glpixeltransfer - Man Page
glPixelTransfer(3G) OpenGL Reference glPixelTransfer(3G)
NAME
glPixelTransferf, glPixelTransferi - set pixel transfer modes
C SPECIFICATION
void glPixelTransferf( GLenum pname,
GLfloat param )
void glPixelTransferi( GLenum pname,
GLint param )
PARAMETERS
pname Specifies the symbolic name of the pixel transfer parameter to be
set. Must be one of the following: GL_MAP_COLOR, GL_MAP_STENCIL,
GL_INDEX_SHIFT, GL_INDEX_OFFSET, GL_RED_SCALE, GL_RED_BIAS,
GL_GREEN_SCALE, GL_GREEN_BIAS, GL_BLUE_SCALE, GL_BLUE_BIAS,
GL_ALPHA_SCALE, GL_ALPHA_BIAS, GL_DEPTH_SCALE, GL_DEPTH_BIAS,
GL_POST_COLOR_MATRIX_RED_SCALE_SGI,
GL_POST_COLOR_MATRIX_GREEN_SCALE_SGI,
GL_POST_COLOR_MATRIX_BLUE_SCALE_SGI,
GL_POST_COLOR_MATRIX_ALPHA_SCALE_SGI,
GL_POST_COLOR_MATRIX_RED_BIAS_SGI,
GL_POST_COLOR_MATRIX_GREEN_BIAS_SGI,
GL_POST_COLOR_MATRIX_BLUE_BIAS_SGI, or
GL_POST_COLOR_MATRIX_ALPHA_BIAS_SGI.
param Specifies the value that pname is set to.
DESCRIPTION
glPixelTransfer sets pixel transfer modes that affect the operation of
subsequent glDrawPixels, glReadPixels, glCopyPixels, glTexImage1D,
glTexImage2D, and glTexImage3DEXT commands. The algorithms that are
specified by pixel transfer modes operate on pixels after they are read
from the frame buffer (glReadPixels and glCopyPixels) or unpacked from
client memory (glDrawPixels, glTexImage1D, and glTexImage2D). Pixel
transfer operations happen in the same order, and in the same manner,
regardless of the command that resulted in the pixel operation. Pixel
storage modes (see glPixelStore) control the unpacking of pixels being
read from client memory, and the packing of pixels being written back
into client memory.
Pixel transfer operations handle four fundamental pixel types: color,
color index, depth, and stencil. Color pixels are made up of four
floating-point values with unspecified mantissa and exponent sizes,
scaled such that 0.0 represents zero intensity and 1.0 represents full
intensity. Color indices comprise a single fixed-point value, with
unspecified precision to the right of the binary point. Depth pixels
comprise a single floating-point value, with unspecified mantissa and
exponent sizes, scaled such that 0.0 represents the minimum depth buffer
value, and 1.0 represents the maximum depth buffer value. Finally,
stencil pixels comprise a single fixed-point value, with unspecified
Page 1
glPixelTransfer(3G) OpenGL Reference glPixelTransfer(3G)
precision to the right of the binary point.
The pixel transfer operations performed on the four basic pixel types are
as follows:
Color Each of the four color components is multiplied by a scale
factor, then added to a bias factor. That is, the red
component is multiplied by GL_RED_SCALE, then added to
GL_RED_BIAS; the green component is multiplied by
GL_GREEN_SCALE, then added to GL_GREEN_BIAS; the blue
component is multiplied by GL_BLUE_SCALE, then added to
GL_BLUE_BIAS; and the alpha component is multiplied by
GL_ALPHA_SCALE, then added to GL_ALPHA_BIAS. After all four
color components are scaled and biased, each is clamped to
the range [0,1]. All color scale and bias values are
specified with glPixelTransfer.
