@tomic601 , An engineer does not need ears to develop a phase coherent loudspeaker, just a measurement microphone.
In order to be accurate (I did not say sound good) a speaker has to start with a flat frequency response and be time coherent which as unsounds relates assures phase coherence. Unfortunatly, for speaker designers this has to include the room the speakers are set up in. For state of the art accuracy one has to be able to adjust frequency response so that the speakers are flat from 20 Hz to 20 kHz. This also assures that both channels have exactly the same frequency response which is very important for imaging. If a subwoofer is added one has to make sure they are time coherent with the main speakers. The only way all of this can be done is with digital signal processing. Afterwards the frequency response can be adjusted so the system sounds good to the owner. I like mine up 5 or 6 dB at 20 Hz and down 6 dB at 20 kHz. Done this way the difference between speakers is based on the way they radiate into the room and what kind of source they are point vs line.
Generally, speakers with controlled radiation patterns sound better because they cause less room interaction. Horns, dipoles and line arrays come to mind. Omnidirectional speakers require much more room treatment and produce a smaller soundstage like sitting at the back of the hall. This can sound quite natural depending on the recording.