I don't disagree that speakers can and often do have a lot to do with it, but I wouldn't conclude that is the only problem for everyone, or even that it is the paramount issue.
I suspect that Gene has pretty good speakers, yet in his video that Shadyj linked to (see it again
HERE) he said he often has trouble hearing dialog and that he increases the center channel a couple of db due to bad mixes, etc. He and Matthew spent well over an hour talking about issues related to dialog intelligibility. If it were only a speaker issue, I think the video would have been 5 minutes. And take a look at the video at around the 48 minute mark. The first item on the list of things that don't work for purposes of making dialog better is replacing speakers. He provides further clarifications and notes that speakers are very important (and also Gene contradicts the point later by saying to get good speakers), but he also says that too many people run out and buy new speakers when that is not necessarily the problem.
Anyway, as I said, I don't disagree with you entirely. But I think there are other things going on as well that need to be considered.
Gene I believe uses RBH speakers. Gene probably won't want to hear this, but none of the RBH speakers are designed correctly for center channel use, and will exhibit the problems you speak of.
These are the problems that have to be solved for a speaker to work properly as a center channel speaker with intelligible dialog every time.
1). This is most important. No serious aberrations in the time and therefore phase domain can be permitted in the speech discrimination band. That is 400 Hz to 3.5KHz and preferably 4KHz.
2). A correct lobing pattern. That means NO horizontal MTMs.
As far as I am concerned these are absolute rules that can not be broken and achieve optimal speech intelligibility. No breaches can be permitted at all.
This can only be achieved in the following ways.
1). A good full range single driver. I used that very successfully in a previous system.
2). A good coaxial driver with minimal time aberration in the crossover.
3). In a spaced driver system, no crossover on the speech discrimination band unless it is an active speaker, with DSP to correct time alignment.
There is no way I no of to do this.
So in my main system the center is a coaxial design. It has excellent speech clarity.
The family room system is 2.1, but the midrange is crossed at 400 Hz and 4KHz. Speech clarity is excellent. Zero trouble with dialog.
My wife's in wall system which is a 3.1 system with a three way center with vertically aligned mid and tweeter. The mid is crossed at 400 Hz and 4KHz. Again speech clarity is excellent. Neither my wife or I have a problem with speech/dialog intelligibility on any of these systems on any program.
This problem is solvable, and above is the road map.