TLS Guy thanks,
If I understand you correctly, the bottomline is to reverse the polarity of the Center speaker as Audyssey advises? True?
Sent from my SM-G975U using Tapatalk
No that is not true. The problem is that ONLY the center mid is out of phase with the left and right mains. The woofers and tweeter of the center speaker are in phase with the left and right mains. So only the center mid needs its phase reversed.
However you really do need measurements to do this. The problem is created by the phase shifts of analog crossovers, both passive and active. Only digital crossovers with DSP can avoid this problem.
The issue is that if you have identical speakers that are three way, with second order LR or Butterworth crossover for that matter, then without reversing the phase of the mid, then the mid is out of phase with woofers and tweeter, but only in the crossover region, where the drivers overlap. This is because the phase shift is 90 degrees for each section and so 180 for the woofers and mid at overlap. So this causes a null in the crossover region, but only in the region where the drivers overlap. The orthodox treatment is to reverse the phase of the mid, to avoid the null. However if you have two way left and right mains, then the mid of a center with reverse phase will be out of phase with the left and right mains, across all of the pass band out of the crossover regions.
The problem is compounded by that fact the the degree and sharpness of the null is highly variable and you have to see what it looks like.
In my case, the null was narrow, and mainly only measurable on, and close, to the axis of the center. So tolerating the null in the crossover region was not a problem in practice.
Usually second order cross overs are the most popular as there is a constant voltage solution, the LR topography. The first order, with 90 degree phase shift generally gives too much power to the drivers out of their pass band, and there are lobing issues at crossover. The third order, with 270 degrees of phase shift, only has a Butterworth constant power solution, also with axis lobing issues at crossover. The fourth order solution has both constant voltage and power solutions, and no lobing issues. However, it has 360 degrees of phase shift, which puts the drivers a whole cycle out of phase at crossover. In practice it usually sounds OK, but the impulse response is not pretty. The big issue is that the part count, and expense, of a fourth order LR crossover is high, and you won't build that for the price you paid for that speaker.
So your solution is likely to reverse the phase of the mid on the center speaker only. However as stated, this should be done, with full data on the effect of it.
In my studio system, I was very careful to carefully design the crossovers of the front three to be compatible as regards to phase and other matters also.
Here is my three way center with the mid driver not phase reversed. The deep null is at the top end of the driver over lap. The black line is the axis response, off axis the response is good.
The room responses are actually very good all around this large space.
Accepting the problem in the bass mid crossover region sounded far preferable to having the mid out of phase with the bass mids of the left and right speakers. So when the center followed normal practice of phase reversal you could not get the center to blend properly with the left and right speakers. The system sounded much better with the center mid in phase with the left and right speakers and accepting the problem in the crossover region.
This is just one more problem with using speakers across the front, that were not designed specifically to work with each other.