Page 5 of 12

Re: Creating a clothes library: some technical aspects

PostPosted: Fri Feb 10, 2012 5:10 pm
by robertltux
mind sharing your shoes??

Re: Creating a clothes library: some technical aspects

PostPosted: Fri Feb 10, 2012 8:25 pm
by duststorm
Here's all the source files.
I made clothes packages for MH, but they're not finished yet.

As usual, blend source is attached and here is the source gimp file: http://duststorm.ulyssis.be/mh/shoes_te ... ources.zip

Re: Creating a clothes library: some technical aspects

PostPosted: Fri Feb 10, 2012 8:38 pm
by robertltux
thanks do you mind if i use them to make some ballet slippers??

Re: Creating a clothes library: some technical aspects

PostPosted: Fri Feb 10, 2012 11:09 pm
by duststorm
A proper license needs to be found for them I think. At the moment, as they are part of the MH project, they share the same license as everything else that comes out of MH.

Re: Creating a clothes library: some technical aspects

PostPosted: Sat Feb 11, 2012 7:29 am
by Lemon
duststorm, that's just great! You've started that important library-enriching process, which was highly expected by all MH users, I believe :)

Re: Creating a clothes library: some technical aspects

PostPosted: Sat Feb 11, 2012 8:00 am
by ThomasL
Hm. Actually, you have released your clothes under GPL3. At least that what it says at the top of the mhclo files.

The makeclothes script looks for a file called proxy_header.txt in your home makehuman directory, and copies its content at the top of the exported mhclo file, with lines prepended with a #. If this file is not found, some default lines are printed instead.

Re: Creating a clothes library: some technical aspects

PostPosted: Sat Feb 11, 2012 12:38 pm
by duststorm
Aha. Apparently I haven't paid a lot attention to that yet.
Anyway, that seems like a good default.

There's still a lot to fix up before these can be called final. I even don't have masks yet..

Re: Creating a clothes library: some technical aspects

PostPosted: Sun Feb 12, 2012 9:50 am
by ThomasL
duststorm wrote:Aha. Apparently I haven't paid a lot attention to that yet. Anyway, that seems like a good default.


It was not the intention to trick unaware clothes-makers to use GPL license. I just put the info there a long time ago when I was the only one using the script. Setting the author/license info has been moved to the bottom of the panel, where it is more visible. Like the other properties, they are stored permanently when you press the Save settings button, so you don't have to retype it.

Re: Creating a clothes library: some technical aspects

PostPosted: Sun Feb 12, 2012 1:55 pm
by duststorm
And I've created a dfferent type of shoe.
classicShoes.jpg
Classic shoes for MH

The low poly version is 182 quads per shoe.
The high poly version 728 quads per shoe.
topo.jpg
Topo, low poly version

topo_hi.jpg
Topo High poly version

I reused the topo of the sports shoes as base, but modified it. I think I added 1 or 2 loops, the polycount is a little higher than the sports shoes.

Re: Creating a clothes library: some technical aspects

PostPosted: Sun Feb 12, 2012 2:45 pm
by duststorm
Thomas, I have some more questions to fire at you.
Trying to assign the original uv mapping of the basemesh to the shoes gave me some pain.
Of course, mapping the shoes to the feet area is probably (together with the hands) one of the most troublesome cases.
shoe_basemesh_seams.jpg
Difficult to map uv seams

As the picture shows it's quite hard to fit the seams to the shoes decently. I also feel that adding extra edges to the shoes would ruin the topology of the shoe mesh, so I'm wondering if there could be a better approach.

My thoughts: since I'm not using the basemesh uv mapping for texturing, it probably doesn't have to be perfect. It will just have to match the areas of the body it covers as best as possible.
So I was wondering: what if the clothes script could have an option to just copy over the uv coordinates using the mapping of clothes to human mesh vertices it made? So each vertex of the shoes would receive the middle uv coordinate of the basemesh triangle it is matched to.
This would at least give a good starting point which could be edited manually.
What do you think?