With a monophonic source and 0 deg transformation, why are the left and right results different?

For sources on the median plane (0 degrees azimuth), the left and right channels can be processed differently when using the “Human” HRTF set or the CIPIC HRTF data, because those HRTF data are not perfectly symmetric. Most likely this is because the human subjects had slight asymmetries in their head and ears, or perhaps were not

facing perfectly forward when the HRTF was measured. The MIT KEMAR data is perfectly symmetric, thus a source on the median plane will have the same HRTF for left and right ears.

Another factor is room reverberation. The early reflection pattern will be different for the left and right ears if the distances to the left and right walls are different. In fact it sounds bad to have equal distances to the left and right walls with a centered source – the resulting early reflections are monophonic and have no sense of space. All of the Panorama presets have unequal left and right wall distances.

The late reverberation is always different for the two ears in order to produce a sense of space.