From the perspective of someone who has and am still working with stuff at 1000C (myself)
Where you measure temperature is almost never where u actually want to measure it
Cos u dont wanna disturb the process ur trying to effect
Thermocouples drop off or dont work past some range
IR cameras ask u to insert an emissivity and u calibrate that at room temperature but the emissivity might change at high temperature and u cant put a scaling hot thing in an integrating sphere
If ur flowing gases for materials growth they will cool the specimen so the temperature will be different one place from another
I dont know about cold stuff. Only thing ive measured in that range are cold atoms :p
Thermocouples also have some thermal contact resistance unless u weld them on
So that magnifies cooling effects and also creates a lag if ur trying to measure time dependent temperature
