Attachment:

megatexture_r3_1.png [ 1.27 MB | Viewed 434 times ]

Video:
http://www.youtube.com/watch?v=eqmWeZmi ... e=youtu.be

Download(binary + megatexture + source code):
http://www.gamefront.com/files/22586803 ... lic_v2_rar

MegaProject files:
http://www.gamefront.com/files/22586897 ... roject.rar

I was going to make this a tutorial but it would be too long of one because everything required ALOT of changes. To install drag and drop into your doom 3 folder, this is fully compatible with Sikkmod.

Features:

BUGS:

Building a megatexture

Opps forgot the material :/.

Make a new material called megatextures.mtr, this one also takes out the self shadow for the terrain.

Code:

megatexture/planar_valley
{
   solid
   cmTwoSided
   noShadows
   qer_editorimage megatextures\valley_preview.tga
   {
      megatexture valley.mega
   }   
}
//

textures/megatexture/grass_01
{
   twoSided
   nonsolid
   qer_editorimage textures/vegetation/grassfoliage1.tga
   //bumpmap   textures/vegetation/pinetree01_local.tga
   {
      blend diffusemap
      map    textures/vegetation/grassfoliage1.tga
      alphatest .4
   }
}

textures/megatexture/grass_02
{
   twoSided
   nonsolid
   qer_editorimage textures/vegetation/grassfoliage2.tga
   //bumpmap   textures/vegetation/pinetree01_local.tga
   {
      blend diffusemap
      map    textures/vegetation/grassfoliage2.tga
      alphatest .4
   }
}

textures/megatexture/grass_03
{
   twoSided
   nonsolid
   qer_editorimage textures/vegetation/grassfoliage3.tga
   //bumpmap   textures/vegetation/pinetree01_local.tga
   {
      blend diffusemap
      map    textures/vegetation/grassfoliage3.tga
      alphatest .4
   }
}


textures/megatexture/tron_grass
{
   //noselfshadow
   noshadows
   nonsolid
   twosided
   qer_editorimage textures/vegetation/tron_grass.tga
   //{
   //   blend add
   //   map    textures/vegetation/tron_grass.tga
   //   alphatest .5
   //}
    {
      blend GL_SRC_ALPHA, GL_ONE_MINUS_SRC_ALPHA
      map    textures/vegetation/tron_grass.tga
      //alphatest .9
      scale 2,1
      rgb .9      
    }  
   

}

textures/megatexture/tron_tall_grass
{
   //noselfshadow
   noshadows
   nonsolid
   twosided
   qer_editorimage textures/vegetation/tron_tall_grass.tga
   //{
   //   blend add
   //   map    textures/vegetation/tron_grass.tga
   //   alphatest .5
   //}
    {
      //blend diffusemap
      blend GL_SRC_ALPHA, GL_ONE_MINUS_SRC_ALPHA
      map    textures/vegetation/tron_tall_grass.tga
      alphatest .65

      
    }  
   

}

textures/megatexture/tron_bark
{
  //noselfshadow
   noshadows
   nonsolid
   //twosided
 qer_editorimage textures/vegetation/tron_bark_diffuse.tga
 diffusemap textures/vegetation/tron_bark_diffuse.tga
 bumpmap textures/vegetation/tron_bark_local.tga

}

textures/megatexture/tron_wall_mossy_large
{
  //noselfshadow
   //noshadows
   //nonsolid
   twosided
 qer_editorimage textures/structures/brick_wall_moss_bottom.tga
 diffusemap textures/structures/brick_wall_moss_bottom.tga
 bumpmap textures/structures/brick_wall_moss_bottom_local.tga 
}

textures/megatexture/tron_stone_wall
{
  //noselfshadow
   //noshadows
   //nonsolid
  twosided
 qer_editorimage textures/structures/stone_wall_smaller.tga
 diffusemap textures/structures/stone_wall.tga
 bumpmap textures/structures/stone_wall_local.tga 
 //diffusemap textures/structures/stone_wall_smaller.tga
 //bumpmap textures/structures/stone_wall_local_smaller.tga 
}

