Why not horns?

Cool! Thanks Mapman.
Ima check that out.

Prez, I hear what you are saying but I think the manufacturers are using terminology to confuse the layman in an attempt to carve out a unique slot in the marketplace. In other words, marketing BS. A change in arrival time is a change in phase no matter how it is done. They are synonymous. If you read Thiels papers they admit as much.

Either the different frequencies arrive at the ear with the same timing relationship they had when they were put on the recording or they do not. If not it could be that the drivers aren't aligned, That a digital or electronic delay was employed, or there is a phase shift through some reactive device like a crossover.

If they want to distinguish phase shifts caused by crossovers as phase coherency since they are frequency dependent and those caused by driver alignment as time coherency since they are not frequency dependent I'm on board with that, but time alignment and time coherency are the same thing.

By eliminating all reactive components after my amps (no crossover what so ever) and implementing the crossovers digitally before the amps I should only have phase shifts caused by the reactance in the drivers and hopefully the bulk of that is outside the band of frequencies they will be fed.. Each band can also be digitally shifted in time so they should be close to being time and phase coherent to use Thiel's terminology. The purists cringe when you talk about digital processing but so far so good.

One point of clarification, Even first order filters cause phase shift as you approach the cutoff frequency. Thiel claims that they have achieved equal but opposite shifts from the drivers above and below the cutoff so they cancel.

The phase shift is kept low by using very gradual (6 dB/octave) roll-off slopes which produce a phase lag of 45° for the low frequency driver and a phase lead of 45° for the high frequency driver at the crossover point. Because the phase shift of each driver is much less than 90° and is equal and opposite, their outputs combine to produce a system output with no phase shift and perfect transient response.

I'm still trying to wrap my head around that one. If one driver produces a sound shifted in time so it occurs slightly earlier than those in the passband and another produces the same sound slightly later how can that add up to no time change?

.

Mapman, I'm not sure that Ohm's current non-bending wave non-Walsh drivers qualify.

"Mapman, I'm not sure that Ohm's current non-bending wave non-Walsh drivers qualify."

I guess you'd have to take that up with John Strohbeen to know for sure.

Herman, there certainly has been much debate over the merits of time coherency.
Just to clarify one point first. The way they describe it, time "coherent" demands time "alignment". But the reverse is not true.

You're point about phase is certainly correct. But I think there is also a distinction with the "when" of phase. For instance, it's either the 2nd order or 4th order filter (I can't remember which) that is phase coherent in the crossover region. Which is to say, all the peaks and valley's line up. The difference is that some frequencies started before others. The delayed ones are a full cycle behind. They are still phase coherent but not time coherent.

The easiest way I found to grasp it was to note that TIME coherent was at the pinnacle of the hierarchy. Indeed, most people, when refereing to this type of design, say "Time AND phase coherent". But it's redundant to say that since time coherent demands phase coherent. It demands both time alignment and phase alignment. These other two could be achieved individually by means of physical placement or filter makeup. But both, on their own, was only part of the story. Anyway, thats the best way I know to describe it.
Also, I think phase alignment and coherent ARE the same thing. But I'll have to think about that a little more to be sure. No, I think its true. anyway..........

First order filters do have phase shifting but the filter circuit as a whole compensates. The current lag in an inductor is the same degree (hopefully) as the voltage lag in a capacitor. But, I'm going to have to think about that some more before I try to go any further. I want to make sure I don't mistate anything.

The one measurement that is the arbiter of all this is the acoustical step resonse. Two speakers that easily show this is Wilson and Dunlavy. Both are time "aligned".
Wilson: http://www.stereophile.com/floorloudspeakers/920/index6.html
Dunlavy: http://www.stereophile.com/floorloudspeakers/162/index10.html

Looking at the step responses. Though, be careful when reading what JA has to say. I've noticed that over the years he throws all the descriptions around without much measure of consistency.

The short of it is this. The only speaker design that shows a step response (and therefor the most accuracy in the time domain) that mimics the input is a "time coherent" design. Ala Dunlavy, Vandersteen, Meadowlark, etc etc. They are the only ones that approach the right triangle form.

The merits of this? Well, like I said, Hotly debated.

I gotta run. More later.
Cheers!