rogerh@arizona.UUCP (Roger Hayes) (10/18/84)
On the 24-to-8 issue: dithering is quite simple and produces good
results. The technique can be derived starting from the 2 pages in
Foley and Van Dam. I usually dither to 3-3-2 bits (rgb) when I have an
8-bit frame buffer -- does anyone take the trouble to divide the target
color space evenly?
On another issue -- I need a pointer to literature on incremental (ie,
generalized Bresenham) techniques for drawing quadratic curves. I have
a raycaster working (arbitrary quadric shapes, boolean combinations), but
it's deathly slow. To speed it up, I would like to use the information
from one ray to determine the next; I need a quick way to determine the
solution for z of
A*z^2 + B*z*x + C*z + D*x^2 + E*x + F >= 0
at (x+1), given the solution at x. "Quick" means faster than substituting
the new x value and solving directly.
Roger Hayes
University of Arizona
rogerh%Arizona@csnet-relay.CSNET
{ihnp4, decvax, ucbvax}!arizona!rogerh by UUCPgwyn@brl-tgr.ARPA (Doug Gwyn <gwyn>) (10/20/84)
> I need a quick way to determine the solution for z of > A*z^2 + B*z*x + C*z + D*x^2 + E*x + F >= 0 > at (x+1), given the solution at x. "Quick" means faster than substituting > the new x value and solving directly. (You must mean = 0, not >= 0.) Sorry, Roger, after cleaning up the statement of the problem, I find that you must solve a quadratic equation no matter what. The best you can do is to take some shortcuts in evaluating the coefficients of the quadratic.