I think this saying, that eye can not notice anything over about 30fps, came from movies industry. However, computer animation can not be compared to the way movies work (at least old ones), which had a collection of static images and you had to change them at around 30fps so that human eye does not notice the change in frame (the "blinking" effect is gone). In the computer animation, the higher the fps, the more fluid the animation is. Just a simple example. Imagine, that in a game a camera (an imaginary camera, through which the scene is rendered) is turning right. In one second (world time) it will turn 60 degrees. Now, question. How many frames will we see during that "camera turning" animation, if the engine is running at 1 FPS? The obvious answer is one, which will be a scene, rendered when the camera is at 60 degrees from its starting angle. Now just imagine how would this look like in-game. We press the button to turn the camera right, and then after one second (which will feel a very long second) will be turned right by 60 degress. No frames will be showed in between the starting angle and ending angle. I already can imagine that many of us will think that the game is crawling. Oh, and even though the engine will be rendering scenes at 1 fps, we will not see the "blinking" effect, because a monitor refreshes at 60 fps (or 120, depending on the monitor you have). Now, let's imagine the engine runs at 30 fps. Obviously we will see 30 different frames while the camera is turning right (the frame will be rendered every 2 angles). compared to my previous example, the animation will be much more fluid. If the engine ran at 60 fps, we would be seeing 60 different frames while the camera is turning (at every 1 degree angle change), which would make the animation even more fluid. In other words, the more frames per second the engine could render, the more fluid all animations will be. As for DirectX and Opengl, i think there is no much difference. Both are just APIs, and it is up for a developer to use an API in a smart way. I already wrote in some other messages that i was curious to know why the game struggles with the rendering. I used some developer tools to get a log of opengl API calls that CM is doing. From the log i concluded that CM is using a "brute-force" approach when rendering a scene. Several simple optimizations could be done to increase rendering performance. One of those, which every graphics programmer knows, is sorting geometry (before rendering a scene) by material (textures). I just wonder, what would be performance change, if CM engine actually did that. In any case, i perfectly understand that optimizations require time, smart programming, digging through recommendations from nvidia and amd, etc.