If GL_MAP_COLOR is true, each color component is scaled by
the size of the corresponding color-to-color map, then
replaced by the contents of that map indexed by the scaled
component. That is, the red component is scaled by
GL_PIXEL_MAP_R_TO_R_SIZE, then replaced by the contents of
GL_PIXEL_MAP_R_TO_R indexed by itself. The green component
is scaled by GL_PIXEL_MAP_G_TO_G_SIZE, then replaced by the
contents of GL_PIXEL_MAP_G_TO_G indexed by itself. The blue
component is scaled by GL_PIXEL_MAP_B_TO_B_SIZE, then
replaced by the contents of GL_PIXEL_MAP_B_TO_B indexed by
itself. And the alpha component is scaled by
GL_PIXEL_MAP_A_TO_A_SIZE, then replaced by the contents of
GL_PIXEL_MAP_A_TO_A indexed by itself. All components taken
from the maps are then clamped to the range [0,1].
GL_MAP_COLOR is specified with glPixelTransfer. The contents
of the various maps are specified with glPixelMap.
If the SGI_color_table extension is supported,
GL_COLOR_TABLE_SGI is enabled, and the width of the color
table is non-zero, then its contents are used to replace a
subset of the components of each RGBA pixel group, based on
the internal format of the table. All the color table lookups
introduced by the SGI_color_table extension behave
identically: The color components are first rescaled to the
size of the color lookup table to form an index. Then a
subset of the components based on the internal format of the
table are replaced by the table entry selected by that index.
(See glColorTableSGI for more information.)
If the EXT_convolution extension is supported, and
GL_CONVOLUTION_1D_EXT, GL_CONVOLUTION_2D_EXT, or
GL_SEPARABLE_2D_EXT is enabled, then the pixels are convolved
with a filter kernel. (See glConvolutionFilter1DEXT,
glConvolutionFilter2DEXT, glSeparableFilter2DEXT, and
glEnable.) A color lookup is performed after convolution if
Page 2
glPixelTransfer(3G) OpenGL Reference glPixelTransfer(3G)
the SGI_color_table extension is supported and
GL_POST_CONVOLUTION_COLOR_TABLE_SGI is enabled.
If the SGI_color_matrix extension is supported, then the
color components of each pixel are modified by multiplying
them with the matrix at the top of the color matrix stack,
scaling them by specified values, and then biasing
(offsetting) them by specified values. That is, if the
original color components are (R, G, B, A), and the matrix at
the top of the color matrix stack is represented by m1, m2,
..., m16, then the modified color components (R', G', B', A')
are given by:
R' = ((m1*R) + (m5*G) + (m9*B) + (m13*A))*Rscale + Rbias
G' = ((m2*R) + (m6*G) + (m10*B) + (m14*A))*Gscale + Gbias
B' = ((m3*R) + (m7*G) + (m11*B) + (m15*A))*Bscale + Bbias
A' = ((m4*R) + (m8*G) + (m12*B) + (m16*A))*Ascale + Abias
The color matrix is specified using any of the OpenGL matrix
commands when the matrix mode (see glMatrixMode) is GL_COLOR.
The scale and bias terms are specified by executing
glPixelTransfer with pname equal to
GL_POST_COLOR_MATRIX_RED_SCALE_SGI,
GL_POST_COLOR_MATRIX_GREEN_SCALE_SGI,
GL_POST_COLOR_MATRIX_BLUE_SCALE_SGI,
GL_POST_COLOR_MATRIX_ALPHA_SCALE_SGI,
GL_POST_COLOR_MATRIX_RED_BIAS_SGI,
GL_POST_COLOR_MATRIX_GREEN_BIAS_SGI,
GL_POST_COLOR_MATRIX_BLUE_BIAS_SGI, or
GL_POST_COLOR_MATRIX_ALPHA_BIAS_SGI and param equal to the
desired value.
A color lookup is also performed after the color matrix
operation if the SGI_color_matrix extension is supported, and
GL_PROXY_POST_COLOR_MATRIX_COLOR_TABLE_SGI is enabled.
If the EXT_histogram extension is supported, and
GL_HISTOGRAM_EXT is enabled, the pixels are histogrammed.
See glHistogramEXT for more information.
If the EXT_histogram extension is supported, and
GL_MINMAX_EXT is enabled, the minimum and maximum pixel
values are recorded. See glMinMaxEXT for more information.
