Commented 3 months ago2024-01-13 12:32:14 UTC
in vault item: Computer_LabComment #105872
This has potential, keep learning and experimenting! It's a good first map. I actually like that the button for releasing the HEV suit has to be hit with the crowbar. I don't know if it was just me, but when I got to barney, he glitched inside me. Someone more knowledgeble might give some insight as to why that happened there.
Commented 3 months ago2024-01-13 03:27:26 UTC
in vault item: cross_stalk_tributeComment #105870
Please provide feedback here if something needs to be fixed.
Also, if you're running a server, please add it to a map rotation. This actually 'saves' Stalkyard, as no one is playing it now - and running to a Bunker is always fun
Commented 3 months ago2024-01-13 00:20:03 UTC
in wiki page: VERC: Doors With GlassComment #105869
I discovered a trick with using BOUNDINGBOX brush which eliminates the lip math. You make it the same width as the main door and tie it to the glass part. No messing with lip value for the glass door; use the same value as the main door.
Rather empty, I feel like it needed a lot more combat scenarios. A lot of areas could have been great for Combine arenas instead of just Zombies being there. Zombie spawns could have also been better as well, they just appear in front of you and weren't much of a threat. A Fast Zombie or a Poison Zombie here and there would have had some risk to the encounters.
The progression to the water tower was great though, it did actually feel like I went through a journey and being able to see the tower from the beginning and being able to see where you began at the end was nice. I would suggest that for your next map that you focus on combat scenarios. Maybe put little spins on them as well for an extra challenge.
I wrote a short journal entry on how all 15 lumps relate to each other. I used the relational model to visualize it.
Because the referencing of entries by other entries is done with indices into the lump using mostly 16-bit shorts, it places a hard limit on how many of the target lump's entries can exist before their indices becomes out of bounds for the data types used. In addition, Carmack makes a few of these data types do double (sometimes triple) duty e.g. positive values refer to one type, negative values means entirely different type, and 0 means null, so that halves the capacity of a type's max index. clipnodes suffer the most from this; an entire half (the negative half) of addressable values in 2^16 is used to assign only 2 significant values -1 and -2.
I've incorporated the information into the first section of this page. You can read the table as:
The resource A has a limit of X because it is referenced by JKL with data type T, using [Remark]
The resource clipnodes has a limit of 32768 because it is referenced by other clipnodes with data type short, using positive values > 0
This version of the map format is colloquially known as Valve20 to differentiate from the original Quake version
Valve20 is now the de-facto map format all Quake-based and Quake-derived editors and compilers support due to the precision in texture alignment
TrenchBroom uses the OS's clipboard for copy and paste (not the case with JACK/VHE3.x) and uses the map format to represent brush & entity data to be pasted. It also means clipboard-based IPC can be done using the map format.
Commented 3 months ago2024-01-06 06:19:43 UTC
in wiki page: func_detailComment #105850
What I can gather from the community's collective wisdom:
func_detail does not reduce world leaves
func_detail instead reduce VIS portals generated by HLBSP. if without it HLBSP makes VIS portals around detail geometry, with it it just treat the surrounding area as a single VIS leaf
func_detail is gone at the end of HLBSP, and is world geometry after that point
hence HLRAD cannot treat it in any way other than opaque world brush because it's gone by the time it's run
One thing I will say though is that the reflection isn't done well. There should not be a huge gap between objects and their reflection. The reflection of the wire is where this is most obvious and you can even see it in the screenshot. The reflection would show on the floor's surface, so objects ON the floor would be in contact with their reflection.
Commented 3 months ago2024-01-05 12:12:36 UTC
in wiki page: func_tracktrainComment #105848
Some findings about func_tracktrain:
wheels distance is only measured forward from origin. To make the train travel as much on the path as possible, the origin needs to be located at the center of the back axle.
wheel distance defaults to 100.
target of a path_track is fired when the origin passes the path_track.
the position and angle is set wrt the origin point as follows:
trace a line forward.
if there's enough distance to the end of the current path, keep heading
if wheel distance > distance to next path, subtract current distance to next path, then iterate the path forwards until the distance fits the path. the point where distance is 0 is the new heading.
origin moves forward
angle is adjusted to point to the heading. this will in all cases be shorter than wheel distance; the sharper the curve the more noticeable. this also means the front wheel part will always overshoot the path. but the origin is guaranteed to always be on the path.
pretty comfy, nicely detailed and good music choice, the shaking is a bit too strong imo though.
