically checked method !
    First step : sensitivity and frequency response. All below curves are measured with the same output voltage, the E3 output at 1kHz is about 7dB higher than ER6 and 12dB higher than Ex71 (but due to its special curve, the perceived loudness of the EX71 should be nearer 9 to 10dB less than E3). First thoughts : the Sony has a very uneven curve. The Ety has the 2.5 kHz peak to EQ for the missing outer ear (pinna EQ, a bit like diffuse field equalisation). Both Ety and Shure have very smooth curves under 1kHz.
    I wanted to check channel imbalance : good results on all IEM, hereunder is the ER6 L and R comparison.My concern was distortion on balanced armature IEM’s, is the high frequency extension only a by-product of distortion ? So I checked on all three models, remember that the Sony uses a standard dynamic transducer.The figures are quite good, even for the balanced armature models.

    Is the effect of different eartips noticeable on measurements ?

    I also tried to check this point and you’ll see frequency responses of the E3 with various eartips. You can see differences in high frequencies but not in the lows. But I took care to have a good seal, if the seal is not fine, you can get very differents results in low frequencies. See the curves shown at the end.

    After those tests, I wanted to make some basic listenings.
    Low bass extension listened with sine wave : all models give noticeable output down to 25Hz

    High frequency : listening to sine waves, I got about the same limit of 14.5 kHz for all models so it means that it is my own ears limits. How can I hear those frequencies when the measured curves are so steeply cut above 10k ? What’s wrong with my test setup ? Mic and preamp are not the cause,
    Frequency smoothness : listening to sine waves, every model shows different behaviour :
    Shure E3 has peaks at 7.8 and 11.5 kHz
    Etymotic ER6 has peaks at 3.5 and 9 kHz
    Sony Ex71 has peaks at 7.8 and 14 kHz.

    I pointed the peaks and not the dips because peaks are allways more damaging and also because it’s more difficult to equalise dips : you are limited to a few dB’s to avoid distortion and not to overload the system.
    To find the ideal EQ curves (gain and Q) to correct those peaks, listening to pink noise may be the solution.
    But on the Sony, equalising those peaks won’t surely be enough (due to its very uneven freq response) and I see no chance to get a neutral sound. Now, I’ll have to find the good EQ for the Ety and the Shure (at first glance, I would think the Shure has to be equalised to have the 2.5kHz pinna curve that the Ety has from scratch). And I want to know if eqalising two comparable phones gives really the same audible result. Because I’m used to the E3 and I find that it needs some EQ in the high-mid and it is missing some high frequencies. The Etymotic, after a few days of listening, seems better in HF but really needs some EQ.
    More on this later…