Wall Worm Forums

Yes I think you have it. The HL2 skeleton rigs and the CS:S ones are similar but NOT the same... and even one extra bone I think would mess everything up.

If you want to use the CS:S animations; decompile one of the T or CT models, import into max (using wunderboys importer) and delete the mesh so you have just the skeleton. Now take your model and add a skin modifier with the skeleton from the decompiled model. Weight your model the way you want. Then when you recompile just add to your .qc the following

$includemodel "player/cs_player_shared.mdl"
$includemodel "player/ak_anims_ct.mdl"   OR   $includemodel "player/ak_anims_t.mdl"  (depending whether you used CT or T model)

If you want to use the HL2DM animations then you must do the same with a charater from that game instead.

What is it you do want to exactly?

I just spoke to shawn and he explain to me a bit clearer. If you are trying to use the existing animation files to animate your model then all you need to do is add

$includemodel "<path>/<filename>.mdl"

to the qc file.

The only thing is that the skeleton HAS to be identical - normally its best to use a decompiled one.

To be honest Captain, I haven't tried importing player models into Max using the WWMT importers for 3 months or more.  The last time I tried it seems that the WW importers were simply not capable of handling fully skinned, rigged character models and/or animations.

.smds that imported fine using Wunderboys smdimporter would lose all (or most) of the bone info and none of the skinning info would be translated either when I tried with WW. This was using L4D2 character models (which have a slightly newer format than CS:S models) but I think the same applies.

Since I last tried to import a character into Max I have mentioned to Shawn once or twice that it would be nice to have a system that could handle character models correctly - but so far as I am aware he hasn't made any changes to the importers since then. He's added a lot of other things but I don't think he's had time to work on them and I think that for him its not a major priority.

Unfortunately Wunderboy no longer supports smdimporter and the last version of max that it is compatiable with is 2012. So in 6 months when max 2015 comes out max 2012 will stop being supported and we will find ourselves with no importer for supported versions of max at all. :(

I hope Shawn will be able to find the time to get the importers working correctly for character models, up and including the GO engine - but I really have no notion of how much work this might be. I know some people who were working on a mdl decompiler called Crowbar and I think it's quite complicated... so getting an smd importer that is capable of correctly reading all the data in the .dmx and .qc files maybe quite a bit of work.

For the moment if I want to import a character model from an smd file my solution is simply to use max2012 with wunderboys importer http://www.wunderboy.org/apps/smdimport.php ... sometimes the animations can get a little weird, but normally the mesh, bones and skinning info comes out ok.


*So far as I know there is nothing that will allow animation files (.mdl and .ami) to be decompiled into a format that can be imported into Max. Most character animations are kept separately in these files rather than being compiled directly into the model as with weapons.

I just tried recompiling the qc files I sent you by simply pumping them back into StudioMDL and the outputted model seems to be absoultely fine. In model viewer some of the animation sequences seem to glitch out, but the same happens if you load up the original model.

So it seems to be working fine for me, if you have problems with this then I think it must be something you are doing.

I sent you a reply to your pm but I dunno if you got it (it said it was sent but isnt showing in my "sent messages")

Just in case not here is the link to the decompiled ctm_fbi:

https://www.dropbox.com/s/kdosuj8raf7pn7t/ctm_fbi.rar

I am worried it maybe vry similar to mdldecompiler - comapre and lemme know, if this doesnt fix all problems I will take a closer look and talk to Crowbar people.

Mdldecompiler is very out of date, and whilst it can sort of decompile CS:GO/L4D2 models quite a bit of the data is often either missing or wrong. I am not sure that what you see in Model Viewer can completely be trusted always... I was testing a model recently that worked fine in-game but when I looked at the collision in MV parts of the mesh got stretched out for no apparent reason.

Atm I am testing the Alpha version of a new model decompiler for Source called Crowbar... the programmers are designing it to hopefully be up to date with the latest Source models, now I think they are focusing a little more on L4D2 than CS:GO models, but it should still be more suitable than mdldecompiler.

