Name

    SGIX_decimation

Name Strings

    GL_SGIX_decimation

Version

    $Date: 1997/07/18 00:49:00 $ $Revision: 1.4 $

Number

    125

Dependencies

    EXT_convolution affects the definition of this extension
    SGI_color_table affects the definition of this extension

Overview

    This extension adds a decimation operation to the pixel transfer
    path.  

    Decimation occurs after convolution and prior to the post-convolution
    color table.

    The operation is controlled by two positive integer parameters, stepx
    and stepy, that specify the decimation step size in the x and y
    directions, respectively.  During a pixel transfer, the decimation
    operation passes only those pixels which are at a position (i*stepx,
    j*stepy), for integers (i, j), relative to the bottom left corner of
    the image that is being transferred.  All other pixels are discarded. 
    Decimation is applied to the image that results from convolution,
    which may differ in size from the source image if convolution is
    enabled.

    By default, stepx and stepy are both 1.

Issues

    * This spec allows for future extension to 3D decimation.
    
    * What is a reasonable minimum maximum decimation step that we should
      enforce?  Two is probably sufficient.
      
    * Future extensions that redefine the convolution operation at the border
      of an image may affect this specification.
    
    * Decimation doesn't apply to bitmaps because they are not affected
      by the PixelTransfer state.  This is probably best.

    * Decimation is related to interlacing, as defined in the
      SGIX_interlace extension.  In fact, it provides the inverse of
      interlacing.  It appears useful to be able to enable decimation and
      interlacing on the same transfer.  Interactions between these
      two extensions should be examined further.

New Procedures and Functions

    None
    
New Tokens

    Accepted by the <pname> parameter of glPixelTransferf,
    glPixelTransferi, GetBooleanv, GetIntegerv,
    GetFloatv, and GetDoublev:

        DECIMATION_X_SGIX                     0xXXXX
        DECIMATION_Y_SGIX                     0xXXXX

    Accepted by the <pname> parameter of GetBooleanv, GetIntegerv,
    GetFloatv, and GetDoublev:

        MAX_DECIMATION_X_SGIX                 0xXXXX
        MAX_DECIMATION_Y_SGIX                 0xXXXX

Additions to Chapter 2 of the 1.0 Specification (OpenGL Operation)

    None

Additions to Chapter 3 of the 1.0 Specification (Rasterization)

    The specification of the decimation operation is added to the GL
    Specification in section 3.6.2, "Pixel Transfer Modes."  
    
    The decimation operation is added to the GL Specification in section
    3.6.3, "Rasterization of Pixel Rectangles, immediately following the
    operations described in EXT_convolution, and prior to the
    post-convolution color table operation that is described in the
    SGI_color_table extension.  

    Decimation operates on the image of pixels resulting from the
    convolution stage of the pixel transfer path.  If stepx and stepy are
    the  current decimation step parameters, then the decimation
    operation passes only those pixels which are at a position (i*stepx,
    j*stepy), for integers (i, j), relative to the bottom left corner of
    the image.   All other pixels are discarded. If the image size prior
    to decimation is (nx, ny), where nx >= 1 and ny >= 1, then its size
    after decimation will be (floor((nx-1)/stepx) + 1,
    floor((ny-1)/stepy) + 1)).

    Decimation is applied to the image that results from convolution,
    which may differ in size from the source image if convolution is
    enabled.

    In cases where errors can result from the specification of invalid
    image dimensions, it is these resulting dimensions that are tested,
    not the dimensions of the source image.

    If this extension is implemented, decimation is always active, and is
    applied to any pixel transfer, regardless of the command used to
    generate it.  (There is no enable or disable of decimation.)       

    The x and y decimation parameters are specified by calling
    PixelTransfer with <pname> set to DECIMATION_X_SGIX, and
    DECIMATION_Y_SGIX, respectively.  Each parameter value must be a
    positive integer less than or equal to MAX_DECIMATION_X_SGIX and
    MAX_DECIMATION_Y_SGIX, respectively.

Additions to Chapter 4 of the 1.0 Specification (Per-Fragment Operations
and the Framebuffer)

    The operation of decimation during pixel copy and query operations is
    identical to the operation during pixel drawing and texture image
    definition.  This operation occurs immediately after the operations
    described by EXT_convolution, and prior to the post-convolution color
    table operation that is described in the SGI_color_table extension,
    which follow section 4.3.2 (Reading Pixels) of the GL Specification.

Additions to Chapter 5 of the 1.0 Specification (Special Functions)

    None

Additions to Chapter 6 of the 1.0 Specification (State and State Requests)

    The x decimation step parameter is queried using GetIntegerv with
    <pname> set to DECIMATION_X_SGIX.

    The y decimation step parameter is queried using GetIntegerv with
    <pname> set to DECIMATION_Y_SGIX.

    The maximum allowed x decimation step parameter is queried using
    GetIntegerv with <pname> set to MAX_DECIMATION_X_SGIX.

    The maximum allowed y decimation step parameter is queried using
    GetIntegerv with <pname> set to MAX_DECIMATION_Y_SGIX.

Additions to the GLX Specification

    None
    
GLX Protocol

    None 
                    
Dependencies on EXT_convolution

    The decimation operation follows the convolution operation (and its
    scale and bias).  If convolution is not supported, the location of
    the decimation operation with respect to all other pixel operations
    remains the same.

Dependencies on SGI_color_table

    The decimation operation precedes the post-convolution color table
    operation.  If the post-convolution color table operation is not
    supported, the location of the decimation operation with respect to
    all other pixel operations remains the same.

Errors

    INVALID_VALUE is generated if PixelTransfer is called with parameter
    <pname> equal to DECIMATION_X_SGIX and parameter <param> less
    than 1 or greater than MAX_DECIMATION_X_SGIX as queried using
    GetIntegerv with <pname> MAX_DECIMATION_X_SGIX.

    INVALID_VALUE is generated if PixelTransfer is called with parameter
    <pname> equal to DECIMATION_Y_SGIX and parameter <param> less
    than 1 or greater than MAX_DECIMATION_Y_SGIX as queried using
    GetIntegerv with <pname> MAX_DECIMATION_Y_SGIX.

New State

    Get Value                           Get Command                     Type            Initial Value           Attribute
    ---------                           -----------                     ----            -------------           ---------
    DECIMATION_X_SGIX                    GetIntegerv                     Z+              1                       pixel
    DECIMATION_Y_SGIX                    GetIntegerv                     Z+              1                       pixel

New Implementation Dependent State

    Get Value                           Get Command                     Type            Minimum Value
    ---------                           -----------                     ----            -------------
    MAX_DECIMATION_X_SGIX                GetIntegerv                     Z+              2
    MAX_DECIMATION_Y_SGIX                GetIntegerv                     Z+              2