Page 3 of 3

Posted: Sun Aug 24, 2008 10:43 pm
by beeman
madpeople wrote:QUOTE (madpeople @ Aug 24 2008, 04:41 PM) a cartoon bomb?

a light bulb?


both 5 minute jobs, so not fantastic looking
mp...you showoff /laugh.gif" style="vertical-align:middle" emoid=":lol:" border="0" alt="laugh.gif" />

Posted: Sun Aug 24, 2008 11:22 pm
by madpeople
problem is i wouldn't know how to make them look better.

The bomb is just a circle with a off centre radial fill of white to black, and two 2px wide grey lines making the fuse with some black pencilled diagonals and a yellow star on the end.

The light bulb is just a yellow oval with a grey rectangle over the bottom (and a smaller rectangle at the very bottom so it tapers) and some pencilled in black diagonals and black bit at the bottom.

I tend to work very zoomed in (3200% - so it fills the screen) and tweak individual pixles to make it look right

here are some grey lines with some pencilled in darker grey dots

(a girder)

Posted: Mon Aug 25, 2008 12:01 am
by KGJV
@Spinoza
I'm no art expert so they look fine to me when reduced to 16x16.

I should eventually change the code so they display at 32x32. I guess most ppl use high rez desktop these days so 16x16 might be too small.

there is no real hurry for the icons. I posted early so contributors can have time to do them when they can. Do what you can when you can.

group icon = it's just a group of meshes to symbolize a tree hierarchy . In my OP screen I used a generic tree icon, I guess I could keep it.

current "effects" in alleg = engine glow, trail, smoke & rocket effect of afterburners so except for 'trail' the cartoon bomb could fit.
I can eventually use a different icon for each effect type (so 4 icons instead of one) but for simplicity I started with 1 only. Eventually an icon with the letters 'fx' in it could do the job.

for lights, the bubble is fine i guess.

Oh and I could use a program icon for AMT 2 too /wink.gif" style="vertical-align:middle" emoid=";)" border="0" alt="wink.gif" />

@MadP

I'll add the light & bomb icons for now but if we go for 32x32 I'll wait for Spinoza's stuff. once again no hurry here.

yes all modern .x formats files I tried load fine. I use the native DirectX 9 routine to read .X files.

CVH boxes can either be generated (same qhull algo) or imported as .X files. I don't want to add 3D editing of mesh (aka modeling). the only 3D editing in ATM2 are the position / orientation of alleg specific datas.

yes for light, even more parameters (ramp up/ramp down/hold too). You can already set light phases with current tools (.mdlc/makegeo). See my post here but you cant currently read them back and change them from .mdl files (well you might eventually can with milkshape and ATM 1 but I haven't tried). AMT 2 will alllow this.

@Orion:
I want to avoid native/managed interop if possible. If you have any pointer/code for pure managed convex hull that would help. Like I said I looked at CGAL library but it's native only.

Posted: Mon Aug 25, 2008 12:10 am
by KGJV
hum I tired 32x32, it's actually bigger than I thought /smile.gif" style="vertical-align:middle" emoid=":)" border="0" alt="smile.gif" />



not sure we can keep it that way, 16x16 might be the size to go.

Posted: Mon Aug 25, 2008 1:54 pm
by Orion
Hmmm... pure managed and Physics APIs dont exactly go hand in hand

U realize though that SlimDX uses interops, right? ;D

I use a wrapper for Newton Dynamics thats pretty cool, but I understand ur desire for purely managed code (we all strive for it;)). There is probably some stuff out there "for XNA" you could use. That, as I'm sure you're aware, is typically just a matter of removing the XNA references it uses and putting in managed directx (although with Slim DX it might be harder, I havent looked at the naming convention).

P.S. will this be open source? /tongue.gif" style="vertical-align:middle" emoid=":P" border="0" alt="tongue.gif" />

Posted: Mon Aug 25, 2008 5:00 pm
by KGJV
Orion wrote:QUOTE (Orion @ Aug 25 2008, 03:54 PM) Hmmm... pure managed and Physics APIs dont exactly go hand in hand

U realize though that SlimDX uses interops, right? ;D

I use a wrapper for Newton Dynamics thats pretty cool, but I understand ur desire for purely managed code (we all strive for it;)). There is probably some stuff out there "for XNA" you could use. That, as I'm sure you're aware, is typically just a matter of removing the XNA references it uses and putting in managed directx (although with Slim DX it might be harder, I havent looked at the naming convention).

P.S. will this be open source? /tongue.gif" style="vertical-align:middle" emoid=":P" border="0" alt="tongue.gif" />
well SlimDX uses interops inside SlimDX . When you 'reference' SlimDX you're in the managed world only. /tongue.gif" style="vertical-align:middle" emoid=":P" border="0" alt="tongue.gif" />

The thing is that I HATE to write wrappers , that's what I meant when saying "avoid native/managed interop " /smile.gif" style="vertical-align:middle" emoid=":)" border="0" alt="smile.gif" />

Initially I wanted to use XNA, but the "content pipeline" is just a pain for creating editing tools so I went with SlimDX since MDX is deprecated. Anyway the 3D rendering I need is very simple. (I'm even considering using WPF 3D later on ...)

I'll deal with the convex hull code at the end anyway so meanwhile may be I'll find inspiration to do it in a simple way.

yes it'll be open source and available in the FAZ SVN.

Posted: Mon Aug 25, 2008 5:30 pm
by Orion
There are lots of wrappers for Newton, take ur pick /tongue.gif" style="vertical-align:middle" emoid=":P" border="0" alt="tongue.gif" />

I've never tried grabbing vertices out of a convex hull after generating it before, but the idea I had was to is lock your mesh, loop through the resultant buffer and create a float array out of all the vertices, then call 'NewtonCreateConvexHull', then 'NewtonBodyForEachPolygonDo'... which would do a callback for each polygon of a convex hull. I've got code that will do all that except the 'NewtonBodyForEachPolygonDo' call.

Posted: Tue Aug 26, 2008 9:28 pm
by madpeople
will LOD be supported? - allowing people to select multiple .x files and say when that LOD (assign a radius of px on screen the object needs to be smaller/larger than for the selected model to be used) should be used and which the CVH should be generated from / which contains the cvh boxes.