I’ve got a sphere texturized with one texture, now I’m trying to add another over top of the first and make it transparent. Sorta like a cloud layer over a plant texture.
how do I do this?
Either use the ARB_multitexture extension, or draw the sphere twice, using blending the second time.
j
well, a little more detailed explanation whould be to do something like this:
-clearbuffers
-prepare matrices
-glDisable(GL_BLENDING);
-setup texture 1
-render sphere
-glDepthFunc(GL_LEQUAL);
-glEnable(GL_BLENDING);
-setup texture 2
-if (texture 2 has alpha channel) then
- glBlendFunc(GL_SRC_ALPHA,GL_ONE_MINUS_SRC_ALPHA);
-else // do aditive - glBlendFunc(GL_ONE,GL_ONE);
-endif
-render sphere
-swapbuffers
Those blending functions are just two examples, but should be the two most common.
> if (texture 2 has alpha channel)
then - glBlendFunc(GL_SRC_ALPHA,GL_ONE_MINUS_SRC_ALPHA);
Yes, this function is the most common.
But IMHO, in this way you are just wasting useful computational resources! 
(j/k)
You can create your texture with premultiplied pixels. In this case you can put an additive (for example - gloss map) part into the very same texture:
At = A; // dest. alpha value.
RGBt = (A * RGB) + RGBadd
glBlendFunc( GL_ONE, GL_ONE_MINUS_SRC_ALPHA );
Originally posted by coco:
[b]well, a little more detailed explanation whould be to do something like this:
-clearbuffers
-prepare matrices
-glDisable(GL_BLENDING);
-setup texture 1
-render sphere
-glDepthFunc(GL_LEQUAL);
-glEnable(GL_BLENDING);
-setup texture 2
-if (texture 2 has alpha channel) then
- glBlendFunc(GL_SRC_ALPHA,GL_ONE_MINUS_SRC_ALPHA);
-else // do aditive- glBlendFunc(GL_ONE,GL_ONE);
-endif
-render sphere
-swapbuffers
Those blending functions are just two examples, but should be the two most common.[/b]
OK. this sorta worked. I have no idea what you mean by alpha channel for a texture, I assume its somewhat similiar to how a gif has one color that is transparent. anyhow,
I create the one texturized sphere, and then created the cloud sphere surrounding the first sphere, and slightly larger. I then textured it, and did the blending, with the GL_ONE, GL_ONE blendfunc and got a semi- transparent texture with quad lines visible in the non-transparent part of the second texture… am I making any sense? So, basically it did work, but why are the polygon lines showing up in the second texture? If any of you guys could send me a simple sample (w/code) on how to do this, it would help alot.
Serge K:
Well, that’s right. I’m not sure how many hardware implementations will benefit from something like this, but in theory it’s more optimizable by the driver.
dex44:
alpha channel is like a Red channel, or a Green or a Blue channel, but instead of containing color information it contains transparency implementation, where the lower the alpha for a pixel the more translucent it is. For example, in 8bit alpha, alpha=64 will give you 25% opaque pixel, where alpha=192 will give you a 75% opaque pixel (25% translucent). Anyway, the final pixel color depends in other factors, such as alpha testing and blending functions.
An RGBA image differs from a RGB image in that it uses four bytes per pixel instead of three in order to fit the alpha channel.
[This message has been edited by coco (edited 01-11-2001).]
Originally posted by coco:
Well, that’s right. I’m not sure how many hardware implementations will benefit from something like this, but in theory it’s more optimizable by the driver.
It’s not about hardware implementations…
I just wanted to say that one can minimize number of the rendering passes with the smart blending function.
Does anyone has some exemple about this !
If anyone has , please send it in aceka@yahoo.com
Thanks !