hi!
I have this code that copy the pixel right and make a pixel shift, i need to do with OPENGL or a similar library.
With Glgraphics works fine in a speed computer ( iMac -nvidia- 2.4 GHz) but in my poor pc runs slow (acer -ATi graphic card! 1.9 GHz)

Any idea?
thanks!

Code:

import processing.opengl.*;
import codeanticode.glgraphics.*;

GLTexture texP;

int xshift =1;

void setup()
{
  size(1024,768, GLConstants.GLGRAPHICS);
  hint( ENABLE_OPENGL_4X_SMOOTH );  
  int w = width;
  int h = height;

  texP = new GLTexture(this);
  texP.init(width, height);
  noStroke();
  background(0); 

}

void draw()
{
  ellipse(mouseX,mouseY,50,50);
  pixelShift(xshift,0);

}
void pixelShift(int xshift,int yshift){

  texP.putImage(get());
  texP.loadPixels();
  texP.updateTexture();
  for (int y=1; y < height; y++) {
    for (int x=1; x < width; x++){

	if ((x+xshift < width) && (x+xshift > 0)) {
	  if ((y+yshift < height) && (y+yshift > 0)) {
	    texP.pixels[x + (y*width)] = texP.pixels[(x+xshift)+ ((y+yshift)*width)];
	  }
	}

    }
  }
  texP.loadTexture();
  texP.updatePixels();
//image(texP,0,0);
  texP.render(0,0);

}



 

One possible solutions -
(1) Render the image to an off-screen buffer and then use image() to copy part of the buffer to the actual screen. This would not only be faster but would be simple to implement a pixel rotation not just a shift. The downside is the amount of memory needed for the buffer.

(2) The problem with your solution is the double loop with nested if statements and multiplications in the inner loop.
It makes use of the arrayCopy() method. This method allows you to copy part or all of an array into another array. This method method calls a Java native method so is fast and using it could reduce the whole problem down to a single loop to iterate through each row.