flat assembler
Message board for the users of flat assembler.

Index > DOS > Drawing lines on 13h mode

Author
Thread Post new topic Reply to topic
wyvern



Joined: 08 Dec 2011
Posts: 27
wyvern
Hi, im trying to draw a line from (x1,x2) to (y1,y2) with the slope-point line formula: y=slope(x-x1) +y1 .

Mi goal is to use this method first doing float-point-arithmetic and later implement more advanced/eficcient ways to draw a line.

For example, considering (x1=4, y1=23) and (x2=13,y2=21), i can plot the points from x1 to x2 one by one like you can see in the attached image, those points are colored in magenta.

But, the green line is the result of mi procedure to draw the same line, dont know why every point of the line is on the y=0 coords. The code is:

Code:
;--------------------------------
; drawLine
;
; [bp - 2] = temp var / x
; [bp - 4] = y
;
; [bp + 4]  = x1
; [bp + 6]  = y1
; [bp + 8]  = x2 
; [bp + 10] = y2
;-------------------------------
drawLine:      push      bp        
                   mov       bp, sp     
                   sub        sp, 4
                                    
                   mov       si, word [bp + 6]
                   fld         dword [si]                                            
                   mov       si, word [bp + 10]      
                   fsubr     dword [si]                 ;y1-y2
                   
                   mov       si, word [bp + 4]    
                   fld         dword [si]                                                     
                   
                   fist        word [bp - 2]
                   mov       cx, word [bp - 2] 
                   
                   mov       si, word [bp + 8]   
                   fld         dword [si] 
                   fist         word [bp - 2]
                   mov       bx, word [bp - 2]
                   fsubrp    st(1), st(0)                 ;x1-x2
                   
                   fdivp      st(1), st(0)                     ;slope=(y1-y2)/(x1-x2)

       
   drawPoint: 
                   cmp       cx, bx                        ;while x1 <= x2
                   jnle        finish
                   
                   mov       word [bp - 2], cx         ;[bp-2]=cx=x
                   fild        dword [bp - 2]                ;st0=x
                   mov       si, word [bp + 8]         ;                    
                   fsubr     dword [si]                    ;x-x1
                   fmul      st(0), st(1)                  ;slope*(x-x1)
                   
                   mov       si, word [bp + 6]     
                   fadd      dword [si]                    ;y=slope*(x-x1) + y1
                 
                   fistp       word [bp - 4]               ;store "y" in [bp-4]
                  
                   push      word [colour]               ; colour
                   push      word [bp - 4]              ;  y
                   push      word [bp - 2]              ;   x
                   call        drawPixel                  ;plot in (X, Y)
                   
                   inc         cx                             ;x+1
                   jmp       drawPoint                                 
                   
  finish:      
                   mov       sp, bp          
                   pop       bp          
                   ret        10  

;data
x1     dd     4.0 
y1      dd    23.0  
x2      dd    13.0   
y2      dd     21.0  
colour  dw   0002h 
    


When debugged, on the main drawLine loop, all the computed Y values for every X are just fine and are passed to drawPixel, the computed values are:

X Y
4 21
5 21
6 21
7 22
8 22
9 22
10 23
11 23
12 23

This are the same coordinates for the magenta manually ploted points, however just the green line appears.

By the way, im using Turbo Debugger and the "Numeric processor" view is not working (at least on DOSBOX), the STx regs appears as empty or unused all the time, it would be really cool to make it work.


Description:
Filesize: 14.4 KB
Viewed: 3171 Time(s)

img1.JPG



_________________
Thanks
Post 11 Mar 2013, 19:25
View user's profile Send private message Reply with quote
revolution
When all else fails, read the source


Joined: 24 Aug 2004
Posts: 17271
Location: In your JS exploiting you and your system
revolution
Perhaps Bresenham's line algorithm would be better for the task?
Post 11 Mar 2013, 20:19
View user's profile Send private message Visit poster's website Reply with quote
wyvern



Joined: 08 Dec 2011
Posts: 27
wyvern
Hi revolution:

Is not exactly a "task", i want to learn from the basis, sure later i will try the Bresenham's algorithm. But this is mi first attempt to draw a line and i want to make it work.
Post 11 Mar 2013, 20:31
View user's profile Send private message Reply with quote
f0dder



Joined: 19 Feb 2004
Posts: 3170
Location: Denmark
f0dder
revolution wrote:
Perhaps Bresenham's line algorithm would be better for the task?

Some years ago I bumped into an article claiming that DDA is actually faster than Bresenham on modern CPUs - don't have a link handy, but it might be worth looking into.

_________________
Image - carpe noctem
Post 11 Mar 2013, 21:15
View user's profile Send private message Visit poster's website Reply with quote
freecrac



Joined: 19 Oct 2011
Posts: 117
Location: Germany Hamburg
freecrac
I loose the focus mixing integer addresses with floating point values on the stack of the program.

; [bp + 4] = offset of x1
; [bp + 6] = offset of y1
; [bp + 8] = offset of x2
; [bp + 10] = offset of y2

Maybe we can simplify the situation if the delivery of the parameter uses named locations instead of involving the stack of the program?

(I think a floating-point-arithmetic for to calculate integer addresses is not the simplest way.)

Dirk
Post 11 Mar 2013, 21:46
View user's profile Send private message Send e-mail Reply with quote
baldr



Joined: 19 Mar 2008
Posts: 1651
baldr
freecrac wrote:
I think a floating-point-arithmetic for to calculate integer addresses is not the simplest way.
Coordinates are still integer, thus rational (∆y/∆x) calculations can be scaled somewhat.

And yes, that's amazing why someone do prefer to use direct [bp+xx] references instead of equs or some macros (LocoDelAssembly or Benjamin).
Post 11 Mar 2013, 22:54
View user's profile Send private message Reply with quote
bitshifter



Joined: 04 Dec 2007
Posts: 764
Location: Massachusetts, USA
bitshifter
some more info about lines in this thread:
http://board.flatassembler.net/topic.php?p=118603#118603
Post 15 Mar 2013, 03:04
View user's profile Send private message Reply with quote
SeproMan



Joined: 11 Oct 2009
Posts: 55
Location: Belgium
SeproMan
wyvern,

Seems nobody actually looked at your code!
What happens in 'drawPixel' is unknown but I suppose you know if it is OK...
Here are some hints for the code you've submitted:


Code:
fild        dword [bp - 2]                ;st0=x    


It's not a dword at this location!


Code:
mov       si, word [bp + 8]         ;                     
                   fsubr     dword [si]                    ;x-x1    


This addresses X2!

Code:
ret        10    


Should be 'ret 8'

_________________
Real Address Mode.
Post 18 Mar 2013, 16:42
View user's profile Send private message Reply with quote
Display posts from previous:
Post new topic Reply to topic

Jump to:  


< Last Thread | Next Thread >
Forum Rules:
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
You cannot attach files in this forum
You can download files in this forum


Copyright © 1999-2020, Tomasz Grysztar.

Powered by rwasa.