File Formats

Gamestudio and lite-C support numerous file formats for importing and exporting artwork and data.  !!  Keep file names short - the names of entity and texture files (including extension) in WMP and WMB levels are limited to 30 characters, and texture and skin names in WAD and MDL models files are limited to 15 characters. File names must not contain any special characters except "_" or "+". Never, ever, use spaces in your file or path names! Many import filters from other editors will not find texture files when they have spaces or exotic characters in their name.

In the following you'll find specific information about the formats supported:

Image files

• PCX 256 color (8-bit palettized) images for low quality, reduced color textures are supported for compatibility reasons. All 256 color images must share the same color palette per level, given in WED Map Properties. The first color of the palette (color #0) must always be black, the last color (color #255) must always be white. Due to the many restrictions and requirements, we do not recommend to use 256 color images - especially for beginners who don't know about handling color palettes.
• PCX true color (24-bit RGB) images are used for low quality textures. They are internally stored in either 16-bit high color or 4-bit compressed format, depending on d3d_texdepth.
• TGA true color (24-bit RGB) and true color alpha (32-bit ARGB) images are used for high quality textures. They are internally stored in 16-bit high color or 32-bit true color, depending on the setting of d3d_alphadepth. 32-bit ARGB images contain an alpha channel which gives a transparency value for every single pixel (see also: Transparency) . For creating an alpha channel in Photoshop, open the channels window, and click onto the small "new channel" button below. Now an alpha channel is created, and you can paint onto it. When saving the picture, use the 32 bit TGA format.
• DDS images are used for medium quality textures when video memory must be economized. They are supported in the DXT1..DXT5 sub-formats with included mipmaps. DDS uses a compressed 4-bit texture format for saving memory at the expense of image quality. DXT2..DXT5 textures contain an alpha channel and are treated as transparent by the engine (see also: Transparency). DXT1 has a 1-bit alpha channel that could be used for overlays, and is not treated as transparent. Due to their compressed content, DDS bitmaps can not be altered by image manipulation functions (like pixel_to_bmap), can not be used for textures that contain several separate images (like animated sprites or sky cubes), and are not displayed by the older GXL libraries in the WED and MED editors. They can be created with the DirectX Texture Tool, with a Photoshop Plugin, or with ATI's Compressonator.
• BMP true color (24-bit RGB) images behave like PCX images, but are uncompressed and thus have larger file sizes than PCX files. There is also a seldom used BMP 32 bit format, which can be created f.i. by taking screenshots, but is not supported by the engine and by many paint programs.
• PNG images behave like 32 bit TGA images with alpha channel (even if the channel was not explicitly created in the pain program). They are not recommended because they are loaded not by the engine, but by the DirectX library, and are not guaranteed to be automatically restored when the video device gets lost. This means that they can lose their content f.i. when a fullscreen application is minimized.
• JPG images should not be used for textures in realtime games. They combine all disadvantages of the other formats, plus the additional disadvantage of slow loading, high video memory consumption, and bad texture quality. Their only advantage is small file size.
• WAD files are image collections containing several BMP, PCX, TGA, or DDS images, with names and included mipmaps. They are imported in WED and used for level textures. As DDS images are not displayed by WED, they appear as blue squares in the WAD preview.

Some considerations about transparency:

• All image formats that contain an alpha channel (32 bit TGA, or DXT2..DXT5) are treated as transparent by the engine - even if the alpha channel has no meaningful content and even if you don't know what an alpha channel is! Transparency has consequences for sorting and rendering, so don't use an alpha channel texture for an object that you don't want to be transparent.
• Images without alpha channel can be used for overlays with transparent parts. Overlays are sorted like opaque textures. In overlay images, black parts (RGB == 000) are totally transparent, the rest is totally opaque. For nontransparent black use a RGB color value of 8 or above. If d3d_autotransparency is enabled, the transparency color is taken from the upper left pixel of the image. This way an arbitrary color can be used for transparency, avoiding the dark seams around intransparent parts that are otherwise caused by black pixel antialiasing.

