(NOTE: This is in addition to Professor Hausner's A2 FAQ page.)
Questoin1:
Answer:
OK, what matrix can you use to change (1,0)frame_A to (1,-1)frame_O?
Answer: Let's say we have a coordinate
C which looks like
| y
|
|
x ------ frame
C
What is (1,0) of frame_C in frame_O? -- it is (-1,0)frame_O
What matrix can do this transformation?
[0] = [
a? b? ] [0]
[1]frame_O
[ c? d? ] [1]frame_C
[-1]
[ a? b? ] [1]
[0 ]frame_O = [
c? d? ] [0]frame_C
Answer: Take a look at the tutorial
question #3 in
"Tutorial: Geometric Transformations 3"
http://www.dgp.toronto.edu/~ah/csc418/fall_2000/tut/tut_gtrans3.html
The TA should cover this tutorial exercise's solution by Wednesday,
Oct 18, 2000.
Answer: If you are running an animation
which displays
60 frames/sec. From each frame to the next, the object you animate
would change. When the object is modelled using the triangles, for
each frame, the triangle's vertex locations would change. For each
frame, you have to re-compute and re-render and re-project the triangles.
Question 5:Answer: No. Please see the Lecture Notes on "The Viewing Transformation"
http://www.dgp.toronto.edu/~ah/csc418/fall_2000/notes/viewing.html
Under the heading "Building M_cam", the formula for vectors i, j, k uses P_ref and P_eye.
Answer:
The pixel coordinate is in the DCS (Device Coordinate System),
which is related to the flat 2D display below.
For example, if your display has resolutin 5x5
(i.e., width=5, height=5), then you shoot 25 rays from your eye,
one for each pixel.
If your display has resolution 10x10 (width=10, height=10), then
you shoot 100 rays from your eye.
Answer: Yes.