If a sound has two channels and you want to make it sound in front of someone in one instance and behind someone in another instance, what is the difference between the two sounds? Is it the stereo then you can change the panning to affect the left/right location, but what about front/back?
Volume, frequency range, and reverb/delay can be altered to give the illusion of depth. For volume, a louder sound will seem to be closer than a quieter sound so simply making the sound you want to place farther back quieter is the most obvious method. Sounds tend to have less higher frequencies when they are farther away. So filtering out some top end will add to the illusion. Finally, add more spacial effects such as reverb/delay to the sound you want to place farther back. The crucial principle when applying all three methods is that the illusion of depth comes from contrast - you need to place a sound farther away IN COMPARISON TO a different sound that is much closer. In other words, the farther sound should be quieter THAN a different sound, have less high frequencies THAN another sound, and have more reverb THAN another sound.
Hope this made sense.
It's all about the reverb. For "behind", the "direct" sound has to go through the auricle (giving it frequency-specific phase changes and attenuation) with the first reflection from the front being more neutral. For "frontal", the amplitude/phase relations are almost opposite.
This requires keeping tabs on several frequencies and their envelopes. Something like a continuous sine wave is much harder to locate: basically you do that by "ear saccades", moving your head and tracing how the sound reacts to that. Of course, that does not work with a stereo recording unless accompanied by visuals giving you a cue for the purported movement.
Without movement and without a somewhat reverbing environment, front/back location would not really work well. You'd have to rely on wide-spectrum sound sources with known phase relations (motor noise, click trains) and even that would be iffy.
The point being that usually a number of different cues individually adapted to auricle form are employed and combined, and most of them work to some higher or lesser degree even in isolation. But usually they are employed in a complex ensemble.