| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 27.1 | graphics info | CIMBAD::POWERS | I Dream Of Wires - GN | Thu Apr 07 1988 17:25 | 17 |
|
re .0
I have some stuff that does rotation, scaling, etc. of three
dimensional objects. The way this is done is with transformation
matricies. If you are interested in the theory behind this I suggest
the graphics book by Foley & Vandam (sp). I have just a few finishing
touches to put on a program for the st which allows the user to
type in a function y in terms of x and z, and plots it on the screen
in three dimensions, doing hidden line removal. Actually doing
hidden line removal is pretty easy on an st. you can break your
object up into polygons, and use the vdi routine polyline, and have
it fill in the polygon as it is drawn. if you draw the object from
the back to the front, all the lines behind surface get erased.
Bill Powers
|
| 27.2 | Principles off Computer Graphics | UTROP1::TRAMONTINA | | Fri Apr 08 1988 03:54 | 15 |
| re .0
If you are intrested in Graphic algorithms you might want to read
a book called "Principles of Computer Graphics" by Leendert Ammeraal.
He is a teacher at a Techical University in Holland. The book presents
programms written in C, the programms are highly machine independant.
They just need some routines like Plot and Move.
Some people (mostly the man's students, as me) think the book isn't very
easy to read but what you all want is in there.
The book is published by an American publisher so it shouldn't be
too hard to find it.
greetings from a sunny Europe,
Renato.
|
| 27.3 | Thankyou. | CHEST::BADMAN | Better Late Than Forever | Mon Apr 11 1988 07:58 | 6 |
| Thanks for the ideas ... I never thought of polyline as an automatic
HLR routine!
Jamie.
|
| 27.4 | Start at the front | WELMTS::GIBSON | Alan Gibson @WLO | Tue Apr 12 1988 09:47 | 10 |
| re .1:
For plotting the function z=f(x,y), it's actually more efficient
to plot from the front than the back. The principle is that for
each calculated point on a screen column (ie x-coordinate), note
the maximum plotted y-coordinate. If a new y-value is higher, plot
it and increase the maximum. If it's lower, don't plot it.
I can give you a pointer to an algotithm if you're interested.
It's in one of the books published by McGraw-Hill.
|
| 27.5 | plotting methods | CIMBAD::POWERS | I Dream Of Wires - GN | Tue Apr 12 1988 10:55 | 13 |
|
RE .4
Yes, I was aware of that algorithm, but that means i would
have to write a routine to generate all the points between two points
on a screen, make the test, and set each point on the screen
individually. I figured the way I did it was quicker, then setting
each point individually. I believe cause essentially the linea
routines would have to get called for each pixel (very expensive)
as to once for each line. Also, it was much less coding for me
to do.
Bill Powers
|