It’s one of those opinions that get thrown around a lot when a movie comes out where the visual effects, despite a budget that would rival the GDP of some countries, is rather poor. I think I can explain why.
Now, the way I see it, there are two kinds of visual effects: the kind you’re supposed to directly look at and the kind that you aren’t supposed to directly look at. Everything is up on the screen for you to see but for the second kind, your eyes are drawn to another part of the image that the effect is only there to reinforce, not create. Let’s look at two movies, Titanic, and Transformers. Now, despite the large amount of miniatures used to create the illusion of this massive ship splitting in half and going under, there were some scenes where computers were used to enhance the image. Specifically, there were the wide shots of the ship as it was leaving port and all the passengers were waving goodbye. Parts of the ship in that shot were added digitally. But because your focus is on the people, you don’t really notice where the computer effects bleed into the image, if they would bleed at all. In Transformers, the CGI creates the characters, and some entire scenes, themselves. Your focus is on the image created directly by computers. If your suspension of disbelief is held just right, you can ignore some of the imperfections, but something really poor in quality would likely just look ugly and out of place.
Now depending on the desired image or budget constraints, small scale models may be filmed directly instead of using computers to generate the image. The main difference, though obvious, is that one is an actual, physically existing object, and the other is entirely nonexistent, data on a hard drive. That means when you see the image there is a much deeper reaction than compared to CGI. Look at John Carpenter’s, The Thing, compared to the more recently released prequel. Movies that are almost thirty years apart and yet there is a sharp distinction in the atmosphere, and suspense. Partially, that is due to the quality of the screenplay, but for the image itself to stick with you it must be something truly special. In Carpenter’s movie, when the chest opens up with rows of teeth and a severed head grows legs and starts walking away, it’s so completely repulsive because it was a physical object doing this rather than something no one could see or react to. That these things actually existed grant them a whole other level of attachment. Even if the body moving isn’t moving quite naturally, the grotesque nature is why it works so well in horror movies.
Because CGI is an attempt at perfection in an image, our eyes almost instinctively seek out the imperfections in it. When it’s used too much or too little it takes us out of the movie, and when the faults are blatant it almost feels insulting if the rest of the movie can’t pick up the slack. While physical counterparts to CGI have a much more visible seam between the actors, the set, and/or each other, the fact that it isn’t going for that same perfection means that, paradoxically, it doesn’t take you out of the experience. It means for a better time watching. Like in The Expendables, all the blood exploding out of people was digitally added. And it looks very fake and takes away from the experience. But in Django Unchained, squibs go off with practically every gunshot, spraying outright preposterous amounts of blood over everything. Even though the latter film uses effects that would seem faker by comparison, it actually makes the film stronger because with “real” fake blood it’s harder to notice that it’s fake, if that makes any sense. This doesn’t mean that all CGI is bad, nor does it mean that all other forms of effects are good. But having a strong understanding of what would go better where would definitely improve some of the otherwise awkward instances when special effects go wrong.