[hlcoders] HLVIS (Warning, technical details and code)

[Skyler York]

[ Converted text/html to text/plain ]
I've been working on a few new features for HLVIS, and I have been making good
progess.  One of the features is a new maximum VIS option.  In addition to
being able to compile the map the first time around with this new feature, you
can chose to run a MaxDistVis on an already compiled map.

This is my problem.  If you run a MaxDistVis during the initial compile, the
the code flags the correct leafs to be invisible:

------------------------------------------------
// i = leafnum of first leaf, j = leafnum of second leaf
offset_l = i >> 3;
bit_l = (1 << (i & 7));
offset_tl = j >> 3;
bit_tl = (1 << (j & 7));

for(k = 0; k < l->numportals; k++)
{
     for(m = 0; m < tl->numportals; m++)
     {
          l->portals[k]->mightsee[offset_tl] &= ~bit_tl;
          tl->portals[m]->mightsee[offset_l] &= ~bit_l;
     }
 }

----------------------------------------------
It basically eliminates leafs before they are even fully VISed.

However, if you run a MaxDistVis on an already compiled map, for the most
part, the leafs rendered in game are correct, except for a few oddball leafs
that aren't "unset" in the code for some unknown reason.  The detection is the
same, but it doesn't unset them.  What I have to do first is decompress all
the vis information:
-------------------------------------------------
void   DecompressAll(void)
{
     int i;
     byte *dest;
     for(i = 0; i < g_portalleafs; i++)
     {
          dest = g_uncompressed + i * g_bitbytes;
          DecompressVis((const unsigned char*)(g_dvisdata + (byte)g_dleafs[i +
1].visofs), dest, g_bitbytes);
     }
}
------------------------------------------------

I use this code to "unset" leafs:
-----------------------------------------------
// i = leafnum of first leaf, j = leafnum of second leaf
byte* base_vis_l = g_uncompressed + i * g_bitbytes;
byte* base_vis_tl = g_uncompressed + j * g_bitbytes;
offset_l = i >> 3;
bit_l = (1 << (i & 7));
offset_tl = j >> 3;
bit_tl = (1 << (j & 7));
base_vis_l[offset_tl] &= ~bit_tl;
base_vis_tl[offset_l] &= ~bit_l;

------------------------------------------------

...and this code to recompress the vis info:
------------------------------------------------
void   CompressAll(void)
{
     int i, x = 0;
     byte *dest;
     byte *src;
     byte compressed[MAX_MAP_LEAFS / 8];
     vismap_p = vismap;

     for(i = 0; i < g_portalleafs; i++)
    {
        memset(&compressed, 0, sizeof(compressed));
        src = g_uncompressed + i * g_bitbytes;

       // Compress all leafs into global compression buffer
       x = CompressVis(src, g_bitbytes, compressed, sizeof(compressed));
       dest = vismap_p;
       vismap_p += x;
       if (vismap_p > vismap_end)
      {
           Error("Vismap expansion overflow");
       }

       g_dleafs[i + 1].visofs = dest - vismap;            // leaf 0 is a
common solid
       memcpy(dest, compressed, x);
     }
}
--------------------------------------------------

It works at around a 90% - 95% success rate, but those oddball leafs annoy me.

Before I forget, here are some pics of what I mean.

This is the map compiled normally:
http://144.132.16.252/zipster/[2]mdv0.jpg

This is the map compiled the first time around with a max distance of 128:
http://144.132.16.252/zipster/[4]mdv1.jpg

And here is a picture of the original map being MaxDistVis'd after a regular
compile (w/ 128 limit):
http://144.132.16.252/zipster/[6]mdv2.jpg

As you can see, that blasted leaf in the distance is screwing me over!  I
don't know how many of you know much about HLVIS procedures, but I hope I've
given you enough information.  I would go into more detail, but there's just
too much if you don't have an idea of what's happening.

------------------------------------------------------------------------------
MSN Photos is the easiest way to share and print your photos: Click Here[7]

===References:===
  1. http://144.132.16.252/zipster/mdv0.jpg
  2. http://144.132.16.252/zipster/
  3. http://144.132.16.252/zipster/mdv1.jpg
  4. http://144.132.16.252/zipster/
  5. http://144.132.16.252/zipster/mdv2.jpg
  6. http://144.132.16.252/zipster/
  7. http://g.msn.com/1HM505401/15

_______________________________________________
To unsubscribe, edit your list preferences, or view the list archives, please visit:
http://list.valvesoftware.com/mailman/listinfo/hlcoders