View source for VTEX (Valve Texture)

{{Tabs|VTF (Valve Texture Format)|source=1|custom1=VTEX (Valve Texture)|custom1alt=Source 2|main=source}}
{{for|the Source 1 command-line program|[[Valve Texture Tool]]}}
VTEX is the proprietary texture format use by the [[Source 2]] engine. The VTEX format includes a .vtex source file that describes how to compile the texture, and the actual compiled .vtex_c texture file.

For most uses the .vtex_c file is compiled automatically by the tools, and no .vtex source file is generated. In some cases, such as when creating sprites for particle effects or textures for [[env_cubemap_fog]] and [[env_gradient_fog]], no automated tool exists and the .vtex file has to be created manually.

{{todo|Mention [https://github.com/kristiker/source2utils/releases/ dds2vtex] somewhere.}}
{{bug|While certain shaders support non-power-of-two textures, {{cs2|2|addtext='s}} Workshop uploader doesn't properly handle such textures.}}
= Source file format =
The .vtex file is a [[DMX]] file usually created in the keyvalues2_noids text format. It describes how to compile a texture from a set of image inputs. Placing a .vtex file in the content directory of the current add-on will cause the tools to compile it when referenced.

== Template file ==
<syntaxhighlight lang="js">
<!-- dmx encoding keyvalues2_noids 1 format vtex 1 -->
"CDmeVtex"
{
	"m_inputTextureArray" "element_array" 
	[
		"CDmeInputTexture"
		{
			"m_name" "string" "InputTexture0"
			"m_fileName" "string" "materials/my_textures/my_texture.png"
			"m_colorSpace" "string" "srgb"
			"m_typeString" "string" "2D"
			"m_imageProcessorArray" "element_array" 
			[
				"CDmeImageProcessor"
				{
					"m_algorithm" "string" "None"
					"m_stringArg" "string" ""
					"m_vFloat4Arg" "vector4" "0 0 0 0"
				}
			]
		}
	]
	"m_outputTypeString" "string" "2D"
	"m_outputFormat" "string" "DXT1"
	"m_outputClearColor" "vector4" "0 0 0 0"
	"m_nOutputMinDimension" "int" "0"
	"m_nOutputMaxDimension" "int" "0"
	"m_textureOutputChannelArray" "element_array" 
	[
		"CDmeTextureOutputChannel"
		{
			"m_inputTextureArray" "string_array" [ "InputTexture0" ]
			"m_srcChannels" "string" "rgba"
			"m_dstChannels" "string" "rgba"
			"m_mipAlgorithm" "CDmeImageProcessor"
			{
				"m_algorithm" "string" "Box"
				"m_stringArg" "string" ""
				"m_vFloat4Arg" "vector4" "0 0 0 0"
			}
			"m_outputColorSpace" "string" "srgb"
		}
	]
	"m_vClamp" "vector3" "0 0 0"
	"m_bNoLod" "bool" "0"
}

</syntaxhighlight>

== File data format ==
{{note | The lists may be incomplete.}}

The .vtex file is structures as a CDmeVtex list with arrays for CDmeInputTexture input textures and CDmeTextureOutputChannel output channel sets. 

=== CDmeVtex root keys ===
{| class="standard-table"
! style="width: 200px;" | Attribute
! style="width: 60px;" |Type
! Description
|-
| m_inputTextureArray
| element_array
| Array of CDmeInputTexture entries describing each input image.
|-
| m_textureOutputChannelArray
| element_array
| Array of CDmeTextureOutputChannel entries describing sets of output channels.
|-
| m_outputTypeString
| string
| Type of texture to generate.
|-
| m_outputFormat
| string
| Image data format to generate.
|-
| m_outputClearColor
| vector4
| Initial color before input textures are applied to the output.
|-
| m_nOutputMinDimension
| int
| The minimum size of a side of the output texture in pixels .
|-
| m_nOutputMaxDimension
| int
| The maximum size of a side of the output texture in pixels.
|-
| m_nOutputDimensionReduce
| int
| {{todo}}
|-
| m_vClamp
| vector3
| Whether to clamp the S, T and U coordinates respectively.
|-
| m_bNoLod
| bool
| Disables Level of Detail.
|-
| m_bHiddenAssetFlag
| bool
| Marks the output file as a hidden asset.
|-
| m_bCreateLinearColorSpaceApiTexture
| bool
| {{todo}}
|-
| m_nDisplayRectWidth
| int
| {{todo}}
|-
| m_nDisplayRectHeight
| int
| {{todo}}
|-
| m_nMotionVectorsMaxDistanceInPixels
| int
| {{todo}}
|}

