Printable View
Ok, since alot of the games today (next gen) are using motion blur, i have some questions on what exactly and how motion blur is achieved. I know a little of what it might be. Does it consist of some type of alpha layer, and then just keep adding alpha layers? I'm not sure how this is achieved. So if anyone knows, please tell me. Note that I don't want specific code since I am no where near that advanced. But just curious. :p
OMGPOSTSTEALER!Quote:
Originally Posted by Sentral
I was going to post something like this because I was wondering yesterday how this is achieved. It adds quite a lot of depth to the atmosphere when done "right", like in CoD2 if you have a mortar land near you and it blows you off your feet, alá the D-Day landing map.
Most of the motion blur Ive seen is a post process on the rendered scene. If you want specifics than google it or steal the source for the projectoffset codebase.
With shaders:
Frames are usually stored in a renderable texture, and blended with the current frame.
Without shaders (OpenGL, not sure about D3D):
Frames are stored in the accumulation buffer, and blended with current frame.
And that concludes everything I know about motion blur. ;}
Accumulation buffers are used as psychopath said. It is the same process as the feedback in your average Windows MP3 player. A lot of post-processing, a lot of GPU tricks, a lot of headache, and if you ask me a lot of effort to give your gamer's seizures.
Yeah, it's generally not a good idea to use motion bluring excessively. Usually in cinematics or cutscenes it can add a nice touch, or to simulate dizzyness. For example, in Gran Turismo 4, when you switch the camera view to in-car and you crash, motion blur is used to swirl and distort the screen, to loosely simulate what the driver sees. There are probably other scenarios where you would use MB, but I can think of any more at the moment.