It is theoreticly possible to use all very small microphones for recording the hrtfs.
But there are several issues to be handled. Small mics have a bad snr ratio.
I used two Sennheiser KE4-211-2 microphones. The snr is 55dB wich is good for
a small mic. It is electret and needs very(!) thin cable to fit into the ears with foam
to block the ear canal (and a battery and little circuit to energize).
I sugest for interested persons to read some papers or books about that matter
before starting. It depends on what one expects to do with the HRTFs.
Binaural simulations need very complex measurements and more complex
corrections to be simulated realisticly.
First problem is to present the needed sound incidence directions accurate enough.
That is essetial for later reproduction.
The Smyth Realizer works therefore with a headtracker and in ear mics to record
a set of HRTFs but with room information for exact reproduction of that specific
room for only this exact person. It measures also the FR of the used headphone
to invert that (or equalize the HRTFs whatever is done there). Again personalized.
Then you can feed a binaural renderer with the HRTFs and the actual orientation
(as done with the hardware box that has a binaural renderer inside).
Static representation of the HRTFs is not very satisfying as I learned with my
properly recorded (horizontal) HRTFs. One confuses often front and back.
That can be overcome by headtracking. Nevertheless it makes the sound more
speaker like (or a bit like the Sigma but with clearer presentation).
If you want a static sim you can just record the left and right speaker seperatly and
hold your head at exactly the same position and use that for left and right channel.
With more speakers it's clear what to do then.