Screenshot Thread

[Xarg0]

@Apoc
Just out of curiosity, would you mind to explain how your navmesh genertation algorithm works?
I'd first filter the graph for triangles you can't stand on and afterwards check if the connected triangles are traversable, or can you safely asume that all connected triangles you can stand on are traversable?

Edit: Stupid me, ofcourse it's possible that there are obstacles between to triangles you can stand on, but I'd guess you still would only need to check triangles around fences, trees and the like...

[MaiN]

@Apoc
Just out of curiosity, would you mind to explain how your navmesh genertation algorithm works?
I'd first filter the graph for triangles you can't stand on and afterwards check if the connected triangles are traversable, or can you safely asume that all connected triangles you can stand on are traversable?

Edit: Stupid me, ofcourse it's possible that there are obstacles between to triangles you can stand on, but I'd guess you still would only need to check triangles around fences, trees and the like...

They are using Recast & Detour.

[MoBob]

Hi, first post in here... so... Hi!

@Apoc: That looks great and promising

I'm working on something very similar but I have some problems positioning the WMOs correctly. Almost everything looks just fine, but Orgrimmar seems to miss some degrees. It would be great if you could give me a hint on what I'm doing wrong:
I'm using the coordinates of the WMO (looks just fine) and rotate on Y-axis by (orientation2 - 90), Z-axis by (orientation1 * -1) and X-axis by (orientation3) where the orientations are also read from the WMO (ofc).

Thanks and sorry for starting my career here with a question, will try to help where I can

[Cypher]

I reset my rank and stats in MW2 and decided to see how badly I could dominate using only the fail builtin classes. I used the built-in Famas/Shotgun class, except I added an extra restriction which was that I was not going to use the grenade launcher attachment on the Famas.

This was the result of that experiment:



Rifk.

Now I'm back to being bored. I started work on my packer again, but I want a couple of other things to work on alongside it and I can't really come up with much that I'd find fun.

[namreeb]

@Apoc
Just out of curiosity, would you mind to explain how your navmesh genertation algorithm works?
I'd first filter the graph for triangles you can't stand on and afterwards check if the connected triangles are traversable, or can you safely asume that all connected triangles you can stand on are traversable?

Edit: Stupid me, ofcourse it's possible that there are obstacles between to triangles you can stand on, but I'd guess you still would only need to check triangles around fences, trees and the like...

What you can do is filter the triangles that are too steep to traverse. You can also remove regions (as in, interconnected mesh polygons) which are 'too small'. One thing we cannot do yet, however, is specify a traversal cost (or obviously a cost threshold) for a given triangle. Ultimately this will be needed for say determining if we should swim through a lake versus walk around it.

[Xarg0]

What you can do is filter the triangles that are too steep to traverse. You can also remove regions (as in, interconnected mesh polygons) which are 'too small'. One thing we cannot do yet, however, is specify a traversal cost (or obviously a cost threshold) for a given triangle. Ultimately this will be needed for say determining if we should swim through a lake versus walk around it.

There are certain ways to determ if a triangle is part of road a lake or some wilderness, based on those information it's very easy to set up a cost, yet it's some work to optimize these cost settings for good pathing.
Iirc there is some kind of enum in the adt files that specificate weather the triangle set belongs to a road lava or water.
My main concern is about how a meshgenarting algorithm should be shaped, so it's not too costy, if I'd start doing collision checks at a certain position and then contiue to check collisions on all connected triangles I'd have to keep a list of checked triangles or some kind of open list, I could also just iterate through a list of all triangles and check if I can stand on them, and afterwards check for collisions between connected triangles, yet I'd run into the problem that some triangles are to small to stand on so I'd have to first add triangles that aren't to steep together to bigger regions, wich would also create quite a lot of proccessing load.

[Apoc]

Woot! Lighting works now!



Win!

[Cypher]

Nice.

What are you using for all the rendering btw?

XNA? DxLite (or w/e its called)? Something else?

[Apoc]

XNA of course.

[Cypher]

XNA of course.

Failure.

EDIT:

Btw, this is you:

[namreeb]

There are certain ways to determ if a triangle is part of road a lake or some wilderness, based on those information it's very easy to set up a cost, yet it's some work to optimize these cost settings for good pathing.
Iirc there is some kind of enum in the adt files that specificate weather the triangle set belongs to a road lava or water.
My main concern is about how a meshgenarting algorithm should be shaped, so it's not too costy, if I'd start doing collision checks at a certain position and then contiue to check collisions on all connected triangles I'd have to keep a list of checked triangles or some kind of open list, I could also just iterate through a list of all triangles and check if I can stand on them, and afterwards check for collisions between connected triangles, yet I'd run into the problem that some triangles are to small to stand on so I'd have to first add triangles that aren't to steep together to bigger regions, wich would also create quite a lot of proccessing load.

Afaik, the only way to tell if a given triangle is a highway is by having a list of textures which correspond to a highway and checking if that triangle has that texture.

As for traversal cost relating to water versus land, if you specify the cost in terms of traversal time, you should have no problems. Say a land triangle is 1.0, then water would be 0.5 or w/e. The only catch is that I believe WoW projects all triangles onto a 2D surface when it comes to calculating traversal time. That is, a flat triangle that has a maximum span of 3 yards will be traversed no more quickly than a triangle 20 yards tall, but with a maximum lateral span of 3 yards... assuming the resulting slope were walkable =)

[flo8464]

Failure.

EDIT:

Btw, this is you:

Apoc is a girl? :confused:

[Xarg0]

Failure.

EDIT:

Btw, this is you:

Apoc is a girl? :confused:

I don't see no girl in that picture :confused:

[Robske]

Yo Apoc, I'm really happy for your Disco thing and Imma let you finish but this roflcopter is one of the best visuals ever.

http://img26.yfrog.com/img26/5098/wo...3009152747.jpg

[ame="http://www.youtube.com/watch?v=Bq2JYYtBnTw"]YouTube- DirectX: Roflcopter[/ame]
Use HD! Source: NoMorePasting.com

Todays course about finite fields wasn't relevant to my interests so I whopped up this on paper I was going to create a colorfull pulsar/radar initially but went with this

Edit: Lulz, encorporated this into my XYZ hack. This makes going upwards to Illidan/C'Thun THAT much cooler.

The yellow is the WMO (Orgrimmar in this case)

The pink is our nav mesh. And yes; pathfinding works.

Nomnomnom such awesomeness.

Is your mesh in an optimized state? As in a minimal number of convex polygons or do you take the triangles straight from the ADT files, or hell, does Recast do all the optimizing for you?
On a sidenote, would Recast be able to generate a complete mesh out of a regular trianglelist? I've noticed that the demos seem to be using one stored like this:

v x0 y0 z0
v x1 y1 z1
f 0 1 2
f 3 4 5

If that is the case, then I would hug the recast devs.

[Cypher]

Oh my god.

That... is... AWESOME!

Btw, lol @ the "Imma let you finish" shit. Rifk.