The <texture> node

Function

General syntax

<texture
	name=""
	filename=""
	pixel_format="RGB"
	filtering_mode="LINEAR"
	s3tc="FALSE"
	num_mipmaps="13"
	anisotropy="1.0"
	addressing_mode="WRAP"
	function_type="MODULATE"
	play_avi="TRUE"
	create_opacity_map="FALSE"
	create_binary_alpha_map="FALSE"
	create_constant_alpha_map="FALSE"
	constant_alpha_map_value="0.5"
	gen_coord_mode="NONE"
	type="TEXTURE_2D"
	use_fbo_for_rtt="FALSE"
	fbo_attachment_point="FBO_COLOR_ATTACHMENT0"
	check_crc32="0"
	vertical_flip="FALSE"
	horizontal_flip="FALSE"
	use_texture_matrix="FALSE"
	perlin_noise="FALSE"
	checker_pattern="FALSE" >

	
	<cube_map
		face="XPOS"
		filename=""
		uv_map_op="NO_SWAP" />

	<perlin_noise_params
		min_color_r="0.0"
		min_color_g="0.0"
		min_color_b="0.0"
		min_color_a="1.0"
		max_color_r="1.0"
		max_color_g="0.0"
		max_color_b="0.0"
		max_color_a="1.0"
		amplitude="200"
		frequency="15" />

	<size
		width="256"
		height="256" />

	<spin_value
		z="0.0" />
		
	<binary_alpha_map_color
		r="0.0"
		g="0.0"
		b="0.0"
		epsilon="0.0" />
		
	<constant_color
		r="1.0"
		g="1.0"
		b="1.0"
		a="1.0" />
		
	<filling_color 
		r="0.0"
		g="0.0"
		b="0.0"
		a="1.0" />
		
	<inject_opacity_map
		filename="" />
	
		
</texture>

texture element

• name - [STR127] - name of the node. This name will make it possible to refer this node throughout the XML script.
  
  
• filename - [STR255] - name of the texture file to be loaded. The access path is relating to the xml file.
  
  
• pixel_format - [ENUM] - pixel format (or texel):
  • RGB: each pixel of the texture is coded on 3 bytes: 1 byte (1 byte) for each R, G and B component. The value of each component is clamped: [0 - 255] - default value
  • BGR: each pixel of the texture is coded on 3 bytes: 1 byte (1 byte) for each B, G and R component. This format should be used to load AVI files. The value of each component is clamped: [0 - 255] - default value
  • RGBA: each pixel of the texture is coded on 16 bytes: 4 bytes (1 float) for each component R, G, B and A. This format of pixel is to be used as soon as one must act on the texture alpha channel (blending, alpha-test...). The value of each component is clamped: [0.0 - 1.0]
  • RGBA_32F: defines a floating point texture. Each pixel of the texture is coded on 128 bits: 32 bits (1 float) for each component R, G, B and A. The value of each component is non clamped: each component is a true 32 bits IEEE 754 floating point number. Besides, you have to use a pixel shader when rendering using a floating point texture as input. This applies regardless of whether the output buffer is the usual frame buffer or FBO. On some hardware implementation (nVidia) you have to use NPOTD texture (TEXTURE_RECTANGLE_2D) to set RGBA_32F format.
  
  
  
• filtering_mode - [ENUM] - filtering mode of the texture:
  • NONE: no filtering is applied to the texture. It is the fastest mode. Useful for distant textured objects
  • LINEAR: default filtering
  • TRILINEAR: filtering mainly used when dealing with large textured surfaces featuring mipmapping
  The purpose of filtering is to limit the bad visual effects when texture is stretched or shrinked.
  
  
• anisotropy - [REAL] - value of the anisotropic filtering factor. A value of 1.0 corresponds to isotropic filtering (default filtering ) - default value: 1.0
  Remark: The max value of the anisotropic filtering coefficient depends on the graphics controller. The following values are common: 4.0, 8.0 or 16.0. The HardwareInfos tool, delivered with Hyperion, gives this value.
  
