This sounds intriguing. Is this the process you're referring to?
https://sourceforge.net/p/hesuvi/wiki/How-To Record Impulse Responses Digitally/
Or is there another way to do this with in ear mics?
Yes, that's how you would do it. I would just process the analog output from the headphone output of GC7 (while trying to adjust the output volume so that it is within line output specs or near, to avoid distortion).
You can do it with ear canal mics, but that's like doing SXFI (which simulates your ear HRTF) + HRTF of your ears (from the in-ear mics), which is like double HRTFing (not worth it).
The best way to do it, if you can afford it, is to forget about SXFI, just get a 7.X full frequency loudspeaker setup (calibrated), get high quality in-ear microphones, then record the impulse responses of the 7.X loudspeaker output with your mics in your ears .then process this and use it as the personalized HRTF function within HeSuVi. Personalized, accurate, no more guessing.
Bonuses:
- no more generic SXFI guessing - your ears+head+torso HRTF, actual impulse response modelling, not a generic algo
- you can record this in the kind of room you want to "hear" with the final HRTF file : if you record it in your room, the impulse response file and the HeSuVi simulation using this, will sound like your speakers in your room (not some generic algo room made up with maths)
- Near zero millisecond processing delay - just pick the fastest sound card you can find that has the lowest overall latency, and then your sound playback latency will be: Windows driver/audio stack latency (can't change this yourself) + Your selected sound card latency (differences here can be up to +300 ms from card to card) + HeSuVi additional latency (should be within <5ms within most modern fast systems)
- Pick your sound card, pick your DAC, pick your headphone amp, pick your heaphones. No more crappy compromises.
It's cumbersome, expensive and time-consuming, but probably will give you the best results.
More info in this long thread (incl. software, links, examples, instructions):
https://www.head-fi.org/threads/recording-impulse-responses-for-speaker-virtualization.890719/