So, after much experimentation, I found a viable approach for scaling up a male Human head to fit the Qunari body. It's not quite as straight forward as the Human-sized Elf I described in the previous post. This one requires a bit of massaging to get the mesh in just the right shape, size, and position. As a result, there's a bit of deformation from the original head, but it was necessary in order to get everything to line up (Qunari head is red, Human head is green):
You can see there that the most important facial features - the eyes and mouth - line up pretty well. The front and sides of the neck line up as well, although the back of the neck does not. This results in a small gap in this area with the naked body model, but fortunately it's is virtually unnoticeable unless you go looking for it. Anyway, on with the show!
So to start with we need our morph. You can go with whatever shape you like, but for illustration purposes I'm going with everyone's favourite prematurely dead mentor, Duncan:
I chose him for two reasons. The first is that I'm both lazy and unskilled with the morph editor, so the shape presets alleviated both those problems. The second is that I wanted to include a beard in the process to illustrate how to deal with those, and Duncan was the most worthy candidate in that regard. You'll see in the image that I highlighted the specific hair and beard models required for this example (unique hair/beard models that only Duncan uses).
As was the case with the Elf head, save your MRH and export the XML. With that done, you can close the toolset for now. Open up DATool and extract the following models -
Human male:
hm_brd_duca_0/2/3 - Duncan's beard
hm_har_blda_0/2/3 - Human scalp
hm_har_duca_0/2/3 - Duncan's hair
hm_uem_basa_0 - Human eyes
hm_uhm_basa_0/2/3 - Human base head
hm_ulm_basa_0 - Human eyelashes
Qunari male:
qm_har_blda_0 - Qunari scalp
qm_uem_basa_0 - Qunari eyes
qm_uhm_basa_0 - Qunari base head
With that done, close DATool and open up Max/GMax. Following the same procedure as last time, generate your donor head XML files, only this time using hm_uhm_basa_0. Once you have compiled your new head model files, we can begin the scaling process. Much the same as in Part 3, we'll start off with a World Scale change to up-size our model, only this time we'll follow it with a non-uniform scale and move chaser. But more of that in a bit. Reset the scene, then open the Listener and paste/type in:
DAOTools.WorldScale = 1.12
and hit Enter. With that done, import your newly compiled head, confirming first that the import windows states WorldScale from Setup: 1.12. Once that has finished importing, select the GOB and tick the Merge with selected checkbox, then merge in hm_har_blda_0 and hm_ulm_basa_0. Once those are finished, close the importer. Now open up the Listener again and reset the World Scale back to 1. Now we need to export. This is only a temporary file, so call it whatever you like. I named mine DUNCAN_SCALED_HEAD_SCALP_LASHES_0. With that done, reset the scene.
Now import the Qunari base head, qm_uhm_basa_0, making sure first that World Scale is at 1. Once that is finished, merge in the temporary file you just exported and close the importer window. You should now have something like this:
Now we need to do some tweaking. Select your custom head mesh. This should be called HM_UHM_BAS_FaceM1. In the modifier stack, select the editable poly object at the bottom, then from the drop-down list add an XForm modifier (in GMax, this is under, appropriately enough, the Parametric Modifiers section, in Max it's at the bottom of the list). Select the XForm in the stack and expand it, making sure the Gizmo is selected. You should see a box appear around your head. Now select the Non-Uniform Scale tool. You can get this by holding down on the Uniform Scale Tool button until the flyout pops up. The Non-Uniform one is in the middle. With that activated, go down to the co-ords box at the bottom of the screen. Next to that on the left is a little square button. If you mouse-over it you'll get a tooltip labelled Absolute Mode Transform Type-In. Activate this button. Now you can enter values in the co-ord boxes. For the X value, enter 107.6 and hit enter/click in another window. You'll see your head become a bit wider. Leave the Y value as-is. For the Z value, enter 95.1. Your head will become a bit smaller and drop down in height:
Now change to the Move tool. In the Z co-ord box, enter a value of 0.084. Your head should pop back up in line with the Qunari head. With that done you can deactivate the Gizmo in the modifier stack (just click on it). Now you can select the scalp and eyelashes meshes in turn and repeat the process: add XForm modifier, scale, move. Once that is finished and you are happy that everything is scaled and positioned as it should be, you can select the Qunari head mesh and delete it. That should leave you with something like this:
You may be wondering why we went about things in this manner. The reason is because the initial World Scale transform was primarily done to get the Human rig the same size as the Qunari rig, thus allowing the two to be merged together. If you try merging two rigs of different (especially radically different) sizes, bad things will happen. So once that was done, we then needed to apply some additional transforms to get the head in the right dimensions and position. The proportions of the Qunari head require the Human head to be squished top-to-bottom and stretched side-to-side.