=== CDmeInputTexture keys ===
{| class="standard-table"
! style="width: 200px;" | Attribute
! style="width: 60px;" |Type
! Description
|-
| m_name
| string
| Name to refer to this input by.
|-
| m_fileName
| string
| Filename for the input image.
|-
| m_colorSpace
| string
| Image color space: linear or srgb
|-
| m_fileExt
| string
| {{todo}}
|-
| m_nMinBitsPerChannel
| int
| {{todo}}
|-
| m_typeString
| string
| Type of input texture.
|-
| m_imageProcessorArray
| element_array
| Array of image processor algorithms to apply to the input.
|-
| m_bPassThroughToCompiledVtex
| bool
| Pass through the input image directly to the texture file?
|-
| m_n3DSliceCount
| int
| {{todo}}
|-
| m_n3DSliceWidth
| int
| {{todo}}
|-
| m_n3DSliceHeight
| int
| {{todo}}
|}

=== CDmeTextureOutputChannel keys ===
{| class="standard-table"
! style="width: 200px;" | Attribute
! style="width: 60px;" |Type
! Description
|-
| m_srcChannels
| string
| Specifies the used channels of the input texture and their order: r, g, b and a
|-
| m_dstChannels
| string
| Specifies what channels of the output textures map to the specified input channels: r, g, b and a
|-
| m_mipAlgorithm
| CDmeImageProcessor
| Image processor algorithm to use for mipmap generation.
|-
| m_outputColorSpace
| string
| Output color space: linear or srgb
|-
| m_inputTextureArray
| string_array
| Array of input textures by their m_name keys.
|}


=== CDmeImageProcessor keys ===
{| class="standard-table"
! style="width: 200px;" | Attribute
! style="width: 60px;" |Type
! Description
|-
| m_algorithm
| string
| What algorithm to use for this image processor.
|-
| m_stringArg
| string
| Algorithm specific argument.
|-
| m_vFloat4Arg
| vector4
| Algorithm specific argument.
|}


=== Texture types ===
{| class="standard-table"
! style="width: 200px;" | Format
! Description
|-
| 1D
| One dimensional texture.
|-
| 2D
| Two dimensional texture.
|-
| 3D
| Three dimensional volumetric texture.
|-
| 1DARRAY
| Array of one dimensional textures.
|-
| 2DARRAY
| Array of two dimensional textures.
|-
| 3DARRAY
| Array of three dimensional volumetric textures.
|-
| CUBE
| Single cubemap texture with a cross layout.
|-
| CUBEARRAY
| Array of cubemap sides.
|}


=== Image data formats ===
{| class="standard-table"
! style="" | Format
! Description
|-
| DXT1
| Standard RGB block compression (also known as BC1)
|-
| DXT5
| Standard RGBA block compression (also known as BC3)
|-
| ATI1N 
| Single-channel (greyscale) block compression (also known as BC4)
|-
| ATI2N 
| Two-channel block compression (also known as BC5)
|-
| BC6H 
| HDR block compression
|-
| BC7 
| High quality RGB or RGBA block compression {{dota2|not}}
|-
| I8 
| Uncompressed greyscale
|-
| IA88 
| Uncompressed greyscale with alpha
|-
| RGBA8888 
| 
|-
| BGRA8888  
| 
|-
| R16 
| 
|-
| RG1616 
| 
|-
| RGBA16161616 
| 
|-
| R16F 
| 
|-
| RG1616F 
| 
|-
| RGBA16161616F 
| 
|-
| R32F 
| 
|-
| RG3232F 
| 
|-
| RGB323232F 
| 
|-
| RGBA32323232F 
| 
|-
| R11_EAC 
| 
|-
| RG11_EAC 
| 
|}

