Digs
#101 posted by mfx on 2014/02/23 19:28:11
try color without the underscore.
#102 posted by digs on 2014/02/24 04:11:55
No, it turned out a different format from 0 to 255. I used to use from 0 to 1
#103 posted by digs on 2014/02/24 08:10:10
light format "_color" "255 255 255" worked, but not compatible with NetRadiant. Light entity looks like white light.
#104 posted by Spirit on 2014/02/24 08:11:47
255 255 255 IS white. try 255 0 0, does that display red?
#105 posted by digs on 2014/02/24 08:20:53
I mean the format of the parameter. No, "255 0 0" display as white. NetRadiant worked for format "_color" "0-1 0-1 0-1"
No Underscore?
#106 posted by mfx on 2014/02/24 12:45:03
nt
Mfx
#107 posted by digs on 2014/02/24 12:46:25
in documentation with underscore
Working For Me Without.
#108 posted by mfx on 2014/02/24 12:49:01
No joke.
Both Work Here
#109 posted by A passersby on 2014/02/24 14:37:11
I am using Tyrutils-0.15 and both properties work for me to give colored light. That is, I can either write the word color with or without an underscore and the results look identical to me. The numbers have to be in range from 0 to 255.
I just tested this with this example line inside a light entity declaration:
"color" "255 10 10" // it's a red light now
The light program produces a .lit file when I use the color properties in a map file, and I need to copy the .lit file alongside the .bsp file into the same directory for loading into a Quake engine. If I omit the .lit file I still get the lights in the map, but they are always white.
I'm using the most recent release version of Darkplaces to test the map & the lighting. I hope this clarified things for anyone having issues.
Degenerate Edge Error
#110 posted by Orl on 2014/04/14 16:01:28
This is a problem aimed specifically at tyrqbsp, so I would assume only Tyrann would know the answer to it, but if anyone else ever experienced the same problem or have any info regarding it, feel free to chime in.
My current map in progress, when sealed, produces a bogus error at the TJUNC stage of compiling with the message of:
---- Tjunc ----
************ ERROR ************
Degenerate edge at (0.000 0.000 0.000)
I really don't know what to make of this error, since the origin 0 0 0 of the map is located in the void. Curiously enough, when the map is unsealed it compiles without any error.
So without knowing what and where to look for the problem, I've hit a roadblock. I have sent you, Tyrann, a much more detailed description of this problem several days ago, but haven't gotten a response back, and figured perhaps you might see it here first.
I am using the latest version of tyrqbsp, along with Jackhammer as my editor. I am more than willing to provide you the .map file so that you may take a closer inspection of what the problem might be.
Or I suppose you can do what TxQbsp does and treat degenerate edges as warnings instead of errors. But I'd first like to hear your thoughts on what this error might mean, and what steps I can take to fix it.
First Thought
#111 posted by Preach on 2014/04/14 20:01:30
If it's at the origin and you haven't placed any brushes in that area, my guess would be that it's a rogue brush with zero size generated by the editor by mistake. Most editors have something that will check for problems like this and allow you to delete the offending brush.
Also
#112 posted by Qmaster on 2014/04/18 05:15:59
Most editors have something that will create brushes like this.
Snapshots
#113 posted by ericw on 2015/01/19 21:13:10
Here are builds with two experimental features - these are unofficial, not approved by Tyrann or anything :)
-dirt support
-0-1 color format support
Plus some older fixes that are already in tyrutils git:
-lightmap coordinates fix for 64-bit binaries (only affects mac or linux builds)
-fix for writing a corrupted bsp when using wad3 (halflife) textures
win32 osx source
#114 posted by - on 2015/01/19 22:25:40
awesome! thanks ericw!
Thanks Ericw!
#115 posted by quaketree on 2015/01/20 02:56:33
But I have a question. There was a bit of talk about doing some sort of radiosity with light.exe in quake when you first came up with the -dirt idea and it got me to thinking. Is it possible to do one or more of the following:
1) An invisible "Light" brush. I get that it "Can" cause problems with multiple faces and light types if it gets too big and it probably shouldn't be used with anything but a solid light (no flickering, throbbing or maybe even switchable). It works with _sunlight so I was wondering if it was possible to translate that over to other textures or if because of how sky is displayed that's too far afield.