I feel like this could be used as one of those "waiting for players to join" rooms along with the start button that some SC maps have. It'd be pretty cool and fitting to wait in a place like this.
PS: I managed to jump on that window and saw the secret texture : )
It'd be pretty cool to use the visgroups to assign attribution e.g. all of sir's prefabs is assigned a visgroup to his name. This is to comply with the "BY" aspect of the CC license.
textures can be scaled as need because now lightmap can be recalculated after any changes.
You still need to be careful because the maximum lightmap size for a face is apparently 16x16 luxels. A lot of faces you'll find in maps are already at 16 luxels on either (or both) sides. GoldSrc doesn't support lightmaps larger than that size, and scaling down a texture that results in a face with lightmaps over that size crashes the engine. (IDK if it's supported on svengine). Hence my verdict is don't do it.
As for vertex editing, I just don't recommend it at all too. It's way more cumbersome, and way way harder to get satisfactory result than to get completely broken result. I honestly don't know how a face whose vertices are non-planar would act inside the game. If there's face splitting ability in world faces and easier vertex shifting via handles vs textboxes then it might be feasible, but still not recommended because of the non-planar thing. *
The main takeaway is that just because it's doable with the BSP structure doesn't mean the game engine that's going to parse and run the BSP would like it. If the game engine doesn't like it, don't do it.
I still don't have any draft for a part 2 of this tutorial. 😅
EDIT 1: * Tested face vertex editing. It messes up lightmaps elsewhere on a map.
Keep going. I'm eager to play more maps from you.
Thank you.
Also, if you're running a server, please add it to a map rotation. This actually 'saves' Stalkyard, as no one is playing it now - and running to a Bunker is always fun
BOUNDINGBOX
brush which eliminates the lip math. You make it the same width as the main door and tie it to the glass part. No messing with lip value for the glass door; use the same value as the main door.I see a lot of effort here, and effort Needs to be premied!
You're welcome.
Texturing — 4
Ambience — 5
Lighting — 5
Gameplay — 3
The progression to the water tower was great though, it did actually feel like I went through a journey and being able to see the tower from the beginning and being able to see where you began at the end was nice. I would suggest that for your next map that you focus on combat scenarios. Maybe put little spins on them as well for an extra challenge.
Because the referencing of entries by other entries is done with indices into the lump using mostly 16-bit shorts, it places a hard limit on how many of the target lump's entries can exist before their indices becomes out of bounds for the data types used. In addition, Carmack makes a few of these data types do double (sometimes triple) duty e.g. positive values refer to one type, negative values means entirely different type, and 0 means null, so that halves the capacity of a type's max index. clipnodes suffer the most from this; an entire half (the negative half) of addressable values in 2^16 is used to assign only 2 significant values -1 and -2.
I've incorporated the information into the first section of this page. You can read the table as:
It doesn't work in Counter-Strike.
I'll be sure to take a look once I have a PC again
Can anyone provide a screenshot of the effect of the "Fog Affects Skybox" flag?
One thing I will say though is that the reflection isn't done well. There should not be a huge gap between objects and their reflection. The reflection of the wire is where this is most obvious and you can even see it in the screenshot. The reflection would show on the floor's surface, so objects ON the floor would be in contact with their reflection.
It's made for the competition and is ready to run on any server.
I feel like this could be used as one of those "waiting for players to join" rooms along with the start button that some SC maps have. It'd be pretty cool and fitting to wait in a place like this.
PS: I managed to jump on that window and saw the secret texture : )
Keep up the good work man!
There we go.
As for vertex editing, I just don't recommend it at all too. It's way more cumbersome, and way way harder to get satisfactory result than to get completely broken result. I honestly don't know how a face whose vertices are non-planar would act inside the game. If there's face splitting ability in world faces and easier vertex shifting via handles vs textboxes then it might be feasible, but still not recommended because of the non-planar thing. *
The main takeaway is that just because it's doable with the BSP structure doesn't mean the game engine that's going to parse and run the BSP would like it. If the game engine doesn't like it, don't do it.
I still don't have any draft for a part 2 of this tutorial. 😅
EDIT 1:
* Tested face vertex editing. It messes up lightmaps elsewhere on a map.