EVIL goes off to build dah BIG HL² FAQ, please gimme questions to put in the FAQ

feel bad for ppl that make maps, they make a huge ass map, with tons of buildings, they boot the level up, and everything just crumbles to the ground lol
 
::1 piece of wood falls, causing a huge chain reaction completely leveling the entire map again::
 
lol.. ahhwell.. i am off to bed.. cya 2 morrow..
 
The fire is interesting but maybe that explosions are gonna be a lot better :)
 
Well.. I am back :) had a very nice sleep. (I dreamth about those wierd sheepswine like creatures that anakin was riding in Starwars episode 2 in the ooh I love you so much amidala, rolling down the gras scene)

edit: this is the one

http://www.theforce.net/scifi3d/starwars/wip/shaak(Teyon)/shaak.jpg
 
Heh, i just watched that film last night on Box office. The "I love you and now im going to cry" bits in films are so annoying sometimes
 
haha yeah.. the love stuff is the most crappy part.
 
First, about shooting down a wooden house. I think youre probably going to build it one board at a time and stack them on top of each other. Sort of like building a house out of domino's. Otherwise I believe a welded section will probably stay together except where broken by bullets. I highly doubt they are going to include the required physics calculations to simulate stress fractures due to excessive loading. I'd be realy happy to be proved wrong on this one, realy I would :)

My only burning question is:

The Dynamic LOD system. Will the model LOD's be modeled by the model builder or will the mesh detail be reduced by the engine itself? And If the engine does the polygon reduction will it be done on the fly, or will all the models have their LOD calculated ahead of time (like when compiling the map, or when starting a new level) ?
 
Originally posted by Punisher454
My only burning question is:

The Dynamic LOD system. Will the model LOD's be modeled by the model builder or will the mesh detail be reduced by the engine itself? And If the engine does the polygon reduction will it be done on the fly, or will all the models have their LOD calculated ahead of time (like when compiling the map, or when starting a new level) ?

read my post here http://www.halflife2.net/forums/showthread.php?s=&threadid=1154&perpage=15&pagenumber=6
 
read my post here

Ya, I get it that the lod switching will adjusted by the engine to maintain a decent FPS. I understand that. But My question is about the models themselvs.
To have a model with multiple LOD's theres basicly 2 ways to do it.
1. The model maker will build multiple versions of the same model, but with different polycounts and then combine them all into one file. For example, you could have a player model with polycounts of 6000, 5000, 4000, 3000, 2000, 1000, 500, 250, 150, 50, 10, 4 polygons per level. the 4 poly model is the most distant and would just be a skinny pyramid in order to light up 1 pixel on your monitor at extreme range (we probably wont have ranges that far away). Usualy with these types of multiple LOD models the model maker can specify the ranges at which the lod's switch. HL2's LOD system is said to be "Dynamic" so I would assume the model maker does not have distance settings and the engine selects the highest detail model it can without lowering the FPS below a certian level.

2. The model builder just builds one high detail model and HL2 performs the poly reduction on its own. Sounds simple, its for sure alot less work for the modeler, but it also gives you alot less control over the look of your model (poly reduction does not always simplify the mesh in the most desirable way).
IF number 2 is the case, then the next question is:
Are the simplified LOD's generated in realtime or are they generated ahead of time, like when the map is compiled or when the level is loaded up?

The answer to these question will determine if modelers need to do alot more work if they want to have some nearly ready models by the time we get the tools to do more.
--OR---
If model builders building multiple LOD's on their own are doing a tremendous amount of extra work.

Multiple LOD models have been common in Flight Sim's for a very long time. On the 2 different Flight Sims I have worked on models for, the model builder builds all the LOD's and selects the distance at which the switching takes place. HL2's LOD system is dynamic so that takes the distance settings out of the modelrs hands.
 
I think its no2 because no1 is very hard.. and it would totally suck if you are mod supportive!
 
here's another question:

how advanced is the collision detection in Halflife2? for example will bodies stack on each other or will they clip thru each other?
 
If you remember the E3 video, he showed the demonstration of bodies interacting with complex surfaces, if they can interact with complex surfaces, I'm almost poistive you can stack bodies.
 
well...some of the screenshots have a soldiers foot cliping thru a body....but they might have done so that living creatures won't collide with bodies, so that it will be easier for the AI getting around, not having to navigate around bodies....really hope that we won't see much clipping.
 
You must remember: that video is year old content.
 
Yes! Keep in mind that what we saw in the video compared to HL2 is like this compared to HL1. :)
 
Originally posted by guinny
You must remember: that video is year old content.
Exactly. I bet a lot of stuff in the video will change - for the better -, i dont know how it can get any better though.

/me cums just thinking about it
 
Back
Top