I am measuring the speech parameters of an acoustical system (a room) and not a sound system

In this case, distortion should be removed (since the distortion is not from the acoustical system but from the sound system that is used for the measurement).

The distortion can be removed using sine sweep as excitation signal and cutting the last part of the impulse response, but this will also remove some of the noise.

Therefore the noise should be measured separately and be corrected for in the table as discussed in the section above.