This topic contains 4 replies, has 0 voices, and was last updated by  Anonymous 6 years, 5 months ago.

Viewing 6 posts - 1 through 6 (of 6 total)
  • Author
    Posts
  • #203316

    Hi Ron,

    I am currently reading my way through the book (which I am absolutely loving may I add) and you brought up an interesting point about motion blur where you say that contrary to popular belief there is no in focus leading edge of a moving object.

    How is a symmetrical motion blur produced when the leading edge is the furthest forward part of the pixel data, in otherwords the object can’t physically travel ahead of itself (on the leading edge) to expose that part of the frame there?

    I hope that makes sense, after reading what you put I started scratching my head a little and wondered how thats actually possible.

    I apologise if you explain this later in the book I am only up to chapter 4.

    Kind Regards,

    Dave Schott

    #218637

    Hi Dave,

    I’m obviously not Ron but I have had issues understanding this in the past myself.

    I think the easiest way to understand how motion blur works is to simulate how CG fakes motion blur. So go into something like photoshop and do the following. Make a solid circle of a certain colour and then reduce it’s opacity to say 25%. Think of this circle as a motion blur sample.

    Now duplicate this circle 4 times and offset each one further along to one side so it is still overlapping somewhat with the previous one. If you duplicate the circle 4 times with equal spacing and each offset is a fraction of the circle diameter you should see a fully opaque section of the “blurred” sphere in the center surrounded by trailing and leading edges that fall off in transparency.

    With only 4 samples it’s obviously going to be very rough example of motion blur but might give you an idea of how the principle works.

    Think of each circle as a representation of what the camera sees of the sphere at that particular point in time during that frame/exposure. On a real camera the samples in time (exposure) and the motion increments of the sphere would be infinite which leads to a very smooth blur.

    Essentially, the furthest forward leading edge of the motion blur IS the latest sample of the sphere’s position at the end of it’s movement during that frame, but because it may only be exposed for a very short amount of time (25% in the example) and therefore have little additive effect overall it will be less visible and therefore lead to the illusion that the sphere is really in the center of the blur where the effect is most opaque.

    So the sphere is never traveling ahead of itself you are simply seeing a history of it’s movement during that frame recorded in an infinite number of fractional exposures (not sure if that’s a correct way to explain this :S) that accumulate over the length of the movement.

    MotionBlur%20Theory.jpg

    I really hope that made sense and I didn’t confuse you any further, in any case I look forward to hearing from Ron himself and see if he can set me right or confirm in a more concise way how this whole thing works.

    Regards

    Linus

    #218636

    Hi Linus,

    that was a really concise explanation, it really makes sense actually now you have broken it down. I was imagining the denser core part to be the actual object with some unexplained blur in front when actually the front part IS the object and the denser core is just the overlap and multiple exposures of the samples.

    Thanks for putting in the effort with those drawings to explain, much appreciated.

    Dave.

    #218638

    no worries I’m glad it made sense 🙂

    Linus

    #218639

    Shona Hallen
    Member

    Thanks Linus

    #218640

    Anonymous

    good sounds…
    can you share any site for best seo

Viewing 6 posts - 1 through 6 (of 6 total)

You must be logged in to reply to this topic.