  
• addressing_mode - [ENUM] - specifies the behavior of the texture on its edge and how the texture is repeated when the tiling factor (uv_tiling) applied to the vertices is greater than 1.0:
  • WRAP: texture is repeated in the 2 direction U and V - default value
  • MIRROR: texture is repeated in the 2 direction U and V with mirror effect
  • CLAMP: texture is not repeated any more and the edges of texture are not taken into account
  
  
  
• function_type - [ENUM] - texture function. This texture function specifies how the texture is mixed up with the entry color in the 3d pipeline of the graphics controller. The entry color can be either the vertices color, or the previous texture in the case of multitexturing.
  texture_function describes the operation carried out on the 2 entries of a texture unit:
  
  
  
  The following values are accepted:
  • MODULATE: texture is multiplied by the entry color: Output = A0 * A1 - default value
  • REPLACE: texture is not influenced by the entry color: Output = A0
  • ADD: texture is added with the entry color: Output = A0 + A1
  • ADD_SIGNED: texture is half-added with the entry color: Output = A0 + A1 - 0.5
  • SUB: texture is substracted from the entry color: Output = A0 - A1
  • INTERPOLATE_CONSTANT: the final color is the result of the interpolation between the texture and the color resulting from the previous texture unit: Output = (A0 * A2) + (A1 * (1.0 - A2)). Here A2 is a constant color which must be specified by <constant_color>
  • INTERPOLATE_TEXTURE_ALPHA: the final color is the result of the interpolation between the current texture and the texture that is on the previous texture unit: Output = (A0 * A2) + (A1 * (1.0 - A2)). Here A2 is a variable color which must be specified by <inject_opacity_map>
  
  
  
• gen_coord_mode - [ENUM] - specifies the way in which the texture coordinates (U and v) are generated for each vertex of a 3d object:
  • NONE: U and v coordinates are left just as it is with the initialization of the 3d object. default value
  • SPHERE_ENV_MAP: U and v coordinates are automatically generated by the 3d engine to create a spherical mapping effect. This kind of mapping allows to simulate objects that seem to reflect the environment
  • CUBE_ENV_MAP: U and v coordinates are automatically generated by the 3d engine to create a cubic mapping effect. This kind of mapping allows to simulate objects reflecting the environment. This type of mapping will be much more realistic than the SPHERE_ENV_MAP mode, but is also greedier in term of GPU resources (presence and update of cubic-map)
  
  
  
• type - [ENUM] - type of texture:
  • TEXTURE_1D: texture 1D with normalized addressing (texture coordinates ranging from 0.0 to 1.0).
  • TEXTURE_2D: texture 2D with normalized addressing (texture coordinates ranging from 0.0 to 1.0) - Default value.
  • TEXTURE_RECTANGLE_2D: texture 2D (NPOTD) with non-normalized addressing (texture coordinates ranging from 0.0 to image max_size) Caution: this Attention, this kind of texture is not supported by all graphics cards models. use the HardwareInfos tool to test your graphics card compatibility.
  • TEXTURE_CUBE_MAP. cubemap type texture. This type of selection automatically initializes gen_coord_mode with CUBE_ENV_MAP.
  
  
  
• use_fbo_for_rtt - [BOOLEAN] - enables (TRUE) or disables (FALSE) the use of this texture with the default FBO (Frame Buffer Object). Use this boolean when you plan to use this texture as reflexion texture - default value: FALSE
  
  
• fbo_attachment_point - [ENUM] - specify the attachment point in the FBO.
  • FBO_COLOR_ATTACHMENT0: color buffer 0.
  • FBO_COLOR_ATTACHMENT1: color buffer 1.
  • FBO_COLOR_ATTACHMENT2: color buffer 2.
  • FBO_COLOR_ATTACHMENT3: color buffer 3.
  
