Humus - Comments

More pages: 1 ... 3 4 5 6 7 8 9 10 11 12 13 ... 21 ... 31 ... 41 ... 51 ... 61 ... 71 ... 81 ... 91 ... 101 ... 111 ... 121 ... 131 ... 141 ... 151 ... 161 ... 171 ... 181 ... 191 ... 201 ... 211 ... 221 ... 231 ... 241 ... 251 ... 261 ... 271 ... 281 ... 291 ... 301 ... 311 ... 321 ... 331 ... 341 ... 351 ... 361 ... 365

Query FailedAaaaaa
Saturday, November 19, 2016

Very nice I really like your demos

huating
Wednesday, October 26, 2016

Hi,i wonder which tool you use to create these cube maps? i want to create some cube maps by myself.

thank you~

Humus
Thursday, October 20, 2016

Yes, a real game engine wouldn't want to do global lightmaps for the entire level with all lights everywhere. For say a typical indoor shooter, you could just subdivide it into chunks, such as a room. But yes, even with a decent subdivision strategy, you will still consume a factor X greater amount of memory than an old-school lightmap. In the best case, X would approach something like the average number of lights hitting an average texel, but likely a bit higher than that. On the other hand, this technique only stores a single 8-bit channel, which could easily compress to BC4. And compared to some recent lightmapping techniques where they store SH or SG coefficients, the storage for this in a typical application would probably be less than that. But with that said, I don't think this technique is for every game, but it could certainly be useful for some.

Michael
Tuesday, October 18, 2016

Hi, I see each light produces one .dds for the lightmap, does that mean if you have 100 lights this method ends to 100 .dds ? How this method scales for a real application ? Thanks

Nuninho1980
Saturday, July 30, 2016

I've new GTX 980 Ti SC+ again and I got min 599 fps - look my old GTX 780 Ti Classified.

crazii
Sunday, June 12, 2016

I agree with the article. Usually after projection, the .w contains the original view space Z/-Z, that's why W buffer is linear.
after perspective divide, the z/w is actually a/Z where Z is original view space Z.

Actually, hardware choose a/z (depth buffer distribution) right because linear interpolation can be easily and correctly done in screen space. so of course SV_POSITION.z is linear in screen space.

But, when talking about "linear" we usually are talking about functions in formats of f(Z)=aZ+b, related to the original Z.

for f(Z) = a/Z + b stored in z buffer, where Z is the original view space Z: even though it has a linear distribution in screen space, f(Z) is NOT a linear function to Z.

Jeb Bradwell
Sunday, June 5, 2016

Solid implementation. This is definitely a great direction

Jameel
Friday, June 3, 2016

When do you think you'll finish the rest of the cubemaps "in the pipeline"?


General
	News 3D Pictures Textures Articles Cool stuff
Other
	FAQ About Humus
Other 3D sites
	OpenGL.org GameDev.net Beyond3D Rage3D
Programming info
	OpenGL extension registry NeHe