textures/megatexture/tron_wood_brick_wall

{
 qer_editorimage textures/structures/wood_brick_wall_smaller.tga
 //diffusemap textures/structures/wood_brick_wall.tga
 //bumpmap textures/structures/wood_brick_wall_local.tga 
 diffusemap textures/structures/wood_brick_wall_smaller.tga
 bumpmap textures/structures/wood_brick_wall_local_smaller.tga 
}




textures/megatexture/valley_water

{
   qer_EditorImage      textures/pbmax/water1_local.tga
   {
      //mirrorRenderMap   128 128               // size of the texture to draw the mirror view in 
      blend   blend
      alpha         0                  //100% reflective, 0% transparent
      translate      0.5, 0.5
      rotate         0.5
      scale         5, 5
      program      fresnel_water_pixel.vfp         //shader program
      vertexParm   0   time * 0.008 , time * 0.0      //scroll
      vertexParm   1   10                  //magnitude
      fragmentMap   0    env\mega_up.tga   //the mirror/reflection image
      fragmentMap   1   textures/hell/clouds_local.tga   //normal map for distortion & "per pixel" fresnel
      
   }
}

Cool. I will definitely give this a crack and post results.

One question, that is slightly unrelated. Does your CDK have global illumination?

Yes I was using beast but later got in realtime GI using the unreal 4 voxel method, it's slow as shit though.

Ooh beast, nice! So I guess that's gone now or will that make a come back? Also awesome job here on the megastuff it's cool beans.

The only problem I see with Beast is that it's probably got a huge license fee.

https://github.com/lsalzman/tesseract

might be something of interest. Tesseract lighting looks really good for an amateur product, almost as good as Cryengine on some of the maps I tested.

Dashiva wrote:

The only problem I see with Beast is that it's probably got a huge license fee.

https://github.com/lsalzman/tesseract

might be something of interest. Tesseract lighting looks really good for an amateur product, almost as good as Cryengine on some of the maps I tested.

I'm focusing on a 100% realtime engine rather than baking everything, as the next gen tech will focus on interaction time rather than prettier graphics, so I'm moving away from things that can add to interation time.

Yeah, rebuilding lighting in UDK is like watching paint dry. Over and over. And over. The question is, if the UE 4 voxel lighting is too expensive, what are you going to go with? Cryengine-esque deferred lighting?

Yea, on second thought if a RT solution can run fast enough and scale well I'll change my answer. Perhaps, I'll feel differently once I drop my 8800 for something faster although I'll probably just end up with a 560 so no idea if even that would be good enough. Keeping in mind that anything using CDK will likely be at least a year after your first solid release so by then most people should have pretty good hardware anyway.

I just... I guess I just have to learn to let go of light maps... my little safety net for horrifically placed lights.

Cryengine 3 does it really well, no lightmaps or anything.

Dashiva wrote:

Cryengine 3 does it really well, no lightmaps or anything.

CryEngine 3 does a lot of things very well, aside from being at a price point that developers are able to afford

.

Quote:

Yeah, rebuilding lighting in UDK is like watching paint dry. Over and over. And over. The question is, if the UE 4 voxel lighting is too expensive, what are you going to go with? Cryengine-esque deferred lighting?

I was thinking of maybe light propagation volumes http://www6.incrysis.com/Light_Propagation_Volumes.pdf

Quote:

I just... I guess I just have to learn to let go of light maps... my little safety net for horrifically placed lights.

F*cking artists

.

Personally I don't have a problem forcing developers to be on high end hardware, because games now a days average 4+ years to make and a substantial amount of whatever your target audience is, will have good enough hardware to do all this stuff in realtime.

http://imgur.com/doz5c

Seems to fail on shader compile or something.

