Four mid range drivers. The lower mid is confused by comb filtering. Male vocals are around 300 Hz. But vocals consist of lots of partial harmonics at 600, 900 and 1200 Hz etc. At 600 Hz you have a wavelength of roughly two feet. At 1200 Hz you have a wavelength of roughly a foot.
These distances are roughly what you have between each mid range driver and therefore you will get cancellations (as peak cancels trough) and peaks (as peak matches peak) across a broad range of frequencies in the lower mid range. This is called a "comb filter" as it looks just like a comb when you plot amplitude reposne against frequency.
This means your lower mid range frequencies (male vocals)will have a small strong sweetspot directly in front of the speaker and a very narrow dispersion pattern in this frequency range (in an anechoic environment).
When you add the room's side walls, floor and ceiling then the reverberant energy arriving at the listener is likely to be "comb filtered".....some frequencies will suck out completely and others will reinforce and be quite strong.
You ears/brain can normally compensate for some of this and you can actually hear and make sense of reverberant sound, as your ears are about 6 inches apart and get different signals. In fact reverbant sound provides not just SPL energy but also cues as to stereo image and position of instruments in a stereo sound stage.
In your case the reverberant field may be confusing due to all these drivers producing the same frequencies. The fact that you are positioned in front of the speakers does not help the reverberant field.
Try to imagine each of your side walls as a mirror. Then imagine the position of the virtual speaker that you would see in the mirror. The farthest mid range drivers of the virtual speaker will likely be 6 inches to a foot farther from your ear than the nearest ones. Also try to imagine what happens off the floor and ceilings.
Solution => thick carpet floor treatment and very generous speaker toe-in may help (beam the speakers at your listening position from the room corners if necessary). Acoustic treatment of the side wall and ceiling first reflections may also be needed