I think the main problem with voices , at least for me, is variety. If the game had for example: islander, gritty, deep, young, old, middle age, tone influence, demeanor, softcore, hardcore, etc I don't think it would be as much of a problem.
Question is - could technology from Watchdogs: Legion somewhat address it? I didn't play the game, but from what I understand they generated many character voices from few pre-recorded take due to voice modulation software. I don't know how effective it was - I heard it was mixed, sometimes it worked sometimes it didn't.