Color index Each color index is shifted left by GL_INDEX_SHIFT bits,
filling with zeros any bits beyond the number of fraction
bits carried by the fixed-point index. If GL_INDEX_SHIFT is
negative, the shift is to the right, again zero filled. Then
GL_INDEX_OFFSET is added to the index. GL_INDEX_SHIFT and
GL_INDEX_OFFSET are specified with glPixelTransfer.
Page 3
glPixelTransfer(3G) OpenGL Reference glPixelTransfer(3G)
From this point, operation diverges depending on the required
format of the resulting pixels. If the resulting pixels are
to be written to a color index buffer, or if they are being
read back to client memory in GL_COLOR_INDEX format, the
pixels continue to be treated as indices. If GL_MAP_COLOR is
true, each index is masked by 2n - 1, where n is
GL_PIXEL_MAP_I_TO_I_SIZE, then replaced by the contents of
GL_PIXEL_MAP_I_TO_I indexed by the masked value.
GL_MAP_COLOR is specified with glPixelTransfer. The contents
of the index map are specified with glPixelMap.
If the resulting pixels are to be written to an RGBA color
buffer, or if they are being read back to client memory in a
format other than GL_COLOR_INDEX, the pixels are converted
from indices to colors by referencing the four maps
GL_PIXEL_MAP_I_TO_R, GL_PIXEL_MAP_I_TO_G,
GL_PIXEL_MAP_I_TO_B, and GL_PIXEL_MAP_I_TO_A. Before being
dereferenced, the index is masked by 2n - 1, where n is
GL_PIXEL_MAP_I_TO_R_SIZE for the red map,
GL_PIXEL_MAP_I_TO_G_SIZE for the green map,
GL_PIXEL_MAP_I_TO_B_SIZE for the blue map, and
GL_PIXEL_MAP_I_TO_A_SIZE for the alpha map. All components
taken from the maps are then clamped to the range [0,1]. The
contents of the four maps are specified with glPixelMap.
The remaining pixel transfer operations described above are
then applied to the resulting RGBA components.
Depth Each depth value is multiplied by GL_DEPTH_SCALE, added to
GL_DEPTH_BIAS, then clamped to the range [0,1].
Stencil Each index is shifted GL_INDEX_SHIFT bits just as a color
index is, then added to GL_INDEX_OFFSET. If GL_MAP_STENCIL
is true, each index is masked by 2n - 1, where n is
GL_PIXEL_MAP_S_TO_S_SIZE, then replaced by the contents of
GL_PIXEL_MAP_S_TO_S indexed by the masked value.
The following table gives the type, initial value, and range of valid
values for each of the pixel transfer parameters that are set with
glPixelTransfer.
Page 4
glPixelTransfer(3G) OpenGL Reference glPixelTransfer(3G)
_______________________________________________________________________________
|_____________________________________|_________|_______________|_____________|
| GL_MAP_COLOR | Boolean | false | true/false |
| GL_MAP_STENCIL | Boolean | false | true/false |
| GL_INDEX_SHIFT | integer | 0 | (-oo,oo) |
| GL_INDEX_OFFSET | integer | 0 | (-oo,oo) |
| GL_RED_SCALE | float | 1.0 | (-oo,oo) |
| GL_GREEN_SCALE | float | 1.0 | (-oo,oo) |
| GL_BLUE_SCALE | float | 1.0 | (-oo,oo) |
| GL_ALPHA_SCALE | float | 1.0 | (-oo,oo) |
| GL_DEPTH_SCALE | float | 1.0 | (-oo,oo) |
| GL_RED_BIAS | float | 0.0 | (-oo,oo) |
| GL_GREEN_BIAS | float | 0.0 | (-oo,oo) |
| GL_BLUE_BIAS | float | 0.0 | (-oo,oo) |
| GL_ALPHA_BIAS | float | 0.0 | (-oo,oo) |
| GL_DEPTH_BIAS | float | 0.0 | (-oo,oo) |
| GL_POST_COLOR_MATRIX_RED_SCALE_SGI | float | 1.0 | (-oo,oo) |
|GL_POST_COLOR_MATRIX_GREEN_SCALE_SGI | float | 1.0 | (-oo,oo) |
|GL_POST_COLOR_MATRIX_BLUE_SCALE_SGI | float | 1.0 | (-oo,oo) |
|GL_POST_COLOR_MATRIX_ALPHA_SCALE_SGI | float | 1.0 | (-oo,oo) |
| GL_POST_COLOR_MATRIX_RED_BIAS_SGI | float | 0.0 | (-oo,oo) |
|GL_POST_COLOR_MATRIX_GREEN_BIAS_SGI | float | 0.0 | (-oo,oo) |
| GL_POST_COLOR_MATRIX_BLUE_BIAS_SGI | float | 0.0 | (-oo,oo) |
|GL_POST_COLOR_MATRIX_ALPHA_BIAS_SGI | float | 0.0 | (-oo,oo) |
|_____________________________________|_________|_______________|_____________|
glPixelTransferf can be used to set any pixel transfer parameter. If the
parameter type is Boolean, 0.0 implies false and any other value implies
true. If pname is an integer parameter, param is rounded to the nearest
integer.
