I may be wrong (in which case, please argue with me!), but I think the opposite is more likely true, i.e., that measurement of a CIEM
built for you is more likely to be representative of what you hear than measurement of a UIEM would be. All UIEMs are going to have different fits in people with different diameter and length of ear canal. The IEC 711 spec was really designed for deep-insertion hearing aids and only has the half-wave resonance peak above 10 kHz with deep insertion. At larger distances between driver and eardrum (pretty common with a lot of modern shallow-insertion IEMs), that peak can easily drop below 10 kHz. From Morten Wille's white paper:
So it's a crapshoot trying to measure UIEM's treble response - it can be all over the map. It's not their fault. Or the coupler's fault. It's our fault, as a species, for having a large standard deviation in ear anatomy.
Most CIEMs are created from an impression or scan that goes to the second bend in the ear canal, so you've at least removed everything prior to that from the equation, and we can be fairly confident that there's less variability from the 2nd bend->eardrum than there is over the entire ear canal. Amplitude and spread are still likely to be a bit ambiguous/uncertain because the coupler mic is normal to the source/canal, whereas actual eardrums are angled, but as long as you get a good seal (mounting putty should be perfectly adequate) with the CIEM tip at the coupler reference plane, I think you could be more confident about the location of those canal resonance peaks than you could with a UIEM measurement.
Representing UIEM data is tricky, because there's a very good chance you'll hear that peak somewhere other than at 8 kHz. I understand
@crinacle had to make a pragmatic choice, and I have to tread lightly here, because 1) I greatly appreciate and respect the work he's done and 2) I can't really think of a better solution, but forcing an 8 kHz peak for each and every UIEM (or CIEM?!?) doesn't properly differentiate between shallow- and deep-insertion IEMs, i.e., one that may have an 8 kHz canal resonance peak, and one that almost certainly doesn't. I suppose one could create a plot that has a superposition of all the above peaks, but how would readers figure out where they are in that spread?
@SilverEars - good idea of yours about us getting all our ears scanned
But for UIEMs, you'd really need a full 3D CAD representation of your ears all the way to the eardrum
and a full 3D CAD model of your UIEM (+ eartip)
and a clever CAD package that could put the two together, because driver->eardrum distance won't only depend on your ear canal length - the shape and diameter of the ear canal will dictate where the UIEM ends up seated.
I don't know what the answer is to all this. ¯\_(ツ)_/¯