Dashiva wrote:

The only problem I see with Beast is that it's probably got a huge license fee.

https://github.com/lsalzman/tesseract

might be something of interest. Tesseract lighting looks really good for an amateur product, almost as good as Cryengine on some of the maps I tested.

It has a nice feature list but Sauerbraten and Tesseract have an extremely poor code quality. I've read open source engines for 12 years now and those are actually the worst.
So if you want to make a game and look for some programmers then you should rather make a big step away from those.

I'm really looking forward for the CDK by Justin. His stuff is very nice. There will be also some nice open engine upgrades by me once I will complete my game.
I also have experience with the Cryengine, UDK and Unity 3D. If you want to create a game wether commercially or for free then I recommend to learn Doom 3 inside out.
You can do pretty much everything you want with it unlike with the others.

Tr3B wrote:

Dashiva wrote:

The only problem I see with Beast is that it's probably got a huge license fee.

https://github.com/lsalzman/tesseract

might be something of interest. Tesseract lighting looks really good for an amateur product, almost as good as Cryengine on some of the maps I tested.

It has a nice feature list but Sauerbraten and Tesseract have an extremely poor code quality. I've read open source engines for 12 years now and those are actually the worst.
So if you want to make a game and look for some programmers then you should rather make a big step away from those.

Yah, I was just referring to it as another example of realtime GI. The cube engines lack any sort of modern art pipeline, I wouldn't use them for anything.

Anyway, anyone get this to work yet? I tried again but it comes up with the same shader compile bug.

jmarshall23 wrote:

No longer based off the Tron's megatexture shader, but rather based off of the megatexture shader in the bfg edition.

Don't credit me for that, all AO's fantastic work!

Tron wrote:

jmarshall23 wrote:

No longer based off the Tron's megatexture shader, but rather based off of the megatexture shader in the bfg edition.

Don't credit me for that, all AO's fantastic work!

Edited the topic, it was great work done by you guys none the less.

. Carmacks original shader just was slightly faster and yielded a slight slight visual increase, so there wasn't that much of a difference you guys did a good job. The speed boost/quality increase came from increasing the clip level size to 2048, and instead of having a levelsizexlevelsize texture for each level, I switched it over to one texture and uploaded the miplevels required for each miplevel, rather than the old implementation for uploading the entire level(even though it technically was 1/4 size of the original but it still filled up the entire texture space), and than the old code generated mipmaps on the fly which obviously had to go.

That mixed with dxt5 compression and realtime lighting makes megatextures usable. Stencil shadows just obviously have to go which I'm pretty close to finishing up the tutorial on.

It's amazing progress you've made, potential for some really nice looking terrain now.

Attachment:

megatexture_r3_1.png [ 1.27 MB | Viewed 434 times ]

Video:
http://www.youtube.com/watch?v=eqmWeZmi ... e=youtu.be

Download(binary + megatexture + source code):
http://www.gamefront.com/files/22586803 ... lic_v2_rar

MegaProject files:
http://www.gamefront.com/files/22586897 ... roject.rar

I was going to make this a tutorial but it would be too long of one because everything required ALOT of changes. To install drag and drop into your doom 3 folder, this is fully compatible with Sikkmod.

Features:

BUGS:

Building a megatexture

Opps forgot the material :/.

Make a new material called megatextures.mtr, this one also takes out the self shadow for the terrain.

Code:

megatexture/planar_valley
{
   solid
   cmTwoSided
   noShadows
   qer_editorimage megatextures\valley_preview.tga
   {
      megatexture valley.mega
   }   
}
//

textures/megatexture/grass_01
{
   twoSided
   nonsolid
   qer_editorimage textures/vegetation/grassfoliage1.tga
   //bumpmap   textures/vegetation/pinetree01_local.tga
   {
      blend diffusemap
      map    textures/vegetation/grassfoliage1.tga
      alphatest .4
   }
}