Likewise, glPixelTransferi can also be used to set any of the pixel
transfer parameters. Boolean parameters are set to false if param is 0
and true otherwise. param is converted to floating point before being
assigned to real-valued parameters.
NOTES
If a glDrawPixels, glReadPixels, glCopyPixels, glTexImage1D, or
glTexImage2D command is placed in a display list (see glNewList and
glCallList), the pixel transfer mode settings in effect when the display
list is executed are the ones that are used. They may be different from
the settings when the command was compiled into the display list.
The color matrix may be used for simple linear color-space conversions.
However, it is also useful for ``swizzling'' (reordering or duplicating)
color components. This is helpful when processing image data that is not
in one of the native OpenGL formats, or when ``packing'' up to four
independent channels of luminance data into an RGBA drawable.
Page 5
glPixelTransfer(3G) OpenGL Reference glPixelTransfer(3G)
ERRORS
GL_INVALID_ENUM is generated if pname is not an accepted value.
GL_INVALID_OPERATION is generated if glPixelTransfer is executed between
the execution of glBegin and the corresponding execution of glEnd.
ASSOCIATED GETS
glGet with argument GL_MAP_COLOR
glGet with argument GL_MAP_STENCIL
glGet with argument GL_INDEX_SHIFT
glGet with argument GL_INDEX_OFFSET
glGet with argument GL_RED_SCALE
glGet with argument GL_RED_BIAS
glGet with argument GL_GREEN_SCALE
glGet with argument GL_GREEN_BIAS
glGet with argument GL_BLUE_SCALE
glGet with argument GL_BLUE_BIAS
glGet with argument GL_ALPHA_SCALE
glGet with argument GL_ALPHA_BIAS
glGet with argument GL_DEPTH_SCALE
glGet with argument GL_DEPTH_BIAS
glGet with argument GL_POST_COLOR_MATRIX_RED_SCALE_SGI
glGet with argument GL_POST_COLOR_MATRIX_GREEN_SCALE_SGI
glGet with argument GL_POST_COLOR_MATRIX_BLUE_SCALE_SGI
glGet with argument GL_POST_COLOR_MATRIX_ALPHA_SCALE_SGI
glGet with argument GL_POST_COLOR_MATRIX_RED_BIAS_SGI
glGet with argument GL_POST_COLOR_MATRIX_GREEN_BIAS_SGI
glGet with argument GL_POST_COLOR_MATRIX_BLUE_BIAS_SGI
glGet with argument GL_POST_COLOR_MATRIX_ALPHA_BIAS_SGI
MACHINE DEPENDENCIES
The following problems are known to exist on InfiniteReality systems.
All should be fixed in the next software release.
1. The GL_INDEX_SHIFT and GL_INDEX_OFFSET pixel transfer operations
are not applied properly in some cases when GL_MAP_COLOR is
enabled.
2. Incorrect results are produced if the I-to-I pixel map contains
entries with negative values.
3. Incorrect results are produced if the I-to-RGBA pixel map or the
RGBA-to-RGBA pixel maps contain values that are outside the
range [0.0,1.0].
SEE ALSO
glCallList, glCopyPixels, glDrawPixels, glNewList, glPixelMap,
glPixelStore, glPixelZoom, glReadPixels, glTexImage1D, glTexImage2D
Page 6