  
  
• perlin_noise - [BOOLEAN] - enables (TRUE) or disables (FALSE) the use of the Perlin noise to fill texture - default value: FALSE
  
  
• check_crc32 - [INTEGER] - allows to perform a crc32 type integrity test on the texture. This kind of test is useful to control if the right file has been loaded. That makes it therefore impossible to falsify a Hyperion scene delivered in a binary format : even if the description code of the scene is in binary format, textures and other 3d objects are always accessible in the ususal files type. Anybody could then appropriate a project by falsifying some textures or models. The crc32 code of a file can be obtained using the Hyperion_CRC32.exe utility - default value: 0
  
  
• vertical_flip - [BOOLEAN] - if set to TRUE, allows to vertically turn over the texture - default value: FALSE
  
  
• horizontal_flip - [BOOLEAN] - if set to TRUE, allows to hozizontally turn over the texture - default value: FALSE
  
  
• use_texture_matrix - [BOOLEAN] - enables (TRUE) or disables (FALSE) the use of the texture transformation matrix. The texture matrix hardware allows to perform, with a very small CPU usage, complex textures transformations. The value of the texture rotation is defined by the pin_value attribute - default value: FALSE
  
  
• checker_pattern - [BOOLEAN] - enables (TRUE) or disables (FALSE) the use of a chequered pattern to fill the texture - default value: FALSE
  
  
• create_opacity_map - [BOOLEAN] - enables (TRUE) or disables (FALSE) the creation of an opacity-map during the texture loading - default value: FALSE
  An opacity-map is a texture in which the value of the alpha channel is the arithmetic mean of the three R, G and B channels. This makes it possible to have an alpha value which varies according to the rgb color.
  
  
• create_binary_alpha_map - [BOOLEAN] - enables (TRUE) or disables (FALSE) the creation of a binary alpha-map during the texture loading - default value: FALSE
  A binary alpha-map is a texture in which the value of the alpha channel equals 1.0 or 0.0. The choice of the correct value depends on the comparison between the rgb color of a texel and the color specified with the binary_alpha_map_color element.
  
  
• create_constant_alpha_map - [BOOLEAN] - enables (TRUE) or disables (FALSE) the creation of a constant alpha-map during the texture loading - default value: FALSE
  A constant alpha-map is a texture in which the value of the alpha channel is equal to the value specified by the constant_alpha_map_value attribute.
  
  
• constant_alpha_map_value - [CLAMPED_REAL] - alpha value ranging between 0.0 and 1.0 for the creation of a constant alpha-map - default value: 0.5
  
  
• s3tc - [BOOLEAN] - enables (TRUE) or disables (FALSE) the creation of a compressed texture according to the S3TC (DXT1 or DXT5) compression diagram - default value: FALSE The texture will be compressed in the DXT1 format if the pixel format is RGB.
  The texture will be compressed in the DXT5 format if the pixel format is RGBA.
  
  
• num_mipmaps - [INTEGER] - makes it possible to specify the number of mipmaps to be created during the texture loading. In this case, filtering_mode must be set to "TRILINEAR " and the value of the anisotropic filtering set with the anisotropy attribute. A value of 0 disables the mipmapping - default value: 13
  
  
• play_avi - [BOOLEAN] - enables (TRUE) or disables (FALSE) the playing of an AVi animated texture - default value: TRUE

cube_map element

• face - [ENUM] - face of the cubemap:
  • XPOS (or right)
  • XNEG (or left)
  • YPOS (or top)
  • YNEG (or bottom)
  • ZPOS (or front)
  • ZNEG (or back)
• filename - [STR255] - location of the texture file of the cubemap face. The path is relating to the xml script.
• uv_map_op - [ENUM] - specifies the operation to be performed on the texture coordinates in order to correct the texture direction:
  • NO_SWAP: operation without effect: leave the textures coordinates unchanged - default value
  • SWAP_U: inverts the U coordinates: texture is reversed in relation to to the axis of the Y.
  • SWAP_V: inverts the V coordinates: texture is reversed in relation to to the x axis.
  • SWAP_UV: inverts the U and V coordinates: texture is reversed in relation to to the Y and X axis.

size element

• mirror texture
• cube map texture
• texture featuring Perlin noise or a chequered pattern

• width - [INTEGER] - texture width in pixels - default value: 256
• height - [INTEGER] - texture height in pixels - default value: 256

perlin_noise_params element

<texture create_binary_alpha_map="TRUE ">
	<binary_alpha_map_color r="1.0" g="0.0" b="0.0" epsilon="0.2" />
</texture>

<spin_value ang_z="0.1"  />