textures/megatexture/grass_02
{
   twoSided
   nonsolid
   qer_editorimage textures/vegetation/grassfoliage2.tga
   //bumpmap   textures/vegetation/pinetree01_local.tga
   {
      blend diffusemap
      map    textures/vegetation/grassfoliage2.tga
      alphatest .4
   }
}

textures/megatexture/grass_03
{
   twoSided
   nonsolid
   qer_editorimage textures/vegetation/grassfoliage3.tga
   //bumpmap   textures/vegetation/pinetree01_local.tga
   {
      blend diffusemap
      map    textures/vegetation/grassfoliage3.tga
      alphatest .4
   }
}


textures/megatexture/tron_grass
{
   //noselfshadow
   noshadows
   nonsolid
   twosided
   qer_editorimage textures/vegetation/tron_grass.tga
   //{
   //   blend add
   //   map    textures/vegetation/tron_grass.tga
   //   alphatest .5
   //}
    {
      blend GL_SRC_ALPHA, GL_ONE_MINUS_SRC_ALPHA
      map    textures/vegetation/tron_grass.tga
      //alphatest .9
      scale 2,1
      rgb .9      
    }  
   

}

textures/megatexture/tron_tall_grass
{
   //noselfshadow
   noshadows
   nonsolid
   twosided
   qer_editorimage textures/vegetation/tron_tall_grass.tga
   //{
   //   blend add
   //   map    textures/vegetation/tron_grass.tga
   //   alphatest .5
   //}
    {
      //blend diffusemap
      blend GL_SRC_ALPHA, GL_ONE_MINUS_SRC_ALPHA
      map    textures/vegetation/tron_tall_grass.tga
      alphatest .65

      
    }  
   

}

textures/megatexture/tron_bark
{
  //noselfshadow
   noshadows
   nonsolid
   //twosided
 qer_editorimage textures/vegetation/tron_bark_diffuse.tga
 diffusemap textures/vegetation/tron_bark_diffuse.tga
 bumpmap textures/vegetation/tron_bark_local.tga

}

textures/megatexture/tron_wall_mossy_large
{
  //noselfshadow
   //noshadows
   //nonsolid
   twosided
 qer_editorimage textures/structures/brick_wall_moss_bottom.tga
 diffusemap textures/structures/brick_wall_moss_bottom.tga
 bumpmap textures/structures/brick_wall_moss_bottom_local.tga 
}

textures/megatexture/tron_stone_wall
{
  //noselfshadow
   //noshadows
   //nonsolid
  twosided
 qer_editorimage textures/structures/stone_wall_smaller.tga
 diffusemap textures/structures/stone_wall.tga
 bumpmap textures/structures/stone_wall_local.tga 
 //diffusemap textures/structures/stone_wall_smaller.tga
 //bumpmap textures/structures/stone_wall_local_smaller.tga 
}

textures/megatexture/tron_wood_brick_wall

{
 qer_editorimage textures/structures/wood_brick_wall_smaller.tga
 //diffusemap textures/structures/wood_brick_wall.tga
 //bumpmap textures/structures/wood_brick_wall_local.tga 
 diffusemap textures/structures/wood_brick_wall_smaller.tga
 bumpmap textures/structures/wood_brick_wall_local_smaller.tga 
}




textures/megatexture/valley_water

{
   qer_EditorImage      textures/pbmax/water1_local.tga
   {
      //mirrorRenderMap   128 128               // size of the texture to draw the mirror view in 
      blend   blend
      alpha         0                  //100% reflective, 0% transparent
      translate      0.5, 0.5
      rotate         0.5
      scale         5, 5
      program      fresnel_water_pixel.vfp         //shader program
      vertexParm   0   time * 0.008 , time * 0.0      //scroll
      vertexParm   1   10                  //magnitude
      fragmentMap   0    env\mega_up.tga   //the mirror/reflection image
      fragmentMap   1   textures/hell/clouds_local.tga   //normal map for distortion & "per pixel" fresnel
      
   }
}

Cool. I will definitely give this a crack and post results.