Restrictions apply to image sizes:

• 3D cards have a texture size limit - mostly 4096, 2048, or 1024 pixels in every direction. For supporting all 3D cards, don't use bigger bitmaps than 1024x1024, except for images containing several frames (for animation or sky cubes), or images with mipmaps. The next-lower mipmap is then automatically used when a 3D card does not support the image size.
• 3D cards can not render arbitrary texture sizes. Normally they require power-of-2 texture sizes (32, 64, 128, 256, 512, 1024...), like 64x128 or 128x512. The minimum texture size normally is 8 pixels. All invalid sizes are automatically extended by the engine to the next power of 2, at cost of video memory, performance, and image quality. Invalid texture sizes on models are only possible when all skins of a model have the same size. In some cases, f.i. for tiled textures, invalid sizes are not possible at all. Therefore, always use power-of-2 sizes for all level, terrain, and model textures. Only for sprites and panels it might make sense to use different sized textures.

Video files

Sound files

• OGG is the recommended format for snd_play and sound sources. It uses Gamestudio's internal OGG player. OGG can also be used with media_play, however an media player OGG filter (installer included in Gamestudio) must then be distributed with the game and installed on the target machines.
• WAV and MID work everywhere, but WAV is large in size and MID low in quality.
• MP3 requires a license fee and thus is not recommended for low budget games.

Script and Effect files

Model files

• 3DS, ASE, and OBJ files contain nonanimated models with multiple meshes and materials, and are supported by almost all 3D packages. Some 3D editors do not use transparency and set the alpha value at zero, so after importing check the material alpha setting. Using a list of 3DS files, vertex animation can be imported. 3DS models can also be exported for using them in external editors. When textures are stored externally, make sure that their format is supported by the editor - such as BMP, TGA, or DDS (see above).
• X is the Microsoft model format. It contains multiple meshes and materials plus bones or vertex animation. Unfortunately, the X format evolved over time into many different, incompatible variants that can normally only be read with the same editor that exported them. Gamestudio can import all X files that are compatible to the Microsoft X format. You can find out compatibility issues with the Microsoft Mesh Viewer that comes with the DirectX SDK. If your model looks and animates correctly in the Mesh Viewer, it will also be correctly imported in Gamestudio.
  X bones animation can be inconsistent when different versions of coordinate transformation systems are used (even within the same file). To be on the safe side, use only bones rotation - no movement and scaling - when you want to export the model to Gamestudio or to other tools.
• FBX is the standard platform independent model exchange format. Exporters are available for by all major editors, such as MAX® or MAYA®. Unfortunately, the FBX format is not compatible to itself; there's a difference between FBX models saved with the 2005 or earlier FBX version, and saved with newer versions (2006 or above). If you are not sure which FBX version is exported, try both. For not animated models, always leave Convert to Points Animation checked in the import dialog. For importing bones, the same restrictions apply as mentioned for X models.
  The following issues were reported with FBX exporters from other editors: MAX® only exports bones that are assigned to at least one vertex. Cinema 4D® properly exports to FBX only when the textures are stored in .TGA format. Milkshape® FBX export was reported to export wrong bones animation. Some editors use a deformation object (such as the Bulge object in Cinema 4D®); make sure that the model is exported without attached deformation objects.
• Milkshape is a popular model editor that supports many formats through external filters. Milkshape filters can be used by MED, and models in Milkshape ASCII format can be directly imported, including bones animation.
• MDL and MD2 files are vertex animated models with a single texture, used for the Quake game series by id software.
• ASC is a nontextured mesh format for importing raw 3D data.

Gamestudio files

• WMP files are created by WED and contain the level lighting, textures and geometry, plus references to additional files for models, terrain, sprites, and sound source. Once they are compiled they become WMB files that can be read by the engine.
• WAD files are created by WED and contain texture collections. External WAD files from Quake and Halflife can be imported.
• MDL files are created by MED and used for animated models. They support bones and vertex animation, multiple meshes and textures, materials and shaders. They are not identical with Quake MDL files, although the engine also accepts models in the Quake MDL format.
• HMP files are created by MED and used for terrain. They contain the height values and the terrain textures.