PDA

View Full Version : GL_EXT_texture_filter_anisotropic



dukey
11-13-2011, 06:01 AM
This extension is 12 years old now. All hardware supports it. Why it not yet part of the core spec ?

ZbuffeR
11-13-2011, 07:35 AM
Patents.
One should be able to implement core OpenGL without having to pay for patents on top of the license fee.

Is it fairly well explained here :
http://arcsynthesis.org/gltut/Texturing/Tut15%20Anisotropy.html

dukey
11-13-2011, 07:56 AM
That's a real shame ..

aqnuep
11-13-2011, 11:03 AM
Here you can find a list of patent issues related to OpenGL and OpenGL ES:
https://www.khronos.org/files/ip-disclosures/opengl/

There are actually patent issues with several extensions: anisotropic filtering (NVIDIA), floating point framebuffers (SGI), tessellation (Matrox), S3TC texture compression (S3), etc.

Alfonse Reinheart
11-13-2011, 11:17 AM
That's a real shame ..

Why? What does it matter?

The functionality is there. Isn't that the important part? Does the fact that it's an extension somehow stop you from using it? Extensions are not lesser parts of OpenGL. They're just theoretically optional parts. Since this one is ubiquitous (http://www.opengl.org/wiki/Ubiquitous_Extensions) (even Intel gets it right), what's the problem with just using it?

dukey
11-13-2011, 05:07 PM
It matters because I missed it entirely, since it was missing from the core spec !

Alfonse Reinheart
11-13-2011, 06:05 PM
Good point. It would be nice if the ARB could at least promote them to ARB status and point them out somewhere in the spec as being "highly recommended".

What's odd is this: RGTC and BPTC are both core. Yet they use almost the exact same compression algorithm as S3TC. Indeed, RGTC uses the alpha compression from DXT5. So why can those be core, while S3TC (and aniso) can't?

mhagain
11-14-2011, 02:47 AM
Interesting news item: http://www.phoronix.com/scan.php?page=news_item&px=OTkxMQ

The S3TC patent may be invalid.