One question, that is slightly unrelated. Does your CDK have global illumination?

Yes I was using beast but later got in realtime GI using the unreal 4 voxel method, it's slow as shit though.

Ooh beast, nice! So I guess that's gone now or will that make a come back? Also awesome job here on the megastuff it's cool beans.

The only problem I see with Beast is that it's probably got a huge license fee.

https://github.com/lsalzman/tesseract

might be something of interest. Tesseract lighting looks really good for an amateur product, almost as good as Cryengine on some of the maps I tested.

Dashiva wrote:

The only problem I see with Beast is that it's probably got a huge license fee.

https://github.com/lsalzman/tesseract

might be something of interest. Tesseract lighting looks really good for an amateur product, almost as good as Cryengine on some of the maps I tested.

I'm focusing on a 100% realtime engine rather than baking everything, as the next gen tech will focus on interaction time rather than prettier graphics, so I'm moving away from things that can add to interation time.

Yeah, rebuilding lighting in UDK is like watching paint dry. Over and over. And over. The question is, if the UE 4 voxel lighting is too expensive, what are you going to go with? Cryengine-esque deferred lighting?

Yea, on second thought if a RT solution can run fast enough and scale well I'll change my answer. Perhaps, I'll feel differently once I drop my 8800 for something faster although I'll probably just end up with a 560 so no idea if even that would be good enough. Keeping in mind that anything using CDK will likely be at least a year after your first solid release so by then most people should have pretty good hardware anyway.

I just... I guess I just have to learn to let go of light maps... my little safety net for horrifically placed lights.

Cryengine 3 does it really well, no lightmaps or anything.

Dashiva wrote:

Cryengine 3 does it really well, no lightmaps or anything.

CryEngine 3 does a lot of things very well, aside from being at a price point that developers are able to afford

.

Quote:

Yeah, rebuilding lighting in UDK is like watching paint dry. Over and over. And over. The question is, if the UE 4 voxel lighting is too expensive, what are you going to go with? Cryengine-esque deferred lighting?

I was thinking of maybe light propagation volumes http://www6.incrysis.com/Light_Propagation_Volumes.pdf

Quote:

I just... I guess I just have to learn to let go of light maps... my little safety net for horrifically placed lights.

F*cking artists

.

Personally I don't have a problem forcing developers to be on high end hardware, because games now a days average 4+ years to make and a substantial amount of whatever your target audience is, will have good enough hardware to do all this stuff in realtime.

http://imgur.com/doz5c

Seems to fail on shader compile or something.

Dashiva wrote:

The only problem I see with Beast is that it's probably got a huge license fee.

https://github.com/lsalzman/tesseract

might be something of interest. Tesseract lighting looks really good for an amateur product, almost as good as Cryengine on some of the maps I tested.

It has a nice feature list but Sauerbraten and Tesseract have an extremely poor code quality. I've read open source engines for 12 years now and those are actually the worst.
So if you want to make a game and look for some programmers then you should rather make a big step away from those.

I'm really looking forward for the CDK by Justin. His stuff is very nice. There will be also some nice open engine upgrades by me once I will complete my game.
I also have experience with the Cryengine, UDK and Unity 3D. If you want to create a game wether commercially or for free then I recommend to learn Doom 3 inside out.
You can do pretty much everything you want with it unlike with the others.

Tr3B wrote:

Dashiva wrote:

The only problem I see with Beast is that it's probably got a huge license fee.

https://github.com/lsalzman/tesseract

might be something of interest. Tesseract lighting looks really good for an amateur product, almost as good as Cryengine on some of the maps I tested.

It has a nice feature list but Sauerbraten and Tesseract have an extremely poor code quality. I've read open source engines for 12 years now and those are actually the worst.
So if you want to make a game and look for some programmers then you should rather make a big step away from those.

