Announcement

Collapse
No announcement yet.

VCTF-RT-TrainCrossing v2

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

    #61
    :/
    i don't see how is this important considering that downtown meshes and textures are only 18 mb and the rest of assets were already myleveled into another maps or were in huge packages

    Comment


      #62
      Another thing you could also try to decrease the map size is to change the lightmap resolution of BSP surfaces in the surface properties depending on how detailed the shadows on that surface need to be to look good.
      Surfaces that are uniformly lit or uniformly shadowed can do fine with a much rougher lightmap resolution (= cranking that lightmap value up, since it's measured in surface texels per pixel on the lightmap) while some areas might look nicer if you give them a more detailed surface.

      Comment


        #63
        Hi UnShame,
        Thanks for continuing to work on this map! I ended up remodeling the kargo barge boat altogether. I've got a much lower tri version that is around 2000 tris for the whole thing as one model (barge base, cab, mast and crates) and only five materials (with simplified collision). I'll upload the new barge along with some lower-tri versions of the UT3 foliage, and lower-tri versions of the UT3 Cargo Crates to use once I've got them ready to go.

        A couple of requests regarding map optimization:
        1. Try what Crusha recommended regarding lightmap resolution on invisible or shaded surfaces.
        2. Convert as many of the "curved surface" brushes to the 'semi-solid' type to limit the BSp torture (manhole covers, wooden dock pillars, curved wlakway in the garden courtyard, and the curved railings and stair/rail bridge section.
        3. Add an anitportal in the wall next to the flagroom that is towards the water (that way the decorations in the flag room will be better hidden from the area where the barge pulls in.
        4. Set a cull distance for all of the individual rail/fence pillar staticmesh objects on the verandas (surrounding where the tables and chairs are)

        Comment


          #64
          Thanks for continuing to work on this map! I ended up remodeling the kargo barge boat altogether. I've got a much lower tri version that is around 2000 tris for the whole thing as one model (barge base, cab, mast and crates) and only five materials (with simplified collision). I'll upload the new barge along with some lower-tri versions of the UT3 foliage, and lower-tri versions of the UT3 Cargo Crates to use once I've got them ready to go.
          great! thanks
          A couple of requests regarding map optimization:
          1. Try what Crusha recommended regarding lightmap resolution on invisible or shaded surfaces.
          2. Convert as many of the "curved surface" brushes to the 'semi-solid' type to limit the BSp torture (manhole covers, wooden dock pillars, curved wlakway in the garden courtyard, and the curved railings and stair/rail bridge section.
          3. Add an anitportal in the wall next to the flagroom that is towards the water (that way the decorations in the flag room will be better hidden from the area where the barge pulls in.
          4. Set a cull distance for all of the individual rail/fence pillar staticmesh objects on the verandas (surrounding where the tables and chairs are)
          I'll come back to this map again after a week or so to finish it. Some more optimization before i move on to making the terrain and hidden rooms won't harm, that's for sure
          I've got some questions though:
          1.How to create a custom mover that will have different MoverGlideTypes in different keys? i want a mover (the boat) to smoothly increase the speed from key0 to key1, move on the constant speed from key1 to key5, smoothly decrease the speed between key5 and 6 etc. I need to break the route of the boat into pieces, ~10 seconds each, because otherwise, if i have just two keys between which the boat is moving, when playing online the boat starts to lag really badly. It looks like the netcode of ut2004 can't handle a mover that moves for more than 15-20 seconds for such a long distance.
          2.How to set a physics volume to affect only projectiles that were launched inside the volume? I want to set up a little system of moving physics volumes that will simulate the horizontal velocity vector on the barge and trains so when you jump you will move along with the barge\train. But when i fire e.g. flak grenade over a train it gains the speed of the train without contact, and it looks weird.
          3.Also, how to set a volume to affect only players?

          Comment


            #65
            A bit of unusual feedback on the bot support. I'm working on a tool that captures player movement and visualizes it in a kind of heatmap. The tool is currently in a proof-of-concept stage, so I wouldn't want to release anything yet, but maybe this result is of use for you:

            This is a plot of a match on the beta 2 version with 16 bots playing on their own for about 20 minutes. I stopped the match afterwards because they didn't make any capture at all.
            The main thing to take away is that the red side seems to get more coverage than the blue side. Red bots also seemed to fly/hover across the water area, while blue bots didn't.

