开发者

Black Screen Effect - OpenGL

I'm new to OpenGL and I've been experiencing the "Black Screen Effect". I've spent ages trying to work out why I'm not seeing anything and I haven't had much success. I'm using LWJGL and here is the piece of code I'm trying to run:

glViewport(0,0,DISPLAY_WIDTH,DISPLAY_HEIGHT);

glMatrixMode(GL_PROJECTION);
glLoadIdentity();
glOrtho( 6200000.0f, 6300000.0f, 350000.0f, 380000.0f, -10000000.0f, 100000000.0f);
gluLookAt(368000.0f, 6250000.0f, -10000.0f, 368000.0f, 6250000.0f, 10000.0f, 0.0f,1.0f,0.0f);
glPushMatrix();

glMatrixMode(GL_MODELVIEW);
glLoadIdentity();
glPushMatrix();

if(ready)
{
glColor3f(1.0f,0.5f,1.0f);
glPointSize(100);
glBegin(GL_POINTS);
    for(int i = 0; i < data.length; i += 100)
    {
        glColor3f(1.0f,1.0f,1.0f);
        glVertex3f((float) data[i][1], (float) data[i][2],0.0f);
        System.out.println((float) data[i][1] + ", " + (float) data[i][2]);
    }
glEnd();

System.out.println("\n\nfinished drawing\n\n");
glFlush();

I am drawing in a different colour that i used to clear the screen.

My data set is quite large (over 100 000 points) so I tried plotting every hundredt开发者_如何转开发h point, but that's not working.

I am also trying to plot points at positions such as (400 000, 6 800 000) would this be presenting me with problems? I'm pretty sure that 32bit floating point numbers should be able to handle these values.

I am pretty certain that a pixel with size=1 will try to plot as 1 pixel on the screen, regardless of how small it is compared with the bounds of the orthographic projection.

Maybe I'm dealing with the projection matrix incorrectly.


First, like said in my comment don't use gluLookAt on the projection matrix. It defines the camera (view) and therefore belongs to the model view matrix. This isn't the cause for your problem and it should also work this way, but it is conceptually wrong.

Next, if you call this code every frame, you push a new matrix onto the stack every frame, without calling glPopMatrix. glPushMatrix is generally there to save the current matrix and restore it later with a call to glPopMatrix, because every other command (like glLoadIdentity, but also gluLookAt and glOrtho) modifies the current matrix (the one selected by glMatrixMode).

Otherwise, you should always keep the size of your scene in relation to the viewing volume (the glOrtho parameters in your case) in mind. At the moment you're looking from point (368000, 6250000, -10000) to point (368000, 6250000, 10000). Together with the glOrtho parameters this should define your viewing volume to be the [368000-6300000 , 368000-6200000] x [6250000+350000 , 6250000+380000] x [-10000000-10000, 100000000-10000] box. If you don't transform your points further by any local transformations, their coordinates should ly in these intervals to be visible. Keep an eye on the minus in the x-interval. This is due to the fact that you actually rotated the view volume 180 degrees around the y-axis, because you defined the view to look from -z to z, whereas GL's default eye-space defines the viewer to look from z to -z (which usually is not that much of a problem with an origin-symmetric viewing volume, but yours is highly asymmetric).

Although your numbers are extremely strange they should be handlable by 32bit floats. But are you really sure you want your points to have a size of 100 pixels (if this is even supported)?

And if you only draw 2D points in an orthographic view, I'm also not sure if you need such a HUGE depth range.

0

上一篇:

下一篇:

精彩评论

暂无评论...
验证码 换一张
取 消

最新问答

问答排行榜