Yah, I was just referring to it as another example of realtime GI. The cube engines lack any sort of modern art pipeline, I wouldn't use them for anything.

Anyway, anyone get this to work yet? I tried again but it comes up with the same shader compile bug.

jmarshall23 wrote:

No longer based off the Tron's megatexture shader, but rather based off of the megatexture shader in the bfg edition.

Don't credit me for that, all AO's fantastic work!

Tron wrote:

jmarshall23 wrote:

No longer based off the Tron's megatexture shader, but rather based off of the megatexture shader in the bfg edition.

Don't credit me for that, all AO's fantastic work!

Edited the topic, it was great work done by you guys none the less.

. Carmacks original shader just was slightly faster and yielded a slight slight visual increase, so there wasn't that much of a difference you guys did a good job. The speed boost/quality increase came from increasing the clip level size to 2048, and instead of having a levelsizexlevelsize texture for each level, I switched it over to one texture and uploaded the miplevels required for each miplevel, rather than the old implementation for uploading the entire level(even though it technically was 1/4 size of the original but it still filled up the entire texture space), and than the old code generated mipmaps on the fly which obviously had to go.

That mixed with dxt5 compression and realtime lighting makes megatextures usable. Stencil shadows just obviously have to go which I'm pretty close to finishing up the tutorial on.

It's amazing progress you've made, potential for some really nice looking terrain now.

Attachment:

megatexture_r3_1.png [ 1.27 MB | Viewed 434 times ]

Video:
http://www.youtube.com/watch?v=eqmWeZmi ... e=youtu.be

Download(binary + megatexture + source code):
http://www.gamefront.com/files/22586803 ... lic_v2_rar

MegaProject files:
http://www.gamefront.com/files/22586897 ... roject.rar

I was going to make this a tutorial but it would be too long of one because everything required ALOT of changes. To install drag and drop into your doom 3 folder, this is fully compatible with Sikkmod.

Features:

BUGS:

Building a megatexture

Opps forgot the material :/.

Make a new material called megatextures.mtr, this one also takes out the self shadow for the terrain.

Code:

megatexture/planar_valley
{
   solid
   cmTwoSided
   noShadows
   qer_editorimage megatextures\valley_preview.tga
   {
      megatexture valley.mega
   }   
}
//

textures/megatexture/grass_01
{
   twoSided
   nonsolid
   qer_editorimage textures/vegetation/grassfoliage1.tga
   //bumpmap   textures/vegetation/pinetree01_local.tga
   {
      blend diffusemap
      map    textures/vegetation/grassfoliage1.tga
      alphatest .4
   }
}

textures/megatexture/grass_02
{
   twoSided
   nonsolid
   qer_editorimage textures/vegetation/grassfoliage2.tga
   //bumpmap   textures/vegetation/pinetree01_local.tga
   {
      blend diffusemap
      map    textures/vegetation/grassfoliage2.tga
      alphatest .4
   }
}

textures/megatexture/grass_03
{
   twoSided
   nonsolid
   qer_editorimage textures/vegetation/grassfoliage3.tga
   //bumpmap   textures/vegetation/pinetree01_local.tga
   {
      blend diffusemap
      map    textures/vegetation/grassfoliage3.tga
      alphatest .4
   }
}


textures/megatexture/tron_grass
{
   //noselfshadow
   noshadows
   nonsolid
   twosided
   qer_editorimage textures/vegetation/tron_grass.tga
   //{
   //   blend add
   //   map    textures/vegetation/tron_grass.tga
   //   alphatest .5
   //}
    {
      blend GL_SRC_ALPHA, GL_ONE_MINUS_SRC_ALPHA
      map    textures/vegetation/tron_grass.tga
      //alphatest .9
      scale 2,1
      rgb .9      
    }  
   

