1. 3D (isometric)
Hi again!
Thanx for the URL!
I made a little procedure to do +-3D Isometric (Starcraft-like) texturing,
But it's very very slow and I need to speed it up!
Can You help me please ??
*********convert* (to use with mode 259)
global function convert(atom x, atom y, atom z)
atom X,Y
X=(x/400)*400+(y/400)*400
Y=(1-x/400)*300+(y/400)*300-2*z
return {X,Y}
end function
*********map3d* (it uses "convert")
include image.e
global procedure map3d(sequence bitmap, integer x1, integer y1,integer
z1,integer x2,integer y2,integer z2
,integer x3,integer y3, integer z3)
atom dis1, dis2, cons1, cons2, cons3, addx1, addy1, addz1, addx2, addy2, addz2,
u1, v1, w1, u2, v2,w2
sequence bmp, dumb1
bmp=read_bitmap(bitmap)
all_palette(bmp[1]/4)
dis1=sqrt((x1-x2)*(x1-x2)+(y1-y2)*(y1-y2)+(z1-z2)*(z1-z2)) -- = Distance between
(x1,y1,z1) and (x2,y2,z2)
dis2=sqrt((x1-x3)*(x1-x3)+(y1-y3)*(y1-y3)+(z1-z3)*(z1-z3)) -- = Distance between
(x1,y1,z1) and (x3,y3,z3)
cons1=2 -- I should'nt use cons1 & cons2, it slow up a lot, but if I don't do
that, the textured space is very ugly
cons2=2
cons3=length(bmp[2])/dis1/cons1
addx1=(x2-x1)/dis1/cons1
addy1=(y2-y1)/dis1/cons1
addz1=(z2-z1)/dis1/cons1
addx2=(x3-x1)/dis1/cons1
addy2=(y3-y1)/dis1/cons1
addz2=(z3-z1)/dis1/cons1
u1=x1
v1=y1
w1=z1
for i = 1 to length(bmp[2]) by cons3 do
u1=u1 + addx1
v1=v1 + addy1
w1=w1 + addz1
u2=u1
v2=v1
w2=w1
for j= 1 to length(bmp[2][i]) by length(bmp[2][i])/dis2/cons2 do
u2=u2 + addx2
v2=v2 + addy2
w2=w2 + addz2
dumb1=convert(u2,v2,w2)
pixel(bmp[2][i][j],dumb1)
end for
end for
end procedure
*********An Exemple* (change the bitmap!)
include get.e
integer key
if graphics_mode(259) then end if
map3d("leeloo.bmp", 1,1,1,100,1,1,1,100,1)
key=wait_key()
********************
2. Re: 3D (isometric)
>Hi again!
>Thanx for the URL!
>I made a little procedure to do +-3D Isometric (Starcraft-like) texturing,
>But it's very very slow and I need to speed it up!
>Can You help me please ??
>*********convert* (to use with mode 259)
>global function convert(atom x, atom y, atom z)
> atom X,Y
> X=(x/400)*400+(y/400)*400
>
> return {X,Y}
>end function
>X=(x/400)*400+(y/400)*400
isn't this the same as X=x+y ?
>Y=(1-x/400)*300+(y/400)*300-2*z
could simplify to Y = (400 - x + y) * 0.75 - z - z
Yikes! I hope my math is OK!!
michael at igrin.co.nz
3. Re: 3D (isometric)
- Posted by MrSmily at AOL.COM
Feb 10, 1999
>>Hi again!
>>Thanx for the URL!
>>I made a little procedure to do +-3D Isometric (Starcraft-like) texturing,
>>But it's very very slow and I need to speed it up!
>>Can You help me please ??
>>*********convert* (to use with mode 259)
>>global function convert(atom x, atom y, atom z)
>> atom X,Y
>> X=(x/400)*400+(y/400)*400
>>
>> return {X,Y}
>>end function
>
>
>>X=(x/400)*400+(y/400)*400
>isn't this the same as X=x+y ?
>
>>Y=(1-x/400)*300+(y/400)*300-2*z
>could simplify to Y = (400 - x + y) * 0.75 - z - z
>
>Yikes! I hope my math is OK!!
>
>michael at igrin.co.nz
Actually, I got something like this for Y:
Y=(1-x/400)*300+(y/400)*300-2*z
Y=300-300*x/400+300*y/400-(z+z)
Y=300-.75*x+.75*y-z-z
Y=300+.75*(y-x)-z-z
MrSmily at aol.com
4. Re: 3D (isometric)
Gwenael wrote:
>Hi again!
>Thanx for the URL!
>I made a little procedure to do +-3D Isometric (Starcraft-like) texturing,
>But it's very very slow and I need to speed it up!
>Can You help me please ??
<snip>
First of all I do not think your routines generate, strictly speaking,
truly isometric views.
Apart from that, they are also inefficient (slow) and often produce
imperfect images. Michael, another New Zealander, already took out his
bloody big bush knife and hacked off about three percent of fat from
the monster. Not enough.
Last night I spent several hours on a slightly different version of
pseudo-isometric tools. They are typically 50 to 100 times faster than
yours, but they are also very clever, finely balanced set, so save the
original copy before you start hacking into it.
Feed your bitmap into the iso() function first. It produces a raw
rectangular image that in turn has to be processed by the second
routine, iso_display() procedure. It will spit out the final, already
trimmed image. If you use these two routines with a good set of tools
to suitably massage the original image (you can try my map.e,
somewhere on the Recent Contribution Page or in the Archives, for
reasonably fast linear/non-linear scaling routines), you may have a
start of a workable system. If I find a free evening or two in the
next few weeks, I might have go at it myself, but don't hold your
breath... Enjoy. jiri
-- file : iso.e
-- author: jiri babor
-- date : 99-02-10
-- email : jbabor at paradise.net.nz
-- topic : fast pseudo-isometry
without type_check
global function iso(sequence bmp)
sequence s
integer m,n
m = length(bmp[1])
n = length(bmp)
s = repeat(repeat(0,m+n),m+n-1)
for y=1 to n do
for x=1 to m do
s[n+x-y][x+y] = bmp[y][x]
end for
end for
return s
end function
global procedure iso_display(integer x,integer y,sequence s)
sequence t,u
integer di1,di2,i1,i2,L
L = length(s[1])
i1=length(bmp)+1
i2=i1
di1 = -4
di2 = 4
for r=1 to length(s) by 4 do
t = s[r]
u = {}
for i=i1 to i2 by 2 do
u &= t[i]
end for
pixel(u,{x+(i1-1)/2,y})
i1 += di1
if i1<1 then
di1 = -di1
i1 += di1
end if
i2 += di2
if i2>L then
di2 = -di2
i2 += di2
end if
y += 1
end for
end procedure
