Stewart’s piece gets really close, but doesn’t quite touch on the issue of composition of the image in games. Outside of cutscenes, it’s immensely difficult to be evocative with imagery while placing full camera control in the hands of the player. If we’re talking cutscenes, the manual labor that Stewart refers to does then indeed become a significant issue. For everyone but the largest studios, doing camera work in games is usually painstaking and extremely limited out of the box, either leaving devs to spend time creating better solutions or being okay with “good enough.”
Being evocative and coherent with visual language during gameplay is a whole other kettle of fish. I’d argue that the problem that games have with visual language is inherently kinesis. Any camera that moves reasonably fast for player interaction is not the kind of camera that is useful for the vast majority of the storytelling. Cameras in games are function first, because they can mar the player experience so badly when they aren’t.
There’s a reason why early RE games and Silent Hill get so much praise for how evocative they are. There’s that common idea that the limited camera in those games is frustrating but part of the experience, and I think you can extrapolate that in reverse to say that freely player operated cameras are the antithesis of being evocative in the way those games are. You can design as much as you want for the player to look where you want them to, but I think asking players to be cinematographers as part of their experience is a bridge too far for most.
Maybe that reads as pessimistic, but I think I arrive at a similar place as Stewart does, in that games need to embrace their gamey-ness to be visually successful. We can learn lessons from what other mediums do, but so much of that goes out the window even when we consider what types of cameras are deemed conventional depending on genre of game.