}

textures/megatexture/tron_tall_grass
{
   //noselfshadow
   noshadows
   nonsolid
   twosided
   qer_editorimage textures/vegetation/tron_tall_grass.tga
   //{
   //   blend add
   //   map    textures/vegetation/tron_grass.tga
   //   alphatest .5
   //}
    {
      //blend diffusemap
      blend GL_SRC_ALPHA, GL_ONE_MINUS_SRC_ALPHA
      map    textures/vegetation/tron_tall_grass.tga
      alphatest .65

      
    }  
   

}

textures/megatexture/tron_bark
{
  //noselfshadow
   noshadows
   nonsolid
   //twosided
 qer_editorimage textures/vegetation/tron_bark_diffuse.tga
 diffusemap textures/vegetation/tron_bark_diffuse.tga
 bumpmap textures/vegetation/tron_bark_local.tga

}

textures/megatexture/tron_wall_mossy_large
{
  //noselfshadow
   //noshadows
   //nonsolid
   twosided
 qer_editorimage textures/structures/brick_wall_moss_bottom.tga
 diffusemap textures/structures/brick_wall_moss_bottom.tga
 bumpmap textures/structures/brick_wall_moss_bottom_local.tga 
}

textures/megatexture/tron_stone_wall
{
  //noselfshadow
   //noshadows
   //nonsolid
  twosided
 qer_editorimage textures/structures/stone_wall_smaller.tga
 diffusemap textures/structures/stone_wall.tga
 bumpmap textures/structures/stone_wall_local.tga 
 //diffusemap textures/structures/stone_wall_smaller.tga
 //bumpmap textures/structures/stone_wall_local_smaller.tga 
}

textures/megatexture/tron_wood_brick_wall

{
 qer_editorimage textures/structures/wood_brick_wall_smaller.tga
 //diffusemap textures/structures/wood_brick_wall.tga
 //bumpmap textures/structures/wood_brick_wall_local.tga 
 diffusemap textures/structures/wood_brick_wall_smaller.tga
 bumpmap textures/structures/wood_brick_wall_local_smaller.tga 
}




textures/megatexture/valley_water

{
   qer_EditorImage      textures/pbmax/water1_local.tga
   {
      //mirrorRenderMap   128 128               // size of the texture to draw the mirror view in 
      blend   blend
      alpha         0                  //100% reflective, 0% transparent
      translate      0.5, 0.5
      rotate         0.5
      scale         5, 5
      program      fresnel_water_pixel.vfp         //shader program
      vertexParm   0   time * 0.008 , time * 0.0      //scroll
      vertexParm   1   10                  //magnitude
      fragmentMap   0    env\mega_up.tga   //the mirror/reflection image
      fragmentMap   1   textures/hell/clouds_local.tga   //normal map for distortion & "per pixel" fresnel
      
   }
}

Cool. I will definitely give this a crack and post results.

One question, that is slightly unrelated. Does your CDK have global illumination?

Yes I was using beast but later got in realtime GI using the unreal 4 voxel method, it's slow as shit though.

Ooh beast, nice! So I guess that's gone now or will that make a come back? Also awesome job here on the megastuff it's cool beans.

The only problem I see with Beast is that it's probably got a huge license fee.

https://github.com/lsalzman/tesseract

might be something of interest. Tesseract lighting looks really good for an amateur product, almost as good as Cryengine on some of the maps I tested.

Dashiva wrote:

The only problem I see with Beast is that it's probably got a huge license fee.

https://github.com/lsalzman/tesseract

might be something of interest. Tesseract lighting looks really good for an amateur product, almost as good as Cryengine on some of the maps I tested.

I'm focusing on a 100% realtime engine rather than baking everything, as the next gen tech will focus on interaction time rather than prettier graphics, so I'm moving away from things that can add to interation time.

Yeah, rebuilding lighting in UDK is like watching paint dry. Over and over. And over. The question is, if the UE 4 voxel lighting is too expensive, what are you going to go with? Cryengine-esque deferred lighting?

