I'm by no means an expert in any of this, so please don't get bogged down in any of the details I might present here. The numbers are only approximate, and I may have ranges fairly wrong as well. Also, I'm sure I am leaving out some details that I have misunderstood in my own reading.
Here's a general presentation of some of the (from my perspective) most important psychoacoustic principles to understand when designing speaker layouts and acoustic treatments, in no particular order.
One of the key ways your ears and brain work together to localize sound is through the time gap and phase change between when your left and right ears hear a sound. For this to work, the sound waves have to be a fraction of the size of your head, or the SPL and phase will be effectively the same for both ears and your brain won't know the difference (a 10cm wave is 3.4kHz - pretty high treble) This only works for sounds side to side. Your brain uses different tricks for up and down and fore and aft (the hardest to discern).
Elevation of sound sources is largely based on the diffraction cancellations created by your outer ears themselves (the pinna). When a sound source is above you, the outer ear creates a cancellation of a range of high frequencies (somewhere around 8kHz, I think). When your brain hears a sound with that portion of the spectrum missing, you hear it as above you - simple as that. It's very reliable - just ask proponents of reflected Atmos speaker implementations.
Why high frequencies? Well, they are less perturbable, in a way. A small wave encounters most anything and is either reflected or absorbed - because the object it encounters is so much bigger than it. In contrast, large low frequency waves refract around most normal sized objects. So, when you want to know if the sound of someone's footsteps is coming from within your bedroom, your brain listens for the high frequency content of the sound to see if it's there - when it's not there, you know the sound came from around the corner.
Commonly, high frequency sounds are associated with the onset of a sound - the sound of the tongue against the teeth and lips when speaking; the fingers on a guitar string as they strum; the shoe leather touching down on the floor before the weight of the step causes the lower frequency boom. As your brain listens for those sounds, it determines the direction of the sound source. If an acoustically reflective surface near a speaker (or any other sound source) directs a reflection to your ears, your brain can be tricked into thinking the source of the sound has moved toward the reflection. This is the foundation of the (generally misguided) argument to absorb early lateral reflections in listening rooms. Especially for music (and most surround channel content) these early high frequency reflections will broaden the image, making them seem more ambient and less pin-point. That's found to be an advantage to most listeners, and why absorption is not the best tool in most cases. If those early reflections were slightly delayed through diffusion, all the better for those goals. I think this is one reason we see more and more diffusion to the sides and behind listening positions in recent treatment plans. But again, this is only relevant to the frequency ranges that your brain uses to localize sounds.
For midrange and low frequency sounds, your brain is much less sensitive to localization issues. Instead, the most pressing concerns are frequency response, frequency extension (bass), and ringing/decay. Frequency extension I mention just because people talk about it with speaker evaluation - lower bass extension is pretty universally preferred - end of discussion.
Frequency response is obviously a function of speaker design, but also and usually more strongly a function of speaker and listener placement. The mess of reflections that your brain is always sifting through make it less sensitive to the "comb filtering" or other frequency response anomalies associated with the mostly destructive interference caused by reflections. Think of these in terms of intervals (octaves and so on - on a logarithmic scale). When the bandwidth of a cancellation dip is 50Hz that's nothing in the treble ranges (a small portion of an octave), while below 400Hz or so that's a huge fraction of an octave - several notes on a scale all swallowed up by a cancellation dip. Where did that cancellation originate? Usually from a strong room boundary within a few feet of either the speaker or listener. This problem is simple, and best addressed through speaker positioning and/or the implementation of a baffle wall, or simply moving your seat away from the wall. Once you are away from the wall by more than a few feet, the frequency of the cancellation has shifted up the scale to where it is less audible.
One of the more nasty (but common) problems with low frequency response anomalies is ringing, caused by uncontrolled standing waves or simply an unhappy set of reflections leading to constructive interference. The increased output leads to a variety of problems; obviously the uneven frequency response is in itself wrong for one, but for two, the high output at one frequency masks nearby (usually higher frequency) sounds - your brain doesn't perceive them at all, even when they are loud enough for you to hear them under other conditions; and three, depending on the reason for the increased output (ringing vs simple constructive interference) there are temporal problems where the bass lingers longer than it should.
So where does that leave us? Here's a couple relevant observations, I think. Frequency response is still king, but keep in mind that's frequency response at your listening position - which is not as simple as the anechoic response of the speaker, as the spectral content of the reflections play a role. So, if you are treating first reflections of the speaker, bandwidth is a concern (though not a huge one, IMO). Mostly, at this point I want to refer you to some well-thought-out best practices: http://www.acousticfrontiers.com/wp-..._standards.pdf
<--- This should be the target. The psychoacoustics explains why and helps with how.
(Sorry for this rambling nonsense, I know this isn't an easy read. There's also a good chance you know most of this already - I hope you don't feel like I'm talking down to you or anything. I'm just trying to shift your perspective to finding solutions to problems. Acoustic treatment design is an engineering exercise in pursuit of a set of goals, so it's always a problem-solving endeavor.)