Proper md2 to mdl conversion?
Proper md2 to mdl conversion?
Hi, as title says, I was trying to convert md2 to Hexen 2 mdl. It was the weapon model(Elven Wand from Heretic). What is the proper way to convert these two formats? Because after I converted it with Quark model editor, both the model and the texture were messed. I tried it the other way, converting it to md3 and then to mdl. While the model was OK, texture was messed again. So is there some better converter or did I missed anything? Thanks.
Re: Proper md2 to mdl conversion?
hexen2 has a different palette. you'll need to find a converter that either explicitly supports hexen2 or has an option to change the palette.
Re: Proper md2 to mdl conversion?
That´s sadly not the case
I´ll post the screenshot which says it better.
I´ll post the screenshot which says it better.
Re: Proper md2 to mdl conversion?
looks kinda like uv maps are getting messed up. i cant really tell what the model is. but you could try noesis.
Re: Proper md2 to mdl conversion?
This is what the model was supposed to look like.
So I guess it is not just UV maps.
So I guess it is not just UV maps.
Re: Proper md2 to mdl conversion?
Have you tried http://icculus.org/qshed/qwalk/ ?
Improve Quaddicted, send me a pull request: https://github.com/SpiritQuaddicted/Quaddicted-reviews
Re: Proper md2 to mdl conversion?
Hm, same result with Qwalk. What is strange is that the model looks fine in model editor, but in game it looks like this.
Re: Proper md2 to mdl conversion?
Hello Mathuzzz,
I think I have seen you in the Quakeone Hexen2 thread.
Good to see you working on H2.
Regarding your question:
I think it is an engine dependent issue.
I quickly converted the Heretic wand model to .mdl (like you did) and got these results:
With FTE engine:
With UQE engine:
With DP engine:
With jhexen2 engine I received a similar result like you have (model is totally messed up).
So I think it is the replacement feature of the engine, that causes this.
Anyway, this is what I did:
- Converted .dmd model from Cain / PiCKleBro into .md2 with "md2tool"
- Converted .md2 model to .mdl with "Noesis" (you can also use "qwalk" for this step)
- Created matching replacement texture with "QME"
It was necessary to create a replacement texture, due to different color palette (just like Spike said !).
Without replacement texture, the colors do not match to Hexen2 palette.
I hope this could help you a little bit.
Seven
I think I have seen you in the Quakeone Hexen2 thread.
Good to see you working on H2.
Regarding your question:
I think it is an engine dependent issue.
I quickly converted the Heretic wand model to .mdl (like you did) and got these results:
With FTE engine:
With UQE engine:
With DP engine:
With jhexen2 engine I received a similar result like you have (model is totally messed up).
So I think it is the replacement feature of the engine, that causes this.
Anyway, this is what I did:
- Converted .dmd model from Cain / PiCKleBro into .md2 with "md2tool"
- Converted .md2 model to .mdl with "Noesis" (you can also use "qwalk" for this step)
- Created matching replacement texture with "QME"
It was necessary to create a replacement texture, due to different color palette (just like Spike said !).
Without replacement texture, the colors do not match to Hexen2 palette.
I hope this could help you a little bit.
Seven
Re: Proper md2 to mdl conversion?
Hi Seven, yes I use jsHexen and I thought the engine might be the case.
Thing is that I think jsHexen is the best engine for Hexen 2. UQE is OK, but has fewer features and with more recent release it has strange movement physics for weapons. Some of them are "flying" in front of you and move differently when you walk. (even though I like the jumping there more ).
Thing is that I think jsHexen is the best engine for Hexen 2. UQE is OK, but has fewer features and with more recent release it has strange movement physics for weapons. Some of them are "flying" in front of you and move differently when you walk. (even though I like the jumping there more ).
Re: Proper md2 to mdl conversion?
glquake has a few limits that if exceeded can result in memory corruption, I guess that's what's happening here. Try simplifying the model down to <512 verts? might be less, not sure what the actual limit is.