So now we have out head sized up correctly, it's time to export it. Before we do though, we need to rename the material applied to the FaceM1 mesh so we can specify some custom textures. I named mine uh_hed_tallduncan. Also, in keeping with practices established in previous posts, change the mesh names from HM to QM. Once you've done that, you can export. Name it something appropriate, remembering to adhere to the DA model filename conventions. I named mine qm_uhm_tallduncana_0. Once the export is finished, reset the scene.
Now we have our base head, we need some hair and eyes. You may be also be wondering about the beard and why we didn't include that as part of the head model. The reason is the same one as why we included the scalp mesh in with the head model instead of the hair model, as the vanilla models do - tints. Having the beard as part of the face model causes problems trying to get it to accept tints, so we'll be including it with the hair model. Let's do that now.
Set your World Scale to 1.12 again. Import hm_har_duca_0, them merge in hm_brd_duca_0. Once that is finished, set your World Scale back to 1. Select the scalp mesh, which should be labelled with the epithet HairM1, and delete it. Now change the hair mesh name from HM to QM. Very importantly, change the beard name from HM_BRD_DUC_BeardM1 to QM_HAR_DUC_HairM3. I think the most critical thing here is the HairM3 epithet (but possibly the _har_ as well). Changing the name will allow it to tint properly. We are now going to do the same temp export that we did for the head model in order to swap to the Qunari rig. Strictly speaking, this is probably not necessary and the scaled Human rig would suffice, but I figure for the LOD0 at least, better to take this extra step to make sure everything aligns correctly. We won't worry for the LOD2/3 models. Export the model with a temporary name, then reset. Import qm_har_blda_0, making sure World Scale is 1. Merge in the temp hair/beard model. Now repeat the scaling process from before for both the hair and beard meshes: add XForm modifier, scale, move. Once that is finished, delete the scalp mesh and then export using an appropriate name. I called mine qm_har_tallduncduca_0.
Now repeat this procedure for the eyes. Set World Scale to 1.12, import hm_uem_basa_0, reset World Scale to 1, export temp file, reset scene. Import qm_uem_basa_0, merge in the temp scaled eyes, scale/move, delete Qunari eyes, change custom eyes mesh name from HM to QM. Export with appropriate name. I called mine qm_uem_talldunca_0.
With that done, we have all our LOD0 models. I won't cover the LOD2/3 creation process here. It is the same as was covered in Part 2 of the guide, but with the additional scaling/move we just did for the LOD0 models. As I said before, you shouldn't need to use the Qunari rig for the LOD2/3 models. Just do the initial World Scale 1.12 import, then do the subsequent manual scale and export at World Scale 1.
Now on to the textures.This follows the exact same procedure laid out for the Elf head, so i won't repeat the specifics. For the Duncan head though I chose a different channel in the stubble tint map for the eyebrows (Blue, for the bushy ones) and also added in the mutton chops from the Alpha channel. I also overlaid the neck area from the Qunari diffuse to try and get a better skin tone match with the body (didn't really seem to do much) and whitened the teeth. For the normal map, I tried the Qunari normal as Duncan seems to have a few wrinkles. Here's a brief summary pic (bit outdated - check source files for final version):
The last remaining task is to create the required GDAs. This follows the same pattern as previously, extending the appropriate race sheets (qunari_hair and qunari_heads) and adding in an additional line to the existing eyes sheet. Then it's on to testing. A lack of suitable Qunari in Origins makes Sten the test victim by default, so I duplicated his UTC and swapped out his morph for my new Duncan head:
Everything seems to be in order. Time to fire up the game for a quick test. Drum roll....
Moving right along... Here's a close-up of the neckline/body join:
And a more modest side-by-side, in daylight this time:
Here are a couple of convo shots:
As I mentioned the other day, it seems that heads don't get automatically removed by helmets like morphs do, which causes some intersection issues:
It seems like it does still remove the hair model at least. I guess that is an insurmountable issue with using head models over morphs. You may need to either not use helmets at all, or try and find one that obscures any intersection (i.e. winged helmets). For Elves particularly, try and avoid giant bat ears like I used on my morph.
Modifié par DarthParametric, 25 août 2011 - 01:47 .
Retour en haut
I guess I'll have to tone down the elf ears. They were a bit big on my elf 
