DRD Team

Understanding that GZDoom is going under some huge changes right now, I'm not going to press this too much since Graf probably has a bigger fish to fry. I believe enjay has also requested something similar, but I'd like to re-iterate this because it would be extraordinarily useful.

A lot of MD3 models utilize multiple skins for one model, and gzdoom will only allow for 1 skin per model, which at the time is severely imparing a project I'm working on. Normally this isn't an issue, since converting multiples skins to a single image isn't all too difficult. However, converting a skin for an animated model is not possible at all, and as such, is causing me problems.

Also, I remember seeing something about this too, but if brightmaps could be applied to model skins, that would be fantastic.

What really needs to be done is the MD3 feature needs to be re-implemented properly 100% (or at least all features except tags and skeletal animation as that wouldn't make sense in a Doom engine).

That includes...

1) Multiple sub objects (groups) and materials per single model
2) Quake 3 shaders (for fancy animated effects on model textures)
3) Vertex lighting (though this is more related to the renderer than the model code alone)

Also an improvement to the MODELDEF structure that would allow the modder to assign a range of frames to a state (for example, assigning frames 12-24 to the See state of the actor) so that defining long, smooth animations wouldn't be as cumbersome as it is right now.

Of course, this has all been discussed way too many times and the general answer is that unless someone with 3-d graphics coding knowledge steps in to help Graf, we won't really see anything added to the current model feature set anytime soon... (which is perfectly understandable -- even more so with the current renderer rewrite )

All I want from modelz is smooth rotation and vertex lighting

Paul wrote:All I want from modelz is smooth rotation and vertex lighting

Both of which could be "borrowed" from the Doomsday source if need be...

Nash wrote:What really needs to be done is the MD3 feature needs to be re-implemented properly 100%

I'd see no point in implementing anything better than what JDRP requires unless you're a very talented artist with lots of spare time and willing to create a brand new set of improved models for Doom.

I'll say it here but this goes for all other model related features as well:

Nothing of this will happen unless I find someone to help me out. The model code is in dire need of a lot of work which I simply can't invest here.

As such it makes no sense to request any model-related enhacements as they won't get implemented if nobody comes forward.

I kinda figured this was the case. I might bug Torr Samaho about this, but the current model support works well enough to be useful.

....though vertex lighting would be really great as having a larger model object being completely lit up by a point light that's just close to the models center sucks

My little demonstration of multiple skins per model, implemented in ST by Torr.

Nice. Can you please post your modeldef lump as an example for me to see how to allocate skins to different groups in the model.

Do I see smooth alpha on the muzzle flash? HOW IS THIS POSSIBLE!!? I thought models can only have 1-bit alpha textures!

Nash wrote:Do I see smooth alpha on the muzzle flash? HOW IS THIS POSSIBLE!!? I thought models can only have 1-bit alpha textures!

For some reason Line_horizon makes it look fine. But if there are particles in view, the MF goes back to 1-bit
And I am extremely sure about this.