=== Image processor algorithms ===
These algorithms can be added to an input m_imageProcessorArray.
{| class="standard-table"
! style="width: 200px;" | Algorithm
! Description
|-
| None
| Do nothing.
|-
| Inverse
| Invert the image channels.
|-
| NormalizeNormals
| Normalize the vectors of a normal map to have a length of 1.
|-
| FillToPowerOfTwo
| Fill the texture to the nearest power of two size.
|-
| FillToMultipleOfFour
| Fill the texture to the nearest multiple of two size.
|-
| EncodeRGBM
| Encode a HDR texture to RGBM: RGB with a multiplier in the alpha channel.
|-
| Mod2XCenter
| {{todo}}
|-
| HeightToNormal
| Generate a normal map from a height map.
|-
| DilateColorInTransparentPixels
| {{todo}}
|-
| ConvertToYCoCg
| Convert the channels to the YCoCg color model.
|}


=== Mipmap processor algorithms ===
One of these algorithms can be selected as the mip processing algorithm for the output.
{| class="standard-table"
! style="width: 200px;" | Algorithm
! Description
|-
| None
| Output without mipmaps.
|-
| Box
| Standard box filtered mipmaps.
|-
| Nice
| Use Source 1 [[NICE_filtering|NICE filtered]] mipmaps.
|-
| BoxInverse
| Standard box filtered mipmaps with inverted colors.
|-
| MultiplyNoMip
| Multiply two input textures together.
|-
| MultiplyBox
| Multiply two input textures together, with box filtering mipmaps.
|-
| AlphaWeighted
| {{todo|Used for transparency masks and color in the same texture.}}
|-
| Normals
| {{todo|Regular tangent space normal maps?}}
|-
| AnisoNormal
| {{todo}}
|-
| AnisoNormalRoughness
| {{todo}}
|-
| AnisoRoughness_RG
| {{todo}}
|-
| HemiOctAnisoRoughness
| Used with hemi-octahedron normal maps and anisotropic roughness maps packed into the RB and GA channels respectively.
|-
| HemiOctIsoRoughness_RG_B
| Used with hemi-octahedron normal maps and regular isotropic roughness maps packed into the RB and G channels respectively.
|-
| HemiOctAnisoRoughness_RG_BA
| {{todo}}
|-
| GGXCubeMapBlur
| Blur cubemap mips to use for variable roughness reflections with the GGX BRDF.
|-
| GGXCubeMapBlur_RGBM
| Blur cubemap mips to use for variable roughness reflections with the GGX BRDF. For RGBM encoded textures.
|-
| RGBM
| Generate mips for RGBM encoded textures.
|-
| Mip2DGradient
| {{todo}}
|-
| PreserveCoverage
| {{todo}}
|-
| MipDepth
| {{todo}}
|-
| MipHeight
| {{todo}}
|-
| MipWidth
| {{todo}}
|-
| AutoLevels
| {{todo}}
|-
| WrapGaussian
| {{todo}}
|-
| HeightCombine
| {{todo}}
|}


= Compiled file format =
{{todo}}

The compiled .vtex_c format can be extracted with the third party tool {{vrf|4}}.

= See also =
[[SteamVR/Environments/Fog_Tutorial|SteamVR Home env_gradient_fog tutorial]]

[[Category:Source 2]]
[[Category:Material System]]
[[Category:File formats]]