Prominent among the current promotional thrusts made for some speaker systems is an emphasis on "phase linearity." The phrase refers to the relative times it takes for different sounds from the same source to reach the ears. The complex were wave pattern of a musical signal entering a speaker system from the amplifier should be maintained as it emerges from the speaker as sound. To the extend that it does, you have "phase linearity." To the extent hat it does not, you have "phase distortion."
On the face of it, phase linearity seems like a good idea (although it certainly is not a new one). On the other hand, many audio insiders feel it represents either an undue emphasis on incidentail performer or something of a snow job. My current feeling leans to the latter. Here's why.
The most cosmetically obvious "design attempt" to achieve phase linearity in some recent models is the repositioning of mid-range units and/or tweeters somewhat behind the woofer in order for all the voice-coils to be in the same vertical plane. This way, the argument goes, all sounds remain in correct time/phase relationship because they all move toward the listener from the same starting line.
How much sense this makes is debatable. Scientific studies have shown that in the critical frequency region of about 5,000 Hz (which is considered the maximum area of sensitivity to phase errors), time delays of up to three milliseconds/delay pattern corresponds to a physical distance of one meter - so the emphasis on correcting for phase errors by moving the driver back a few inches seems like a delusion. Further, the size, shape and mass of different drivers, not to mention the way in which they may be loaded to a room (some directly, some perhaps by vents) can frustrate any "phase coherence" when all the voice-coils are vertically aligned. Moreover, it seems obvious that, at best, the alignment of various drivers in a given speaker system can produce phase coherence at only one listening angle from that system.
In any event, voice-coil positioning alone will not achieve the desired effect. Additional design factors must be included - among them, careful design of the crossover network, and possibly the use of "filler" drivers. Even then, the audible effect may be negligible or apparent from only one listening spot. A widely accepted test for a phase alignment of a speaker system is its ability to reproduce a "square wave" (a complex signal very rich in harmonics, with an almost instantaneous rise-time and similar decay). Because of normal reverberation and reflection properties of a room, it is virtually impossible for any speaker system - regardless of how little or how much design effort was made to achieve phase coherence - to reproduce that square wave perfectly except at one spot, or perhaps at a few specific spots (literally, points in space).
The basic hallmarks of good speaker design remain: Uniform power output at all frequencies; reasonably high sound-pressure-level (SPL) capability to reproduce the dynamic range of program material; low harmonic and IM distortion; good transient response (no "hangover" or "ringing" effects); and good dispersion (no "beaming of the treble range).
As one speaker expert has put it, with these attributes you can forget about "phase linearity" unless the system is so horrible that it produces a displacement equal to one meter or more - which would be something like a speaker system whose tweeter stuck out more than 39 inches in front of the woofer.
Don't let if faze you.