WiP in WiP, post your screenshots!
- Thread starter Arhurt
- Start date
Sel
Banned
- Feb 18, 2009
- 1,239
- 2,570
cryengine
takabuschik
aa
- Apr 14, 2013
- 662
- 344
Make it a sniper SMG mod
nightwatch
aa
- Sep 7, 2012
- 638
- 501
Messing around with using displacements for walls (rather than brush geometry) to avoid invalid solids, light seams, clipping errors, and patchy lighting in areas that necessitate complex curved faces.
Was about 45 min of work but now that I have a method I can probably think of a way to automate it.
That's 45 minutes for what would be 8 brush faces (but technically 128 displacement faces). If anyone knows a way to select specific vertices to move when "painting geometry" in the displacement tool that would be awesome.
Was about 45 min of work but now that I have a method I can probably think of a way to automate it.
That's 45 minutes for what would be 8 brush faces (but technically 128 displacement faces). If anyone knows a way to select specific vertices to move when "painting geometry" in the displacement tool that would be awesome.
nightwatch
aa
- Sep 7, 2012
- 638
- 501
Yes, but then you're limited to one at a time. I want to be able to select specific groups of vertices that are not necessarily close to each other in a spherical pattern, which is what the current tool allows. I'd like to select a line of vertices, for instance.
nightwatch
aa
- Sep 7, 2012
- 638
- 501
It was only 10 minutes to complete the second half, so it's not that bad I guess. Still takes too long without automation though.
a
EDIT: In terms of filesize, a stripped map with only the specific geometry and a light_environment to provide lighting information was 19.5 KB. A func-detailed version in brushwork was 20.9 KB. Assuming that the majority of that filesize is overhead for the basic map structure, it is interesting to note that the displacement version is cheaper. Both will have the same vis information, so why is the brush based version 1.4 KB more?
I'm guessing part of it is lighting information, seeing as displacements and brushes are lit differently. There's also the possibility of different bsp-structure (not vis structure!) for collisions etc, but I'm not sure. Anyone have ideas?
I figure it's 128 tris for the displacement version and 32 tris for the brushwork version, so in terms of basic geometry the brushwork should be cheaper. I'm going to try a fullbright version next.
EDIT 2: After compiling with no lighting I can confirm that the lighting is most of the discrepancy.
Fullbright Displacement: 18.0 KB
Fullbright Brushwork: 17.9 KB
EDIT 3: I compiled with no VIS because I was curious, and the filesize dropped again. Both versions were 16.9 KB. This seems a little silly since theoretically both versions would have been empty cubes (because neither displacements nor details cut vis) but clearly that was not the case.
a
EDIT: In terms of filesize, a stripped map with only the specific geometry and a light_environment to provide lighting information was 19.5 KB. A func-detailed version in brushwork was 20.9 KB. Assuming that the majority of that filesize is overhead for the basic map structure, it is interesting to note that the displacement version is cheaper. Both will have the same vis information, so why is the brush based version 1.4 KB more?
I'm guessing part of it is lighting information, seeing as displacements and brushes are lit differently. There's also the possibility of different bsp-structure (not vis structure!) for collisions etc, but I'm not sure. Anyone have ideas?
I figure it's 128 tris for the displacement version and 32 tris for the brushwork version, so in terms of basic geometry the brushwork should be cheaper. I'm going to try a fullbright version next.
EDIT 2: After compiling with no lighting I can confirm that the lighting is most of the discrepancy.
Fullbright Displacement: 18.0 KB
Fullbright Brushwork: 17.9 KB
EDIT 3: I compiled with no VIS because I was curious, and the filesize dropped again. Both versions were 16.9 KB. This seems a little silly since theoretically both versions would have been empty cubes (because neither displacements nor details cut vis) but clearly that was not the case.
Last edited:
this might just be me, but lightmap grids/scale on displacements are a bit wonky.
i.e. a 64x64 displacement with lightmap scale set to 1 should have 64^2 luxels, right?
according to the 3D lightmap grid mode, the displacement seems to have 65^2 luxels.
a 160^2 displacement with lightmap scale set to 16 should have 10^2 luxels, but again, it looks like it has 11^2 luxels instead.
if you create a normal brush next to the displacement and apply the same texture, it will show the proper 16 units/luxel. for a 160^2 surface, using a displacement increased the number of luxels by 21%. up from 100 to 121.
wait, i misread something... displacements were cheaper?
