Java-Gaming.org Java4K winners: [ by our judges | by the community ]         
Featured games (67)
games approved by the League of Dukes
Games in Showcase (∞)
games submitted by our members



News: Read the Java Gaming Resources, or peek at the official Java tutorials
 
    Home     Help   Search   Login   Register   
Pages: [1]
  Print  
  Fastest method of referencing all pixels in a line?  (Read 1147 times)
0 Members and 1 Guest are viewing this topic.
Offline counterp

Full Member
**

Posts: 235
Medals: 11



« on: 2011-07-15 13:31:00 »

Currently, to find all pixels in a line I am using Bresenham's line algorithm. I converted the method from pseudo-code (on wikipedia) to java, but it's rather slow. Here is the code:

1  
2  
3  
4  
5  
6  
7  
8  
9  
10  
11  
12  
13  
14  
15  
16  
17  
18  
19  
20  
21  
22  
23  
24  
25  
26  
27  
28  
29  
30  
31  
32  
33  
34  
35  
36  
37  
   private void getPixelsInLine(int x0, int y0, int x1, int y1) {
      boolean steep = Math.abs(y1 - y0) > Math.abs(x1 - x0);
      if (steep) {
         int tmp = x0;
         x0 = y0;
         y0 = tmp;
         tmp = x1;
         x1 = y1;
         y1 = tmp;
      }
      if (x0 > x1) {
         int tmp = x0;
         x0 = x1;
         x1 = tmp;
         tmp = y0;
         y0 = y1;
         y1 = tmp;
      }
      int deltax = x1 - x0;
      int deltay = Math.abs(y1 - y0);
      int error = deltax / 2;
      int ystep = -1;
      int y = y0;
      if (y0 < y1)
         ystep = 1;
      for (int x = x0; x <= x1; x++) {
         if (steep)
            mark(y, x);
         else
            mark(x, y);
         error -= deltay;
         if (error < 0) {
            y += ystep;
            error += deltax;
         }
      }
   }


I was wondering if there is a faster method of obtaining all pixels in a line/better implementation?
Offline avm1979

Full Member
**

Posts: 157
Medals: 10



« Reply #1 on: 2011-07-15 15:22:10 »

That should be very fast. Fast enough that you could do it thousands of times a second and not worry about at all.

Unless you're doing something hideously slow in the "mark" function. If this were actually slow, that's where I'd look...

Offline Roquen

JGO Strike Force
***

Posts: 823
Medals: 25



« Reply #2 on: 2011-07-15 16:00:07 »

You're leaving out the most important piece of information: what do you really want to do?
Games published by our own members! Go get 'em!
Offline counterp

Full Member
**

Posts: 235
Medals: 11



« Reply #3 on: 2011-07-15 19:47:35 »

yes it is fast, it seems the problem is linked to the other thread I posted (with the images)
Offline pitbuller

Sr. Member
**

Posts: 340
Medals: 9



« Reply #4 on: 2011-07-16 05:02:34 »

int error = deltax / 2;

Could be:

int error = deltax >>1;



I have allways thinked do Java found these kind of bit shift optimizations automatically?

Offline Roquen

JGO Strike Force
***

Posts: 823
Medals: 25



« Reply #5 on: 2011-07-16 05:44:54 »

Not the same if negative.
Offline zoto

Full Member
**

Posts: 234
Medals: 7



« Reply #6 on: 2011-07-16 19:27:24 »

1  
int error = deltax >>>1;

http://en.wikipedia.org/wiki/Arithmetic_shift
http://en.wikipedia.org/wiki/Logical_shift
Offline counterp

Full Member
**

Posts: 235
Medals: 11



« Reply #7 on: 2011-07-17 01:19:19 »


see post above yours
Offline Roquen

JGO Strike Force
***

Posts: 823
Medals: 25



« Reply #8 on: 2011-07-17 03:28:32 »

1  
2  
3  
  int i = -1;
  int r0 = i/2;
  int r1 = i>>1;


Shifting (alone) isnt't the same for negative.  The result depends on the sub-bit, so compilers won't convert division into a shift unless it can statically determine that all values are positive or zero.
Pages: [1]
  Print  
 
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.16 | SMF © 2011, Simple Machines Valid XHTML 1.0! Valid CSS!
Page created in 0.103 seconds with 20 queries.