2) A sort of "Fake" radiosity tied to specific textures like *lava and *slime by default. The way most engines handles liquids makes it look a bit off when it's in a dark location. Maybe even instead of default textures (like *lava or *slime) a way to set up a user created data file.txt that specifies specific textures that act like a "Light" brush above during the compile process).
Only spitballing here so feel free to ignore it.
How About
#116 posted by anonymous user on 2015/01/20 05:57:10
specifying light emitting textures? Defined by a text file given to the commandline. With all wait/delay functionality. Would help making spotlights for example. Just an idea.
Yeah
#117 posted by ericw on 2015/01/20 08:36:47
sounds like a great feature idea.
I could see either a command-line flag like "-lighttexture {texture light r g b wait delay}" or something, or having a list of these in a text file.
The quake2 light tool has it, (I think it's set as a surface flag?), and at first glance it seems to be intertwined with the radiosity support. It'd definitely be a bigger project than adding the -dirt was, but would be cool to do (as well as having radiosity)
Give Q1rad A Try
#118 posted by Spirit on 2015/01/20 08:36:53
It's Cool
#119 posted by ericw on 2015/01/20 09:05:05
I think i used that on my first map.. but it's old, the source isn't available, and the readme says it's derived from Valve's q1rad so it woudn't be possible to combine with a modern gpl quake1 tool anyway.
The quake2 qrad3 is GPL'ed though: https://github.com/id-Software/Quake-2-Tools/tree/master/bsp/qrad3
"Fake" Radiosity Tied To Specific Textures Like *lava And *slime By De
#120 posted by mh on 2015/01/20 11:18:35
I implemented this in the last version of MHColour, but instead of emitting light the radiosity was used to tint the already existing light data. It's easy enough to do and you can just butcher the code and get something that works up and running quick enough.
MHColor
#121 posted by quaketree on 2015/01/20 13:57:02
I looked at that and while it sort of does the job there is no control over it (re specific textures for example) specifically because it "Does" use pre-existing lights. The user also can't specify any textures to ignore or add.
It worked pretty well for generating .lit files for .bsps where the source files weren't available but quite honestly the hues and intensity of the colors weren't really to my liking about 80% of the time. Too much like Q2. As I've said before with Quake and colored lights, less is almost always more. It ain't the Las Vegas Strip. It's what you might find deep underneath the Las Vegas Strip after a bad night out.
I was thinking of something that didn't use a light entity (in the traditional sense) at all but turned the top face (or any face exposed to "Air" I suppose) of a brush into a light all by itself.
@quaketree
#122 posted by mh on 2015/01/20 17:21:13
In fairness that's all it ever really set out to do. The last version was a good deal more subtle than earlier versions, but I'd guess about 50% of users preferred the stronger colouring.
In general I agree - mapper control of lighting in the tools and at the compilation stage is the correct way to go, and MHColour was never meant to be a replacement for that.
Despite all the above the theory is still applicable - it's just tracing the BSP between points, then checking distance and facing, after all.
#123 posted by Lunaran on 2015/01/20 18:00:49
I'm pretty sure nobody's mentioned this yet.
So I tried out the fancy hacked dirtmapping and got this:
http://lunaran.com/pics/fitz0061.jpg
That's looking at the corner of a room. The diagonally opposite corner looks correct, and the two remaining corners are wrong on one plane or the other, almost as if the sample points are off by one in a positive-axis direction.
Derpmapping?
#124 posted by Zwiffle on 2015/01/20 18:01:17
Hm Weird
#125 posted by ericw on 2015/01/20 19:30:36
I also get some problems with a box map:
-lit -dirty -dirtdebug: http://i.imgur.com/q6ggYIe.png
-lit -dirty -dirtdebug -extra4: http://i.imgur.com/mmEszwv.png
Quakespasm is hacked to show lightmaps with nearest-neighbour filtering.
One issue is, even with extra4, the left and right walls don't get any significant dirt applied at the edge with the back wall. Will look into it..
|