Now whilst this program is in Alpha version I cannot give it to you as I promised the developers I would not distribute it, but I cannot see any harm in me using it to decompile the model and then send you the outputted qc and smd files. I can't promise that will fix all your problems, but it seems like a good place to start.

Tell me which model exactly you are working on and I will decompile and put a dropbox link for you.

Tested this fix and it works fine, thanks Shawn!

Would it be a lot of work to add a material prompt option when importing models? It's just a small point though, only worth doing if it's easy to add.

I like the way that this names the model after the SMD rather than just putting the same generic name for each model - again a small thing but it's a pain to have to change the name each time when importing multiple models. On that note is it possible to add a checkbox option to create a new max layer to put the model into when importing? - the layer name can just be the same as the modelname.

Glad you are happy with wallworm.

Just to clarify a few points for info.

Normally you cant use selfillum in combination with a transparent diffuse map in source, ie you cant have $transparent 1 and $selfillum 1 in the same vmt. You also cannot use separate specular and normal maps for the same basetexture. If you want normal and specular you must add the specular map to either the alpha channel of the normal map or the basetexture, doesnt matter which... if your basetexture is transparent in parts then you can still use its alpha channel for specular reflections, its just that the opaque areas will be completely matt and the transparent areas will be reflected to the amount of the transparency, which limits you more.

Usually to compile normal maps you should turn off "nice" filtering, also a good thing to do for decals if you get "halos" around them as you move away.

All reflective surfaces need $envmap and will only reflect if you place cubemaps and build them (dont forget to build twice in HDR environments). $envmap can be a little tempremental on certain model textures, especially when they are also alpha $transparent 1... you can replace env_cubemap with an environment map texture if necessary and if you get transparency issues try replacing $transparent 1 with $alphatest 1

You can also use the $additive 1 parameter as an alternative to selfilllum in certain situations, and although I dont think it works for models you can also add your texture path/name to the .rad file and make it emit real light.

Anyway, nice looking work, keep it up :)

When you select the "use local origin as world origin" function in WWMT, the pivot point of your model becomes the origin in Hammer (so wherever the pivot point is in relation to your model in max, the orgin will be in relation to your model in hammer)

However there seems to be an issue concering physics props; that the pivot point in max is incorrectly exported, and when you put your model in Hammer the origin isnt where its suppost to be.

There is an easy fix for this:

With your model selected, in the WWMT diaglog click on the "open folder" button next to the Model Folder Path. In that folder you will find the modelname.qc file for the model you are working on. Open this file in notepad. Find the line beginning with "$origin". Now in Max, simply select your model using the "select and move" tool (making sure you are in "View" or "World" mode for the Reference Coordinate System). Now if you look at the bottom of the screen, the x,y and z numbers should correspond with the first 3 numbers wirtten in the "$origin" line in the qc, if they do not, change the 'qc to match the numbers given in Max.

Save the modelname.qc file and the find the modelname.bat file. Click on the modelname.bat file to re-export your model with the correct origin.

Bare in mind that if at any point you re-export your model via the WWMT tools then the .qc file will be updated and the incorrect coordinates re-copied into it. You can prevent this from happening by checking the "lock qc" box in WWMT... however if you do that any changes you make to your model may also not be saved into the qc file.

Happy modeling,

Gmaps

Absolutely amazing workftp://. You people have to try this, you really won't believe the possibilities it opens up... If nothing else it means you can make models whilst still having a social life. So next time you're relaxing in a bar because you didnt have to make the .qc, the mesh, the lods etc, etc etc, just remember Shawn hard work and donate a little so he can  have a. beer too. Oh, did I mention?! You can also make you maps in max!!!!!

Wow, look at those sweet, fluid curves, the elegent cheekbone structure and provocative hips... definitely a chick dinosaur that one... I can almost see the pink lipstick and a crocodile skin handbag (or is that like canabalism for a pliosaur?)

Looking good! Can you put another photos with a different colour, kinda hard to see in detail. I'll have a little boyfriend ready for this one in a few days... :)

Looking sweet :) You should consider turning the spine/ribcage into a separate model and make it large enough for players to run through... you could have a few of those littered around the level, half-buried in the dirt