Getting stuck in base models

[madpeople]

I was just wondering if they were built using newer DX SDKs if the newer formats would be supported automatically even if the code hasn't changed, the load x file may have been something in the dx API.

[LANS]

QUOTE (Raveen @ Apr 14 2010, 03:20 AM) My guess is that it's 3. A fairly simple if time consuming test would be to try and reproduce the effect in game. I think you'll find that you can't with any regularity.

Its easier to prove that its three by testing 1 and 2 for a continuous shape. If I had a .cvh file I could reasonably easily parse it into a piecewise 3-space surface function, and then check that for closed-surface continuity. Matlab++. Again, this is assuming that cvh files define a group of ellipses which define the hitbox(es).

QUOTE (KGJV @ Apr 14 2010, 03:35 AM) These CVHs were made with my AMT plugin which has a lot of 'guessed' stuff in it (it was made before Allegiance source release).

Today, the 'good' way to make CHVs is to use the official tools (cvh/mdlc tools) but this requires to have the models in X file format with the proper structures and names in place (specially for doors and mount points).

So the starting point would be to ask Orion to release GT models sources, convert them in X file if they're not then add the Alleg specific X frames in them, then use the tools.

What is the AMT plugin for/where can I find documentation on it? Who/what is Orion and where do I find him/her/it?

[Orion]

xD

I am an Orion.. I don't know if I have the model source anymore though (it's been like.. 6-7 years I think.. and I've gone through quite a few computers since then).

[KGJV]

Yeah he/it is our Orion. We only have one so be nice with him/it. As you can see, he/it is getting old , he/it forgot where he/it put his things

AMT is a plugin for milkshape 3D. It's in the downloads section. I think our Orion wrote once upon a time a nice tutorial for AMT too. But like I said this is old stuff made before we had the code source of Allegiance.

Check the wiki too, may be someone put some stuff in there.

If you have technical questions about CVH, feel free to ask here.

[LANS]

Thanks, but I don't really need the model source. Thats only important if I want the CVH ellipse definitions to line up with the graphical models themselves, which isn't something I care too much about as long as its good enough, which it is currently. Either I can fix the CVH definitions, or there's an issue with alleg's collision-detection.

So I need 4 things to get started:

1. Time (has to wait until I'm done exams - end of april)

2. CVH files themselves. Are these in my alleg folder, or where can I get the ones which are used?

3. Full documentation on the CVH file format. Its gotta be somewhere.

4. Matlab or another math program. I have access to these at my university's computer labs, but I don't want to buy my own copy. I'll find something, but it'll take time if I want to do this legally.

[KGJV]

QUOTE (LANS @ Apr 14 2010, 03:10 PM) Thanks, but I don't really need the model source. Thats only important if I want the CVH ellipse definitions to line up with the graphical models themselves, which isn't something I care too much about as long as its good enough, which it is currently. Either I can fix the CVH definitions, or there's an issue with alleg's collision-detection.

So I need 4 things to get started:

1. Time (has to wait until I'm done exams - end of april)

2. CVH files themselves. Are these in my alleg folder, or where can I get the ones which are used?

3. Full documentation on the CVH file format. Its gotta be somewhere.

4. Matlab or another math program. I have access to these at my university's computer labs, but I don't want to buy my own copy. I'll find something, but it'll take time if I want to do this legally.

2. yeah the ones in your folder are the same used on the servers (well i hope for everyone's sake:))

3. I once wrote a very short documentation here. Ask if you need more infos.

I also don't think the ellipse definitions are the cause of the problem but the convex hulls themselves.

For clarity, the alleg code uses 2 tiers to detect a collision:
1st the ellipse equation is used to determine if a more detailed, tier-2 detection is required.
If it is then the convex hulls are used.

If the ellipses are too big, it's not a problem, it'll just generate more tier-2 useless detections (except when shields are taken into account: the ellipses are always used then,there is no tier-2 so an 'ellipse hit' means a 'hull hit').
If the ellipses are too small, it could be a problem but AFAIK, AMT uses the bounding box as ellipse equation so *it should* be safe (but still worth verifying just in case).

The usual issues with CVH files are scaling (radius in CVH doesn't match model radius and scaling goes wrong) and the bitchy docking doors.

[LANS]

So Id guess that ship-station collisions are tier 2 then?

[Raveen]

Not necessarily, because the station most likely had shields up which (if I read Kage right) makes it a tier 1 collision only.

[madpeople]

Stations don't use their shields for collisions - it' why you can fly so close to them and only hit them when you hit near where the graphical model is (i.e. where the CVH is).

Ottherwise you would collide with them a long way out because they would be in a big shield bubble. Which would make docking rather tricky not being able to get near the doors and all
Which reminds me, was the carrier docking / shields raised for frendlies bug been fixed yet?

[raingriffin]

to be honest i've never ever done any CVH using the official tools. i've always used blender, my head and AMT. go guess...

now if you were to explain to me how to do use said tools, you'd probably help me gain the will to work for alleg back cause tbh i have none atm.
After all, you know, i have a couple dozens models of a new faction to cvh, and most of them are round somehow gah.