Ok, I have to speak in order to stop people like Stevecham from making incorrect claims.
Sound speed is definitely freq dependent. But in order to truly make a speaker time correct, you will need to have infinite number of drivers to cover each freq and align them on different time axis. This is is not possible since each driver covers a certain freq range, not one freq, and the designer "picks" a freq to time align it. Not the whole freq range since it's not possible. Therefore reviewer tests the same way, one freq on tweeter axis. So if the reviewer doesn't test the designer's freq, it will not appear time correct.
As for phase correct, the worst thing to do is to use anything higher than 1st order crossover like in Thiel. This is a common sense and often practiced by well regarded manufacture including Sonus Faber and Dynaudio. I challege Stevecham to solve an op-amp in 4 pole equation, it will involves 1st to 4th order slope and will show it's impossible to keep phase correct with high order x-over. I hold a MS in EE and have been in semicondutor field for 10+ years, it's almost embarrassing to hear people like Stevecham making unsound claim without any technical background.
And thank God there is people like Stevecham on earth to keep company like Thiel alive who picks equipment by one data set, not by his own ears.
And this site is great for people who listen with their ears, not for people who bash other manufactures without knowing what they are talking about.