Mdl vs mdl. Tga vs Dds

Hello! Title could seems crazy but I have some questions about NWN files.
Ascii vs binary mdl. Pros and cons of each?

Dds. Pros and cons? I converted some large tga to dds. 95% of them work well, but some of them gave me problems, expecially tga used for lights or graphic effects. In addition, a very small percentage is now 180° rotaded.

Thanks :slight_smile:

There are multiple threads on these topics. I’ll attempt to summarise:

  • Compiled (binary) mdl files load faster than ASCII, but the difference is imperceptible except in the case of some very high poly models (of which there are more these days)

  • ASCII models have the advantage that they can simply be amended with text editors without the overhead of decompile / recompile or a complicated modelling tool

  • DDS files are more compact and generally work better than TGA

  • NWN Crunch is the tool of choice for making DDS and TGA for NWN these days, as it doesn’t seem to have those issues (though tools like TGA Flipper are still available if needed)

1 Like

Thank you very much for this very complete answer!
Yes, I am using NWN Crunch, but some problems exist. Some rare case of flipped textures and, more important, some textures do not work anymore…

There is a caveat with DDS files. They are a lossy format. That means they don’t hold as much information as TGA. In certain cases they can display incorrectly. I had some high resolution textures with very fine lines and in certain colour combinations these lines were smeared out to several times their intended size.

Also there are certain small graphics in NwN that must be in TGA format, not Direct Draw Surface (aka DDS).



One more word on MDL files…

ascii mdl files, if fancy mapped, allow the game engine to calculate and apply the tangents to the meshes, allowing the texture to display heightness, reflection, etc.

binary mdl files do not allow this, UNLESS the mdl file was compiled WITH the tangents in place using the in game compiler.

None of the binary vanilla mdl files will display a normal/specular/etc. map without first being decompiled to ascii format and loaded into the game.


my bed is a mess, awash with the sweat i wake with deep in the night, having lost untold amounts of sleep over these troubling questions… ‘fancy mapped’ ? ‘specular’ ? ‘mesh tangents’ ? all of these – and MORE – resound with gravitas… and there are more, and ever more posts and projects and contributions talking about this… and yet – WHAT DOES IT MEAN ??!
google is no help, preferring to send me to etsy and other merchandising holes… duckduckgo quacks in my face… qwant tells me i should go examine my bottom and charm and spin (up/down)…
can anyone please point me to something that explains what all of this is about ? [preferably with pictures… pretty ones… yeah…]

“Fancy mapping” is a community slang term for using layered textures through a material file (MTR) to provide a textured surface with depth to fake a 3D aspect.

Normal maps “fake” the lighting of bumps and dents.

Specular maps determine how that lighting reflects off those bumps and dents.

Heightness maps “fake” the depth of those bumps and dents by making raised (white) surfaces look higher than recessed (black) surfaces.

Tangents are simply points that tell the game engine where to stretch the texture to to get these fake 3D effects from a 2D texture.

If you use Discord, go to and on the screenshots page (starting about midway down) you’ll see some examples of fancy mapping at work. Sorry there aren’t any before/after shots, but by now, anyone that’s been around the community as long as we have probably has the 2D look of the vanilla textures imprinted on their brain.

1 Like

That’s quantum mechanics and as Pstmarie’s explanation shows, fancy maps are nowhere near that complicated. For a very quick example take a look at the Fake/Concealed Door Placeable EE I did a couple of years ago. That is just a flat model (aka a plane) but by the use of normal and specular maps you get the impression (in game) that it is 3D.

For an explanation of just normal and specular maps with before and after pictures there’s this page from the game Second Life’s community site.


1 Like

It sounds way more complicated than it is. Simply put, in ages past videogames would use a 3d model, and envelop it in a single picture file, which would add detail, color and texture. The models were, therefore, very simple, while the complexity fell upon texture image files, which became bigger and bigger, or fragmented into several parts, in order to add as much detail and resolution as possible. Such went on for a long time, until more modern games came around. Technology evolved along with developers, who realized they could significantly reduce the weight put on a single image file to envelop a model by using “image overlays” or different “map” files over said model. They could divide different aspects of the original texture that react to lighting. Normal maps were the first, which show bumps and actual texture relative to sources of light. Specular maps show how shiny and reflective stuff should be under light depending on the stuff the object is made of (Metal should be shinier and more reflective than stone, for example). Tint maps determine different colors for the same, base texture, which means, instead of having 4 or more image files for a model, each bearing a different color, you can use a single Tint Map that includes all desired colors. Imagine these new “maps” as overlays to the original texture image (which is usually called a Diffuse Map nowadays).

thanks for all the great information, everyone!
i think i see what you’re talking about, @Pstemarie – esp re- the ‘fancy shields’ and some of the floor textures in your re-works. so these wouldn’t look so textured w/o the fancy-mapping, then ?
@Tarot_Redhand is the sort of fancy mapping you used for the 3D door effect not possible in 1.69 ?

aha ! this is the method used in bioware’s nwn heads, right? [now if only they could do that in a way that wouldn’t make the female heads ‘grow beards’… O_O …or is that also a matter of some sort of texture mapping as well ? ]

so for example colour 61 on the nwn colour palette has a very high ‘specularity’, i guess. [61 is the one that looks like flowing mercury]

as for the rest, it’ll take me some time to mull it over. thanks for your answers and giving me something interesting to chew on. :slight_smile:

No. The routines to make it happen just don’t exist in 1.69 only EE (and to some extent NwN 2).


Oh no no. See, most of what I shared was about more modern games. Original NWN is part of the old guard. I believe “maps” were first used around 2006, back when Oblivion and NWN2 came out, so our classic NWN didn’t see these advancements until the Enhanced Edition. The heads do go the “single texture enveloping a model” approach. However, Bioware used some tricks from the era to get around the limitations. If you’ve noticed, some models in the game use a different image file to bmp, dds or tga. Heads in particular use something called “plt” files, which are an early form of tint maps, which means they hold the possibility of multiple colors in a single image file. There is a little utility in the vault called PLTEditor that can help you open those files so you can have an idea of how they work. As for specularity, Bioware used another trick applied to tga files based on their alpha channels. Usually, alpha channel in a tga image determines transparency, but NWN has an option to turn it into reflectiveness (something called “environmental map”) through its 2da files. This means that something like an armor, having some level of transparency applied to its tga texture, would have the game interpret it as “shininess,” as something made out of metal would look. This you can see by opening said texture in an image editor like photoshop.