Yea, on second thought if a RT solution can run fast enough and scale well I'll change my answer. Perhaps, I'll feel differently once I drop my 8800 for something faster although I'll probably just end up with a 560 so no idea if even that would be good enough. Keeping in mind that anything using CDK will likely be at least a year after your first solid release so by then most people should have pretty good hardware anyway.

I just... I guess I just have to learn to let go of light maps... my little safety net for horrifically placed lights.

Cryengine 3 does it really well, no lightmaps or anything.

Dashiva wrote:

Cryengine 3 does it really well, no lightmaps or anything.

CryEngine 3 does a lot of things very well, aside from being at a price point that developers are able to afford

.

Quote:

Yeah, rebuilding lighting in UDK is like watching paint dry. Over and over. And over. The question is, if the UE 4 voxel lighting is too expensive, what are you going to go with? Cryengine-esque deferred lighting?

I was thinking of maybe light propagation volumes http://www6.incrysis.com/Light_Propagation_Volumes.pdf

Quote:

I just... I guess I just have to learn to let go of light maps... my little safety net for horrifically placed lights.

F*cking artists

.

Personally I don't have a problem forcing developers to be on high end hardware, because games now a days average 4+ years to make and a substantial amount of whatever your target audience is, will have good enough hardware to do all this stuff in realtime.

http://imgur.com/doz5c

Seems to fail on shader compile or something.

Dashiva wrote:

The only problem I see with Beast is that it's probably got a huge license fee.

https://github.com/lsalzman/tesseract

might be something of interest. Tesseract lighting looks really good for an amateur product, almost as good as Cryengine on some of the maps I tested.

It has a nice feature list but Sauerbraten and Tesseract have an extremely poor code quality. I've read open source engines for 12 years now and those are actually the worst.
So if you want to make a game and look for some programmers then you should rather make a big step away from those.

I'm really looking forward for the CDK by Justin. His stuff is very nice. There will be also some nice open engine upgrades by me once I will complete my game.
I also have experience with the Cryengine, UDK and Unity 3D. If you want to create a game wether commercially or for free then I recommend to learn Doom 3 inside out.
You can do pretty much everything you want with it unlike with the others.

Tr3B wrote:

Dashiva wrote:

The only problem I see with Beast is that it's probably got a huge license fee.

https://github.com/lsalzman/tesseract

might be something of interest. Tesseract lighting looks really good for an amateur product, almost as good as Cryengine on some of the maps I tested.

It has a nice feature list but Sauerbraten and Tesseract have an extremely poor code quality. I've read open source engines for 12 years now and those are actually the worst.
So if you want to make a game and look for some programmers then you should rather make a big step away from those.

Yah, I was just referring to it as another example of realtime GI. The cube engines lack any sort of modern art pipeline, I wouldn't use them for anything.

Anyway, anyone get this to work yet? I tried again but it comes up with the same shader compile bug.

jmarshall23 wrote:

No longer based off the Tron's megatexture shader, but rather based off of the megatexture shader in the bfg edition.

Don't credit me for that, all AO's fantastic work!

Tron wrote:

jmarshall23 wrote:

No longer based off the Tron's megatexture shader, but rather based off of the megatexture shader in the bfg edition.

Don't credit me for that, all AO's fantastic work!

Edited the topic, it was great work done by you guys none the less.

. Carmacks original shader just was slightly faster and yielded a slight slight visual increase, so there wasn't that much of a difference you guys did a good job. The speed boost/quality increase came from increasing the clip level size to 2048, and instead of having a levelsizexlevelsize texture for each level, I switched it over to one texture and uploaded the miplevels required for each miplevel, rather than the old implementation for uploading the entire level(even though it technically was 1/4 size of the original but it still filled up the entire texture space), and than the old code generated mipmaps on the fly which obviously had to go.

That mixed with dxt5 compression and realtime lighting makes megatextures usable. Stencil shadows just obviously have to go which I'm pretty close to finishing up the tutorial on.

It's amazing progress you've made, potential for some really nice looking terrain now.