1  |   Order by default

Jul 31, 2006, 8:52pm

To let people know:
Create solid off does NOT contribute to lightening your rendering load. It
simply makes the object nonsolid for your avatar. The polys are still
checked by the enging for collision. However, collision IS a drain on
resourses and only the faces that you need to colide with should have
collision enabled, only, this must be done in the RWX. At the end of the
model simply put "collision off" in the last clump. It should be the last
clump because once collision is defined as off, it cannot be turned on
again. The end result of this meathod is that these polys never get to the
collision checker and thus saves system resources. I define collision for
at least some part of almost every object I make...it definately eases the
load. Just keep an eye out for stuff that you wont need to hit :)

BTW The last clump also has to be a child clump as well, dont know why, but
just make sure that it is nested in your model begin / model end somewhere,
in a clump begin / clump end :)

Also, you may consider ADDING surfaces just for collision. For example,
lets say you have a beautifull corinthian column (and its way high polys you
sinner :P ), build a box around it only as wide as the column (no top or
bottom of course, thats a waste!). In the RWX put the whole column in the
bottom child clump, turn off collision and thrn turn opacity to 0 for the
box. Vioala! You spared the poor exhausted collision detection all that
work and believe me you will not need that much detail just to be able to
slam into the object :) fine fine, if you REALLY want to be able to slide
arounde the column you could use a hexagonal cylinder or leave collision on
for just the shaft itself (without the base and capital) if the shaft is not
so heavy. You may be surprised that adding these few polys will likely
increase your framerate.

happy building!
-oleyo


[View Quote]

Aug 1, 2006, 5:52pm

Actually I have not taken a look at this in a long time so I cant verify or
deny the current potential performance enhances with solid off so take my
first sentance with a grain of salt. However the rest of my post still
applies and is definately good practice when building. Especially since
there are plenty of objects that must be solid but dont have to be entirely
solid.

-Oleyo

[View Quote]

Aug 1, 2006, 6:19pm

One thing to keep in mind is that AW is TOTALY different from games like
doom UT and so on when it comes to rendering. Because the whole point of AW
is to sort of stream objects and textures to you, it is designed around
constantly checking the state of the world at your position at any given
time (almost all games dont have to do this, the environments are almost
entirely static, but in AW you might move any object next to me and i will
see it happen) Also, FPS games like UT use what is called a bsp for the
level. When a level is compiled (when the level is actually built by
someone) it runs a calculation on the entire map and creates a table of what
polygons can be seen from EVERY single "cell" in the level. This
calculation can take a long long time depending on how complex the level is.
Thus when you are in the game the engine simply consults the table for each
cell you are in to see what is visible and what is not and saves all the
work from the engine. That is one huge trick that we just could never take
advantage of...it is not that the renderware engine cannot give good fps.
One thing that could help us would be different levels of detail for objects
and definable visibility, I discuss these a little bit in a reply to the
minimum visibility post.

Happy building!
-oleyo

[View Quote]

Aug 1, 2006, 6:00pm

Hmm well the point is for people to make their own avatars. Dont really
have the resources to make a hundred avs at the moment without ignoring
other concerns. :) And this feature is fuctional, but as many have pointed
out there are some technical and non technical questions to be settled first
like:

What physical size limit should be imposed?
What poly limit?
What limits on num of textures and texture sizes?
What should be allowed / dissalowed?


And as most have already imagined each avatar will need to be aproved
personally and added. As most are aware, issues involving censorship are
always touchy. There is just too much potential for abuse with offensive
avatars....not sure we are ready for that yet :)

[View Quote]

Jul 31, 2006, 7:57pm

I agree with you guys here. I often am frustrated by the limit imposed on
developers who have the ability to build discretely and to build
efficiently. My greatest desire though is not to simply increase the view
limit but to have a level of detail property for objects or at least a
definable draw distance. Then you could build say, a large tower with a
very long visibility, say 240m which i think is as far as the browser can
check for objects, or across a single "zone" of 9 sectors (I think it might
be able to check blocks of 5x5 or 20 sectors, but I am foggy in this area).
Either way you could save LOTS of resources by making your objects in the
tower or details render at only 30m or 20m giving you more flexibility in
world construction. It is likely that this will be the form that any change
in draw distance will take, though there has only been small experimentation
and discusion in this area. That being said, it is strange to me that the
minimum view that you can set is less than the maximum that the browser
allows you to set your view. After all, you can build your world how you
see fit within the scope of the browsers capabilities. For example, we
never prevent people from dropping in ugly 40,000 poly 3dsmax objects from
the web into their worlds if they are bent on doing so ;)

-Oleyo

[View Quote]

1  |   Order by default