• Advertisement

RasterGuy

Member
  • Content count

    4
  • Joined

  • Last visited

Community Reputation

110 Neutral

About RasterGuy

  • Rank
    Newbie
  1. Thank you Lightness, That definitely makes sense and kind of ties everything together. I really appreciate your input =)
  2. Thanks Bob, I think that is what I was trying to come to. That helps a lot, starting to see the forest now, I think.
  3. [quote name='Brother Bob' timestamp='1348700869' post='4984154'] The focal length and aspect ratio are determined from the parameters used in the equation in the book. For example, the aspect ration is [i](r-l)/(t-b)[/i] and the focal length (as you appear to be defining it) is [i]n/(t-b)[/i]. [/quote] Thanks Brother Bob, I guess I need to go over it some more. It seems like, in the book, the near plane is actually replacing the view/projection plane. Is that right? I apologize if this is obvious, but I seem to stumbling on it for one reason or another.
  4. Hi guys, I'm having a little trouble wrapping my mind completely around the perspective projection and the relationship to the view plane. A little background:[list] [*]I'm working on a 3D Software Renderer. [*]I'm reading Lengyel's "Mathematics for 3D Game Programming and Computer Graphics" (I've added a link at the end to a preview in Google Books, has a good chunk of what I'm reading at least, pp. 111 and on.) [/list] So, I understand that a basic perspective projection can be achieved with (obviously the matrix is over kill): [ 1 0 0 0 ] [ 0 1 0 0 ] [ 0 0 1 0 ] [ 0 0 1/e 1 ] Where e is the distance to the view-plane(focal length) as a function of the horizontal FOV. This would correctly project points/vertices, but would not perform any clipping. I know that the OpenGL projection matrix transforms all of the points into homogeneous clip-space (transforming the view-frustum to a cuboid), and makes clipping to the frustum much more simple. (The matrix can be seen on pp. 124 linked below, labeled Mfrustum) I also see how the focal length determines the normals from the frustum planes, as well as the aspect ratio's role in the frustum plane normals. What I don't understand completely, is that, both e(focal length) and the aspect ratio seem to have no contribution to the perspective projection matrix. Is the actual projection a final step after transforming to clip-space (and doing the clipping)? Is OpenGL doing something extra behind the scenes? (recall, I'm working on a software renderer) More over, what is the view-plane's relation to the view-frustum? I hope my question makes sense, if not I will be happy to rephrase or elaborate. I appreciate and knowledge you can share. Book Preview: [url="http://books.google.com/books?id=bfcLeqRUsm8C&pg=PA115&lpg=PA115&dq=frustum+plane+normal&source=bl&ots=FqTvf6tUfB&sig=habQb5--rTggjW1ilsbDvv-4eD8&hl=en&sa=X&ei=5nxjUO_fO-XW2AX2zIDAAg&ved=0CEYQ6AEwBDgK#v=onepage&q=frustum%20plane%20normal&f=false"]http://books.google.... normal&f=false[/url]
  • Advertisement