There are two factors to consider with respect to cabinet shape. There is both the internal shape and the external shape. I express these ideas as opinion and will offer support for my ideas.
Internal cabinet shapes other than a properly dimensioned box are marketing hype, and a fashion statement. If you like them for aesthetic reasons that's fine. They have little to do with the overall quality of sound reproduction. I offer four items of support.
1. Given the top end hifi speakers available, about 95% are built in a box. If the box were truly deficient, more manufacturers would be using other shapes.
2. All enclosures/cabinets/rooms have internal standing wave modes. Odd internal shapes don't reduce the modes in a speaker enclosure. Odd internal shapes merely make those internal standing waves more difficult to calculate. This is discussed at length in the Master Handbook of Acoustics.
3. "...slanting the walls to avoid parallel surfaces... does not remove colorations; it only makes them more difficult to predict." Christopher Gilford, Acoustics for Radio and Television Studios, (1972)
4. When discussing ideal room shapes Mr. Everest writes, "for example, parabolic shapes yield beautifully sharp focal points and cylindrical concavities less sharp but nonetheless concentrated. Even polygonal concave walls... result in colorations of sound in some areas at the expense of others. The popularity of rectangular rooms is due in part to economy of construction, but t has its acoustical advantages. The axial, tangential, and oblique modes can be calculated with reasonable effort and their distribution studied." F. Alton Everest, The Master Handbook of Acoustics, (2001)
Asserting the superiority of the plain old rectangular cabinet is a strong statement, but try this. Go to your local hifi shop, or write to the manufacturer of one of these non-box speaker manufacturers. Ask them what their box modes are. You might get an answer, but you will likely get the "deer in the headlights" look. This is because the manufacturer hasn't paid the engineer the couple-thou $ to accomplish this. The manufacturer has paid that couple-thou $ to the marketing department, and to the cabinet shop. It is likely the marketing department that initiated the cool shaped (non-rectangular) enclosure - not the Mechanical Engineer.
There are, nonetheless, good speakers in odd-shaped enclosures. This is not due to the enclosure. It is because the design was executed with a proper crossover, good drivers, and a solid/damped cabinet.
External shapes are a different story. Baffle diffraction is a significant part of loudspeaker design. A rounded external shape/corner can be somewhat advantageous, but the radius must be 1/4 lambda (wavelength) for the waveform to be affected. The egg-shaped enclosure page has reification for the rounded enclosure. The flat baffle also has an effect. This is called baffle step. Go to these links. Other folks do a better job explaining baffle issues than me. One point of interest is the following:
Harry Olson Ph.D., in Applied Acoustics, 1939, p204 wrote, "It has been found experimentally that the diffraction of sound by a a cube is practically the same as that exhibited by a sphere of the same volume".
If anyone has solid reference for the 1/4 lambda/radius principle - I would appreciate that reference.