Are you sure that the material order in Blender 2.79 actually mattered (and that it wasn't instead decided by the internal polygon order, where adding new polygons with new materials would of course make them come after previously-existing polygons, with Blender 2.80 maybe importing those polygons all willy-nilly or just backwards or such, messing up the order)?

I remember a similar "material order" issue for Source (not GoldSrc), from long before Blender 2.80 was released.


Does GoldSrc support $TextureGroup? If so, you might be able to specify the materials in order before the meshes are loaded, like this:For Source, the above QC would put the "Opaque" material first in the model, then "Transparent 1", then "Transparent 2", regardless of the order that they're defined in "Material_Tech.smd", allowing you to easily layer materials in any render order.
But I'll admit, I have no idea about whether or not this works for GoldSrc.

引用自 Ro예：

I mean, not such an attachment, Such like object that attaching on guns.

Have you added $BoneMerge into the QC yet, for every bone?

引用自 Ro예：

- i need to make bone that don't have vertex group for attachments.

If you need to use $Attachment, add it into the QC, and compile the model again.

If bones don't have a purpose, they won't be in the model.
But $Attachment is a purpose, making the bones come back.

引用自 bubstock9：

- What's weird is that it's clearly a problem with the .vmt -

It's not; It's instead the iris VTF.

Using something like VTFEdit, enable the VTF's "Clamp S", "Clamp T", and/or "Clamp All" flags in the "Image" tab, then save the texture and reboot Source Filmmaker.

Use the "Blender 2.74+" > "2.10.2" link of http://steamreview.org/BlenderSourceTools/archives/ to get the latest version of the Blender Source Tools for Blender 2.79.


If you don't trust the above link, switch to the "Announcements" tab of this group, click on the "Download" link of an announcement to go to http://steamreview.org/BlenderSourceTools/ , and then click "Links" > "Official" > "Release archives" from there.

引用自 ḾᶉḾĻǤҎŀãẏ：

-Image, showing a character opening their mouth, with one of the vertices of the lower lip extending up to the upper lip-

Move those vertices a bit farther apart on the base mesh.

Flexes (both DMX flexes and VTA flexes) are matched based on how close a vertex on the base flex is to a vertex on the model. If two vertices on the base model are in the exact same location, then only one of those vertices will ever be "matched", since neither of them are closer to something than the other and only the "closest" vertex is matched to a flex.

引用自 ḾᶉḾĻǤҎŀãẏ：

- it turns out that in this screenshot: -Link- and there you can see that his teeth are black, -

No, we can't; That Google Drive item is private. After clicking on that link, I simply see "You need permission - Want in? Ask for access, or switch to an account with permission. [...] ["Request access" button] ["Switch accounts" button]".

Try to right-click that item in your Google Drive, then choose "Get shareable link" to make it non-private and also automatically be given the proper shareable URL, though your link probably already leads to the correct URL.

What version of the Blender Source Tools are you using, and can you link to the exact SMD file in question?

That you haven't updated your installation of the Blender Source Tools to version 3.0.3 yet.

Can you verify that the mesh SMDs still have less than 100 bones when you make a new Blender session and import one of the mesh SMDs into it?

I know that, at least for Source, the model compiler itself can discard bones that don't have any vertices connected to them if none of their child bones are linked to any vertices either, even though those bones may still be in the SMDs themselves.
Meanwhile, for mesh-less animations, the model compiler can of course only assume that every single bone means something important.


Additionally, what versions of Blender and the Blender Source Tools are you using?


I'm sure that Artfunkel would appreciate knowing those things.