SDL + OpenGL?

[dandymcgee]

I'm trying to figure out how to make use of OpenGL's power properly through SDL. Keep in mind, the most I've ever done in OpenGL is made a rotating cube, and I didn't even understand portions of that simple code.

First question: glOrtho vs. gluPerspective - Would it be more beneficial in the long run to use gluPerspective with a 2D tile engine, or does an orthogonal view provide everything necessary to rotate, zoom, etc.?

Basis: I attempted to create a 2D quad and texture it, but upon rotating it (any angle) it would immediately disappear. Does this always happen or was I just doing it wrong?

Second question: When texturing should I load to an SDL_Surface then convert it to OpenGL format, or is there some way to load it directly?

I apologize for any ridiculous assumptions I may have made, but I'd like to try to avoid picking up any bad habits while I'm still clueless.

[Ginto8]

I'm trying to figure out how to make use of OpenGL's power properly through SDL. Keep in mind, the most I've ever done in OpenGL is made a rotating cube, and I didn't even understand portions of that simple code.

First question: glOrtho vs. gluPerspective - Would it be more beneficial in the long run to use gluPerspective with a 2D tile engine, or does an orthogonal view provide everything necessary to rotate, zoom, etc.?

Basis: I attempted to create a 2D quad and texture it, but upon rotating it (any angle) it would immediately disappear. Does this always happen or was I just doing it wrong?

Second question: When texturing should I load to an SDL_Surface then convert it to OpenGL format, or is there some way to load it directly?

I apologize for any ridiculous assumptions I may have made, but I'd like to try to avoid picking up any bad habits while I'm still clueless.

#1: glOrtho vs. gluPerspective - orthographic projection is the ideal for anything 2D. Perspective projection is just to make depth line up (at least, you don't need it when you don't need realistic-looking depth).

#2: the textured 2D quad rotation: are you redrawing it every frame? glRotate only rotates the matrix, not the objects in it. Also, how are you calling it? To draw a rotated 2D object, you'd do

Code: Select all

glRotatef([angle], 0.f, 0.f, 1.f); // rotate [angle] on the Z axis
// drawing code goes here

#3: You could load it into an SDL_Surface then pass surface->pixels to glTexImage2D, or load it directly into OGL using something like libpng. Personally, I'd use libpng to load it into an SDL_Surface, then optimize and pad (my crappy graphics card doesn't support newer versions of OGL, so it can really only use textures with power of 2 dimensions) the surface. After that, I'd pass the padded image's pixels to glTexImage2D and free the surfaces.

#4: ridiculous assumptions are fine. Though some may say otherwise, most of those people would also get angry at you for having bad habits. Some people just love to tick others off, I guess.

[dandymcgee]

Perspective projection is just to make depth line up (at least, you don't need it when you don't need realistic-looking depth).

Well the goal of the question was to decide whether or not I need realistic depth to achieve any sort of zooming.

#2: the textured 2D quad rotation: are you redrawing it every frame? glRotate only rotates the matrix, not the objects in it. Also, how are you calling it? To draw a rotated 2D object, you'd do

Code: Select all

glRotatef([angle], 0.f, 0.f, 1.f); // rotate [angle] on the Z axis
// drawing code goes here

It was rotating and updating fine, but it was only visible on the screen when it's angle was zero (flat toward the screen). The code is gone now, so I can't provide an example presenting the error.

#3: You could load it into an SDL_Surface then pass surface->pixels to glTexImage2D, or load it directly into OGL using something like libpng.

I'll keep that in mind, thanks.

Bam, schooled by a 12-year-old.

Indeed.

[Ginto8]

Perspective projection is just to make depth line up (at least, you don't need it when you don't need realistic-looking depth).

Well the goal of the question was to decide whether or not I need realistic depth to achieve any sort of zooming.

It is completely possible to do zooming in orthographic projection; it really only takes a multiplication of the x, w, y, h of the texture's mapping area to get realistic zooming (though the x, w, y, h of the camera also have to be multiplied)

[dandymcgee]

It is completely possible to do zooming in orthographic projection; it really only takes a multiplication of the x, w, y, h of the texture's mapping area to get realistic zooming (though the x, w, y, h of the camera also have to be multiplied)

So basically you just draw it smaller, right?

[MarauderIIC]

Or bigger. This is what we do. But I don't know that we use perspective or orthographic at all, rather we just don't allow camera movement. Rather, we scale according to a certain factor (and correct by it). We don't even allow camera movement on the z-axis, although I'm not totally sure why we don't use this for zoom. Hmm. Probably something to do with DC's rendering system and stuff we already had in place, etc.

[dandymcgee]

Or bigger. This is what we do. But I don't know that we use perspective or orthographic at all, rather we just don't allow camera movement. Rather, we scale according to a certain factor (and correct by it). We don't even allow camera movement on the z-axis, although I'm not totally sure why we don't use this for zoom. Hmm. Probably something to do with DC's rendering system and stuff we already had in place, etc.

Genious! Thanks Marauder. :P