The terms depth, image, soundstage etc. are bandied about in describing a systems ability (or inability) to recreate a reasonable facsimile of musicians playing their instruments in your listening room. Another way of describing this illusion is to say that the speakers "disappear" sonically. In order to recreate this illusion there are a few simple rules.
Rule # 1 The recording must be made correctly. When more than 2 microphones are used in a stereo recording, you're snookered. All the King's horses and all the King's men cannot unravel a mix. Another term often used in audio is coherent. It becomes a physical impossibility to create a coherent stereo illusion when more than one mic is used for each channel.
Rule # 2 The electronics in the system play a surprisingly small part in creating the illusion. It takes a wretched power amp to totally screw things up. It is not necessary to spend vast sums of money on exotic thermionic based units. Also, the crossover, assuming it's not a total dog's breakfast, has little to do with recreating the illusion of an image. (see Rule #4).
Rule # 3 A conventional pair of front firing loudspeakers cannot recreate a natural stereo illusion. It always sounds like a sophisticated public address system. I suppose the continued success of front firing units is due to the proliferation of PA systems wherever you go. A "live" jazz event isn't really live when everything is squeezed through a PA system. Hell, even the opera is now mic'd since most of the citizens expect things to be "loud".
Rule # 4 The dirty little secret of the recreating stereo illusion with a loudspeaker is: 360 degree dispersion at all frequencies into a solid angle. When a two way system is used the woofer and the tweeter must be snuggled as close together as possible. Both units must exhibit 360 degree dispersion at all frequencies into a solid angle. To achieve this task the drive units are forced to behave in this fashion. One need only look at the detail of an omnidirectional microphone to see how it "receives" it's signal. A grid is fixed over the capsule forcing it to receive from all directions. At the other end of the chain we simply turn the thinking inside out and "broadcast" the signal by forcing it to disperse into a solid angle. Many speakers will have drive units on different faces of the box firing six ways to Sunday. This won't work. It may sound a bit more "spacious" but since the drive units are too far apart and there is not a solid dispersion pattern at all frequencies due to lobe effects, we end up with yet another PA system. A full range single drive unit suffers the same problem. When mounted in a conventional front firing fashion the effects of a discontinuous signal launch of all frequencies will get us back to PA land. A single drive unit firing upwards and then forced to launch its signal into a 360 degree angle at all frequencies will work. The bad news of course, is that a single driver will always trade off frequency extremes - this is true no matter how it's mounted. Any talk of the basic "monkey coffin" type of loudspeaker having great imaging is nonsense on stilts.