            Comment


              #66
              teh game is over 5 years old, you guys and gals live in the "age of fibre" and you complain about the map size becuase it is big,
              - noting that the reason is big because this OP probably made an Effort on the map and uses custom assets

              - @wormbo - you complaining about file size ???- does this not just merely highlights you as a hypocrit ??? [given you have been here for 6 years and have huge content files and mods and probably should actually be happy ppl are still active enough to make big files sizes]
              ^^ that is a fair question not intended as an insult or hate-- jsut curious why you feel anxiety towars this TINY size map file [based on current tech]


              - you kids and you technology need to learn how beneficial this new technology is to old games like UT and how the files size in this day and age - is TOTALLY 100% irrelevant
              -- optimize your map OP - and good luck, i always liked the idea of this map- But it was horribly made so i will DL when your finished, and probably give you hints and tips after the waste men have finished with there hating

              - Good luck- stay active and remove the cicada

              - given that epic failed alotby trying to save file size space by removing normals etc- and have materials that contain material expressions that do not connect to any nodes etc -- i think you should forget about the file size and just say thanks - and offer help...

              Comment


                #67
                Apart from being totally uncalled for, you do not measure file size compared to other technology. This game will need to handle it, and this game is not current technology, as you pointed out yourself. In fact, we didn't even expect him to do unusual hackish tricky stuff to reduce the size.

                Tell me please, where I created huge content files. And please use a UT2004 example. Crusha already said it: Content packages exist for a reason - to be shared by multiple mods/maps. If all maps myLeveled all their content, we'd indeed end up with map sizes of more recent games. That, however, should be considered a problem with more recent games. The reason we poke at the map size is online play - the map needs to be downloaded at least once by the player. A map this size still takes one to several minutes to download. Players likely choose to switch servers instead of waiting for that. The content packages are already available and used with another map, so there's a chance people already have them and don't need to download them again.

                Comment


                  #68
                  A bit of unusual feedback on the bot support. I'm working on a tool that captures player movement and visualizes it in a kind of heatmap. The tool is currently in a proof-of-concept stage, so I wouldn't want to release anything yet, but maybe this result is of use for you:

                  This is a plot of a match on the beta 2 version with 16 bots playing on their own for about 20 minutes. I stopped the match afterwards because they didn't make any capture at all.
                  The main thing to take away is that the red side seems to get more coverage than the blue side. Red bots also seemed to fly/hover across the water area, while blue bots didn't.
                  nice tool
                  i see bots in the blue team don't go for the cicada... that explains why they don't go over the water - they're using only cicada for that

                  Comment


                    #69
                    Originally posted by Wormbo View Post
                    The reason we poke at the map size is online play - the map needs to be downloaded at least once by the player. A map this size still takes one to several minutes to download.
                    Indeed. Actually is my internet connection not the most stable one, so it tends to disconnect on constant large downloads about every 50 MB. The more individual packages there are, the more likely it is that I can actually successfully download a map from a server or resume the download after a disconnect.

                    Comment


                      #70
                      so it tends to disconnect on constant large downloads about every 50 MB.
                      that's great because the map is 40mb when in uz2 xD

                      Comment


                        #71
                        Originally posted by UnShame View Post
                        nice tool
                        i see bots in the blue team don't go for the cicada... that explains why they don't go over the water - they're using only cicada for that
                        You might want to also place HoverPathNodes for the Mantas then. About the ship and trains - I'm not sure bots will figure them out, but you can probably experiment with lift pathing here.

                        Comment


                          #72
                          i placed some hoverpathnodes over the water and connected them with roadpathnodes but bots just ignored them so i removed it

                          Comment


                            #73
                            @UnShame: Here is a link to a map file with the optimized barge mesh along with several other remodeled/optimized/converted decoration static meshes (it ONLY includes the new stuff so you should be able to open it up, copy whatever you want to use and paste it into your map).

                            Since my old computer could barely run the previous version I ended up optimizing a few other meshes too (barrels, shipping crates, train tracks, the little Azalea bushes from UT3, the manhole covers, the wood dock pilings, and I converted the background Tree Line mesh/texture from UT3 for use in the map), feel free to use or not use them. I was able to cut the train track mesh down by roughly 75% triangle count, and I made a new model for the energy track emitter bases (no point in wasting almost 1000 triangles and 2 512x512 textures on something most players won't spend a lot of time looking at right?). The benches (which were something like 800 triangles) were remodeled down to something like 200 tris. Simplified collision models were included where appropriate.

                            I also included a few little emitter experiments to fake/simulate speedtree sprites (they are unlit though so they don't work particularly well). Here are a few pictures of the new stuff:
                            <-> <->

                            Comment


                              #74
                              Yeah, back in the days when I extracted those meshes Blender wasn't able yet to use the Decimate tool without destroying the mesh's UV layout, so I was unable to reduce the triangle count and had to settle for meshes with reasonably low default counts.

                              Comment


                                #75
                                Originally posted by meowcat View Post
                                I also included a few little emitter experiments to fake/simulate speedtree sprites (they are unlit though so they don't work particularly well). Here are a few pictures of the new stuff:
                                <-> <->

                                The fake-speedtree-emitter is a great idea, meowcat!
                                A good idea for remakes of UT3 maps/UT3 map ports as the quality of UT2004's tree meshes seems a bit off-place compared to the rest.

                                Comment

                                Working...
                                X