I think nodraw surfaces still get considered for lightmapping despite being "invisible" but they do provide physics collisions to players, props, and projectiles and also block light (iirc). solution a) use the skip tool material. b) increase the lightmap scale for parts of the geometry that shouldn't be visible to a large number, like 128-1024.
i.e. a 64x64 displacement with lightmap scale set to 1 should have 64^2 luxels, right?
according to the 3D lightmap grid mode, the displacement seems to have 65^2 luxels.
a 160^2 displacement with lightmap scale set to 16 should have 10^2 luxels, but again, it looks like it has 11^2 luxels instead.
if you create a normal brush next to the displacement and apply the same texture, it will show the proper 16 units/luxel. for a 160^2 surface, using a displacement increased the number of luxels by 21%. up from 100 to 121.
wait, i misread something... displacements were cheaper?
I think nodraw surfaces still get considered for lightmapping despite being "invisible" but they do provide physics collisions to players, props, and projectiles and also block light (iirc). solution a) use the skip tool material. b) increase the lightmap scale for parts of the geometry that shouldn't be visible to a large number, like 128-1024.
takabuschik
aa
- Apr 14, 2013
- 662
- 344
What you're saying is that if I'll make all nodraw surfaces with absurd lightmap scale (1024 units/luxel for example), I'll drastically reduce filesize?
Inari's finale in 60fps webm glory
On blue's win.
When the pit opens
On red win (WIP, more particles incoming)
On blue's win.
When the pit opens
On red win (WIP, more particles incoming)
Last edited:
@ nightwatch
Try to look at the different lumps in the .bsp. That let's you see what exactly takes more space.
https://developer.valvesoftware.com/wiki/Source_BSP_File_Format#Lump_types
Try to look at the different lumps in the .bsp. That let's you see what exactly takes more space.
https://developer.valvesoftware.com/wiki/Source_BSP_File_Format#Lump_types
I stole Aly's idea and made my own WebM thingies for the finale of koth_probed
Blu win
Red win
3 Minutes of the map doing it's thing
All of this still needs some tweaks still, going to do that tonight/ tomorrow.
Blu win
Red win
3 Minutes of the map doing it's thing
All of this still needs some tweaks still, going to do that tonight/ tomorrow.
We've seen Aly's props in Glassworks. You two are totally the maps to beat for the MvA contest. When are we gonna see a real collaboration? 
Also, is anyone ever so slightly annoyed that your tractor beam isn't over the point? Why not move the point if you can't move the space ship?
Also, is anyone ever so slightly annoyed that your tractor beam isn't over the point? Why not move the point if you can't move the space ship?
- Aug 7, 2014
- 1,241
- 1,025
Alarms for the lasers opening up, sounds for the UFOs/Mothership coming in, and screeching when the mothership is going down would add a lot. Right now it's sort of silent and eerie.
There is currently audio for the Mothership (it's a low rumble) and I just realized the UFO audio isn't kicking on when it's supposed to, so I'll fix that. And there is the extra audio that comes in as the mothership goes down as well, a sort of engine dying off noise.
I tried to not go overboard with audio because once you add in 12 vs 12 people fighting/ making explosions + voice comms it can get a bit overwhelming sound-wise. I learned that lesson from Wubwubwub.
I'll look into something for the lasers, though. It could use something pointing them out a bit more.
I tried to not go overboard with audio because once you add in 12 vs 12 people fighting/ making explosions + voice comms it can get a bit overwhelming sound-wise. I learned that lesson from Wubwubwub.
I'll look into something for the lasers, though. It could use something pointing them out a bit more.
- Feb 18, 2012
- 246
- 407
NEW and FUN screenshots from felgen:
I doubt I will continue the development of ctf_felgen, since it has been reduced to a chore for me, especially since I have no chance reaching the deadline.
Those "chunky" debris from the mothership kinda make it looks like it consists entirely of clay or brick, instead of plates and metal. It looks more like a brick slammed against a wall, rather than a hollow structure blowing up.
The other finals is pretty cool though. Especially how the laser is revealed.
I doubt I will continue the development of ctf_felgen, since it has been reduced to a chore for me, especially since I have no chance reaching the deadline.
Those "chunky" debris from the mothership kinda make it looks like it consists entirely of clay or brick, instead of plates and metal. It looks more like a brick slammed against a wall, rather than a hollow structure blowing up.
The other finals is pretty cool though. Especially how the laser is revealed.
Last edited: