This is a very overdue blog post.
I normally do not go to movie theaters (or "cinema" in British English) to see movies (or "films" in British English). I had never seen a 3D movie. Hence, when Jurassic Park 3D came out, I gave it a try, and I learned something interesting about vision science. I was more doing an experiment than watching a movie. Again, it is about how we sense depth, a useful topic in computer vision and robotics.
Note that I am not an expert of computer vision or vision science. Please comment for anything I said wrong or naive.
According to the theory that I learned in physics class, we sense depth because of the difference between the views from our two eyes. Hence, in order to make a 3D movie, you can use two cameras (usually forming a small angle) to capture and then play such that the two eyes see different scenes, e.g., via polarized glasses. However, this is exactly the case.
Note that I do not think Jurassic Park 3D was made using two cameras. I think they used computers to generate different views and give audiences a feeling of 3D.
First of all, our eyes have the ability to adjust focus (even for people wearing prescription eyes all the time, like me) while cameras cannot focus on all objects all the time. When the camera is focused on dinosaurs, objects very far or close are blurry. When it is 3D, it becomes very annoying. I felt that I was seeing things as usual. But my eyes just couldn't adjust to turn those blurry objects into clear ones. It is really funny that sometimes I felt I was watching a 1920s movie where actors were playing in front a screen playing a background movie (e.g., sitting in a fixed car with moving background to mimicking driving the car).
Second, our eyes really don't have good ability of depth sensing for some objects (at least this is my theory), such as fog or hair. I feel that our eyes have good sense of depth for large piece objects. I guess the production team also had difficult time for objects like fogs or hair - how do they know the depth of each water drop in the fog? It is easy to say that "the actor is in FRONT of the dinosaur" by common sense.
The last thing I notice was that our eyes have the ability to tell depth by how big the difference between two eyes is. In the end of the movie, the credits popped up. I could feel that the letters were "floating" in front of the background. (Actually my roommate complained that this was the only part that looked 3D to him.) How do we do this? I assume that the difference between two eyes are proportional to the depth and we can convert from difference to depth.
In conclusion, I think it is very difficult for us to make or see a real 3D movie, due to the focus issue I mentioned above. The only possible way to do so is to fully use computers to generate and render - such that every pixel is not blurry. This is very possible for Sci-Fi movies which heavily rely on computer graphics.
Again, a very naive blog post. Comments and critics are welcomed.