Name

    OES_vertex_half_float

Name Strings

    GL_OES_vertex_half_float

Contact

    Aaftab Munshi (amunshi@ati.com)

Notice

    Copyright (c) 2005-2013 The Khronos Group Inc. Copyright terms at
        http://www.khronos.org/registry/speccopyright.html

IP Status

    None.

Status

    Ratified by the Khronos BOP, July 22, 2005.

Version

    April 22, 2007 (version #3)

Number

    OpenGL ES Extension #38

Dependencies

    This extension is written against the OpenGL 2.0 specification

Overview

    This extension adds a 16-bit floating pt data type (aka half float) 
    to vertex data specified using vertex arrays.  The 16-bit floating-point
    components have 1 sign bit, 5 exponent bits, and 10 mantissa bits.
    
    The half float data type can be very useful in specifying vertex attribute 
    data such as color, normals, texture coordinates etc.  By using half floats 
    instead of floats, we reduce the memory requirements by half.  Not only does 
    the memory footprint reduce by half, but the memory bandwidth required for
    vertex transformations also reduces by the same amount approximately.
    Another advantage of using half floats over short/byte data types is that we 
    do not needto scale the data.  For example, using SHORT for texture coordinates 
    implies that we need to scale the input texture coordinates in the shader or 
    set an appropriate scale matrix as the texture matrix for fixed function pipeline.
    Doing these additional scaling operations impacts vertex transformation
    performance.

Issues

    1.  Should there be a half-float version of VertexAttrib{1234}[v] functions

    RESOLUTION: No.  
    
    There is no reason to support this, as these functions are not
    performance or memory footprint critical.  It is much more important that the
    vertex data specified using vertex arrays be able to support half float data
    format.

New Procedures and Functions

    None

New Tokens

    Accepted by the <type> parameter of VertexPointer, NormalPointer,
    ColorPointer, SecondaryColorPointer, IndexPointer, FogCoordPointer, 
    TexCoordPointer, and VertexAttribPointer

        HALF_FLOAT_OES      0x8D61

Additions to Chapter 2 of the OpenGL 2.0 Specification (OpenGL Operation)

    Add a new section 2.1.2.  This new section is copied from the 
    ARB_texture_float extension.

    2.1.2  16-Bit Floating-Point Numbers

    A 16-bit floating-point number has a 1-bit sign (S), a 5-bit
    exponent (E), and a 10-bit mantissa (M).  The value of a 16-bit
    floating-point number is determined by the following:

        (-1)^S * 0.0,                        if E == 0 and M == 0,
        (-1)^S * 2^-14 * (M / 2^10),         if E == 0 and M != 0,
        (-1)^S * 2^(E-15) * (1 + M/2^10),    if 0 < E < 31,
        (-1)^S * INF,                        if E == 31 and M == 0, or
        NaN,                                 if E == 31 and M != 0,

    where

        S = floor((N mod 65536) / 32768),
        E = floor((N mod 32768) / 1024), and
        M = N mod 1024.

    Implementations are also allowed to use any of the following
    alternative encodings:

        (-1)^S * 0.0,                        if E == 0 and M != 0,
        (-1)^S * 2^(E-15) * (1 + M/2^10),    if E == 31 and M == 0, or
        (-1)^S * 2^(E-15) * (1 + M/2^10),    if E == 31 and M != 0,

    Any representable 16-bit floating-point value is legal as input
    to a GL command that accepts 16-bit floating-point data.  The
    result of providing a value that is not a floating-point number
    (such as infinity or NaN) to such a command is unspecified, but
    must not lead to GL interruption or termination.  Providing a
    denormalized number or negative zero to GL must yield predictable
    results.

    Modifications to section 2.8 (Vertex Arrays)

    Add HALF_FLOAT_OES as a valid <type> value in Table 2.4.
        
    For <type> the values BYTE, SHORT, INT, FLOAT, and DOUBLE indicate
    types byte, short, int, float, and double, respectively; and the values
    UNSIGNED_BYTE, UNSIGNED_SHORT, and UNSIGNED_INT indicate types ubyte,
    ushort, and uint, respectively.  A <type> value of HALF_FLOAT_OES represents
    a 16-bit floating point number with 1 sign bits, 5 exponent bits,
    and 10 mantissa bits.


Errors

    None

New State

    None

Revision History

    June 15, 2005    Aaftab Munshi    First draft of extension.
    June 22, 2005    Aaftab Munshi    Renamed HALF_FLOAT token to HALF_FLOAT_OES
    #3, 2007/04/22   Jon Leech        Add version date and extension number.